Exocytic plasma membrane flows remodel endoplasmic reticulum–plasma membrane tethering for septin collar assembly

Abstract

Endoplasmic reticulum (ER)–plasma membrane (PM) tethering is crucial for the non-vesicular lipid transport between the ER membrane and the PM. However, the PM-associated ER can impede the PM binding of cytoskeletons and other organelles. It is poorly understood how the competition between the ER and cytoskeletons/organelles on the PM is resolved. Here, we show that, upon septin collar assembly, ER-PM tethering proteins are excluded from the yeast bud sites, and the PM-associated ER is locally detached from the PM. Our results suggest that PM flows by polarized exocytosis extrude PM proteins, including ER-PM tethering proteins, from the bud sites. When the reorganization of the ER-PM tethering was inhibited by exocytosis repression, septin localization was restricted to the PM sites poor in ER-PM tethering proteins. This study proposes machinery reconciling ER-septin competition on the PM, providing mechanistic insights into the spatial organization of PM-associated organelles and cytoskeletons.

Exocytic PM flows resolve competition between large structures on cell membrane.

INTRODUCTION

In eukaryotic cells, a portion of the endoplasmic reticulum (ER) is attached to the plasma membrane (PM) by ER-PM tethering proteins and protein complexes that are bound to and/or inserted into both ER membrane and PM (1–5). The ER-PM tethering is crucial for the non-vesicular lipid transport between the ER membrane and the PM and for the regulation of intracellular ion homeostasis (4–6). Recent studies have shown that the ER-PM tethering competes with binding to and/or assembly on the PM of large structures including cytoskeletons, organelles, and membrane invaginations (7–12).

Septins are membrane-bound cytoskeletons conserved in a wide range of eukaryotes except for plants (13–15). Septins bind to the PM and contribute to its curvature, serving as a scaffold for the proteins including cytokinesis-related proteins and cell polarity regulators (13–19). In budding yeast, septins initially form diffusible patches at the presumptive bud-forming sites and mature into immobile collars of ~1-μm diameter at the mother-bud necks (18, 20, 21). Recently, septins were shown to localize to the ER-free region of the PM and restrict the movement of the PM-associated ER, suggesting physical competition between these two structures (22, 23). A significant part, 20 to 50% in budding yeasts, of the PM is covered by the ER (24, 25). The PM-associated ER has small gaps (24), and endocytic PM invagination of ~50-nm diameter was restrictively formed in the ER gaps (9, 26). However, it is unclear how the large septin collars assemble at the selected bud sites precisely, avoiding interference from the PM-associated ER.

In this study, we investigated the ER dynamics during septin collar assembly at the yeast bud sites. Our results suggest that PM flows by polarized exocytosis extrude PM proteins, including ER-PM tethering proteins, that preexist at the bud sites. The exclusion of ER-PM tethers resolves the competition between the ER and septins at the bud sites, enabling intact septin collar assembly. These results provide mechanistic insights into the reconciliation of organelle-cytoskeleton competition on the PM.

RESULTS

Competition for PM binding between the ER and septins

To determine whether PM-associated ER impedes the PM binding of septins, we attempted to strengthen the ER-PM tethering by overexpressing ER-PM tethering proteins. Ist2 is an endogenous ER-PM contact site protein that tethers the ER to the PM by its ER membrane–spanning domains and PM-binding polybasic motif (Fig. 1A) (3, 27, 28). We expressed Ist2 fused with mScarlet-I (mSc-I) from the inducible GAL1 promoters in high-copy plasmids. Induction of mSc-I–IST2 expression expanded the cortical ER and mislocalized an mNeonGreen (mNG)–tagged septin subunit, Cdc10-mNG, to the cytoplasm (Fig. 1, A to D, and fig. S1, A and B). By contrast, expression of either a tandem repeat of the polybasic motif of Ist2 (PB^Ist2×2), which strongly binds to the PM but not to the ER (3), or Ist2 lacking PB motif (Ist2ΔPB) resulted in little or no cytoplasmic localization of septins (Fig. 1, A to D). The expression level of the truncated proteins was higher than or comparable to that of mSc-I–Ist2 (fig. S1A). Thus, overexpression of mSc-I–Ist2 induces cytoplasmic localization of septins, which requires both PM-binding PB motif and ER membrane–spanning non-PB domain. We also tested whether overexpression of an artificial ER-PM tethering protein, Sec71TMD–mSc-I–PH^PLCδ×2, altered septin localization. This protein is composed of an ER transmembrane domain (TMD) of Sec71 (29), mSc-I, and a tandem repeat of rat phospholipase δ (PLCδ) Pleckstrin-homology (PH) domain that binds to phosphatidyl inositol 4,5-bisphosphate in the PM (30, 31). By its ER-membrane spanning and PM-binding domains, this protein was expected to bind to both ER membrane and PM (Fig. 1A). When expressed from chromosomal TEF1 promoters, Sec71TMD–mSc-I–PH^PLCδ×2 localized to the cell periphery and overlapped with a part of the cortical ER, suggesting its localization to the PM-associated ER (fig. S1C). Expression of this protein from the GAL1 promoters in high-copy plasmids expanded the peripheral ER and increased cytoplasmic localization of septins (Fig. 1, A to D, and fig. S1, A and B). Next, we examined whether the overexpression of ER-PM tethers blocks septin binding to the PM before budding or detaches septins from the bud necks of budded cells. At the point of 8-hour expression of mSc-I–Ist2 or Sec71TMD–mSc-I–PH^PLCδ×2, 72 or 60% of the cells with cytoplasmic septins was unbudded, respectively (n = 173 and 105). In addition, time-lapse imaging revealed that 81% of the cells in which de novo cytoplasmic septins emerged was unbudded (n = 31) (Fig. 1E). These data suggest that overexpression of ER-PM tethers blocked septin-PM associations in unbudded cells. Furthermore, we observed stronger growth defects for cells expressing mSc-I–Ist2 or Sec71TMD–mSc-I–PH^PLCδ×2 than the cells expressing their truncates (fig. S1D), which is consistent with the essentiality of septins in vegetative growth (32, 33). Together, these results suggest that septins and the PM-associated ER compete with each other for binding to the PM.

Fig. 1. Septin dissociation from the PM by overexpression of ER-PM tethering proteins.

(A) Localization of septins under the overexpression of ER-PM tethering proteins or their truncates. Each mSc-I–tagged protein was expressed from the GAL1 promoters in high-copy 2-μm plasmids. After transcription activation of the GAL1 promoters by galactose addition, cells were incubated for 8 hours. Septins were visualized with Cdc10-mNG. We note that mSc-I–PB^Ist2×2 localized to nucleus-like structures in addition to the cell periphery and Sec71TMD–mSc-I was enriched in nuclear envelope-like structures rather than whole ER. Arrowheads indicate cytoplasmic septins. Scale bar, 2 μm. (B to D) Percentage of cells with cytoplasmic septins. ER-PM tethering proteins or their truncates were expressed from the GAL1 promoters in high copy plasmids for 4 hours (B), 6 hours (C), or 8 hours (D). mSc-I–positive cells (≥30) were examined in each experiment. Data are represented as means + SEM. n = 5. ***P < 0.001 (Tukey-Kramer test). (E) Maximum intensity projection images of Cdc10-mNG under mSc-I–Ist2 overexpression. Arrowheads indicate cytoplasmic septins. Scale bar, 2 μm. DIC, Differential Interference Contrast.

Local ER detachment from the bud site PM

Bud neck localization of septins is crucial for cytokinesis and regulation of cell polarity (13, 18, 32, 34). Therefore, we speculated that there are machineries that resolve the ER-septin competition and enable septin collar formation at the ER-attached PM. To examine whether ER-PM tethering is remodeled at the bud sites, we observed ER-PM tethering proteins, Ist2 and Tcb3 (3, 27), tagged with mNG during polarity establishment. An mSc-I–fused exocyst subunit, Exo84–mSc-I, was used as a marker of cell polarity establishment. Both mNG-Ist2 and Tcb3-mNG were excluded from the haploid bud sites upon Exo84–mSc-I polarization (Fig. 2, A to C, and fig. S2A). Furthermore, Sec71TMD–mSc-I–PH^PLCδ×2 was also excluded from the bud sites (Fig. 2D and fig. S2B). Next, we visualized the whole ER with mNG-fused Hmg1 truncate (Hmg1^1-702) (35), and the ER was dissociated from the Exo84–mSc-I–accumulating cell periphery (Fig. 2, E and F). We also performed dual-color imaging of mNG-Ist2 and Cdc10–mSc-I, and the exclusion of mNG-Ist2 preceded the accumulation of Cdc10–mSc-I at the bud sites (Fig. 2G and fig. S2C). Together, these results indicate that ER-PM tethering proteins are excluded from the bud sites, which presumably dissociate the PM-associated ER from the PM and resolves ER-septin competition.

Fig. 2. Local ER detachment from the PM at the bud site.

(A) The behavior of mNG-Ist2 at the Exo84–mSc-I–accumulating bud site. An arrowhead indicates the exclusion of mNG-Ist2 from the bud sites. Scale bar, 2 μm. (B) Fluorescence intensity change of mNG-tagged ER-PM tethering proteins at the bud sites and non-bud sites during Exo84–mSc-I polarization. n = 16 (Ist2) and 15 (Tcb3). ***P < 0.001 (paired t test). (C) Changes in mNG-Ist2 gap sizes at the bud sites during Exo84–mSc-I polarization. n = 16. ***P < 0.001 (paired t test). (D) Fluorescence intensity change of Sec71–mSc-I–PH^PLCδ×2 at the bud sites and non-bud sites during Exo84-mNG polarization. n = 14. ***P < 0.001 (paired t test). (E) The behavior of ER (Hmg1^1-702–mNG) at the Exo84–mSc-I–accumulating bud sites. Arrowheads indicate detachment of the ER from the PM. Scale bar, 2 μm. (F) Fluorescence intensity change of Hmg1^1-702–mNG at the bud sites and non-bud sites during Exo84–mSc-I polarization. n = 16. ***P < 0.001 (paired t test). (G) Three-dimensional (3D) reconstructed time-lapse images of mNG-Ist2 and Cdc10–mSc-I. An arrowhead and arrows indicate exclusion of mNG-Ist2 and accumulation of Cdc10–mSc-I, respectively. Scale bar, 2 μm.

Involvement of exocytosis for exclusion of ER-PM tethering proteins from the bud sites

We found that the exclusion of mNG-Ist2 was also observed at other polarity sites; proximal and distal bud sites in diploid cells, shmoo tips, and repair sites of local cell peripheral damage induced by intense laser irradiation (fig. S3, A to D). This raised the possibility that the ER was detached from the PM by machineries that were commonly utilized in the polarity establishment and/or maintenance at these sites. The PM-associated ER is detached from the PM at sites with active exocytosis in fission yeasts, although how exocytosis remodels the PM-associated ER is unclear (7, 36). Because exocytosis is also highly active at the polarity sites of budding yeasts (Fig. 2, A and E, and fig. S3) (34, 37, 38), we assessed whether exocytosis is involved in the exclusion of ER-PM tethering proteins.

We repressed exocytosis by degrading an exocyst subunit, Sec6, with an auxin-inducible degron (AID), which induces extensive vesicle accumulation in the cytoplasm (39–41). Tandem triple repeats of mini AID (mAID×3) were introduced into the C terminus of Sec6. In the presence of auxin analog 5-Phenyl-1H-indole-3-acetic acid (5-Ph-IAA), mAID tags are ubiquitylated by exogenously expressed ubiquitin E3 ligase OsTIR1^F74G, and the ubiquitylated proteins are degraded by the proteasome (39, 40). The addition of 5-Ph-IAA induced degradation of Sec6-mAID×3, resulting in severe growth defects (fig. S4, A to D).

Using this system, we tested whether exocytosis inhibition mitigated the exclusion of ER-PM tethers. Because only 20 to 50% of the PM is covered by the ER (24, 25), numerous preexisting cortical ER gaps make it difficult to quantify the de novo gap formation under exocytosis inhibition that perturbs the localization of cell polarity markers (42–44). To overcome this problem, we artificially increased ER-PM tethering by expressing mNG-Ist2 from the strong TEF1 promoter so that mNG-Ist2 covered the entire PM (Fig. 3A) (27). After the release from the α factor–induced G1 arrest, we observed the gap formation of mNG-Ist2. The sizes of de novo gaps in 5-Ph-IAA–treated cells were smaller than dimethyl sulfoxide (DMSO)–treated cells (Fig. 3, A and B). Because exocytosis inhibition prevents cell growth, cell cycle synchronization might be inefficient in Sec6-mAID×3–degraded cells. To exclude the possibility that the smaller gap formation in Sec6-mAID×3–degraded cells was due to inefficient cell cycle synchronization, we also examined shorter α factor treatment in control (DMSO-treated) cells, and similar results were obtained to long α factor treatment (Fig. 3B). To test whether the smaller gap formation in Sec6-mAID×3–degraded cells can be explained by the disturbed polarity, we observed mNG-Ist2 gap formation and/or expansion at the polarity sites that were visualized with an mSc-I–fused polarisome component, Spa2 (Fig. 3, C and D). Sec6-mAID×3 degradation declined populations of cells with single polarity sites from 91 to 67%, suggesting the partial disturbance of cell polarity (n = 23 and 27). Therefore, we quantified the gap sizes in cells with single polarity sites, and the mNG-Ist2 gaps of 5-Ph-IAA–treated cells were smaller than that of DMSO-treated cells (Fig. 3D). These results suggest that exocytosis is involved in the efficient exclusion of ER-PM tethering proteins from the bud sites.

Fig. 3. Involvement of exocytosis for exclusion of ER-PM tethering proteins from the bud sites.

(A) 3D reconstructed time-lapse images of mNG-Ist2 under the repression of exocytosis by Sec6-mAID×3 degradation. mNG-IST2 was expressed from TEF1 promoters. Cells were treated with α factor for 4 hours and incubated with dimethyl sulfoxide (DMSO) or 5 μM 5-Ph-IAA for the final hour of the α factor treatment. After the washout of α factor and the addition of pronase E to digest remaining α factor, the cells were observed with a microscope. Scale bar, 2 μm. (B) Maximum diameter of de novo mNG-Ist2 gap formed in each cell during 1 hour of observation. For short synchronization, cells were treated with α factor for 3 hours instead of 4 hours. Data are represented as means + SEM. n = 55 (DMSO), 60 (DMSO, short sync), and 58 (5-Ph-IAA). ***P < 0.001 (Tukey-Kramer test). (C) 3D reconstructed time-lapse images of mNG-Ist2 and Spa2–mSc-I under the repression of exocytosis. mNG-IST2 was expressed from TEF1 promoters. Cell cycle was synchronized with α factor, and exocytosis was inhibited by degrading Sec6-mAID×3 with 5 μM 5-Ph-IAA. Scale bar, 2 μm. (D) Maximum diameter of mNG-Ist2 gaps at the Spa2–mSc-I–polarized sites. The cells with single Spa2 polarization sites were examined. Data are represented as means + SEM. n = 21 (DMSO) and 18 (5-Ph-IAA). ***P < 0.001 (Welch’s t test).

Exocytosis-dependent exclusion of PM-associated proteins from the bud sites

Exocytosis locally remodels the PM by supplying membrane lipids and proteins from secretory vesicles (37, 45). Upon the vesicle fusion, PM tension stretches the fused vesicles (45, 46). This stretch generates the PM flows and extrudes mobile PM lipids and proteins that preexist at the exocytic sites (45, 47–49). Thus, we hypothesized that the exocytic PM flows also contribute to the exclusion of ER-PM tethering proteins from the bud sites.

To test whether PM binding is sufficient for the exclusion, we first observed multiple PM proteins, including PM-binding motif of Ist2, during Exo84–mSc-I polarization. Two PM-spanning proteins, Pma1-mNG and Pdr5-mNG, were excluded from the bud sites before visible bud emergence (Fig. 4, A and C, and fig. S5, A and C). However, contrary to our hypothesis, the fluorescence intensity of mNG-PB^Ist2×2 at the bud sites did not change during Exo84–mSc-I accumulation (Fig. 4, B and C, and fig. S5C). Two other PM-associating domains, an mNG-tagged C-terminal palmitoylation domain of Ras2, mNG-Ras2C, and mSc-I–PH^PLCδ×2 also did not show fluorescence decline at the bud sites (Fig. 4C and fig. S5, B and D). Therefore, some but not all of the PM-binding proteins were efficiently excluded from the bud sites.

Fig. 4. Exclusion of PM-associating proteins from the bud sites.

(A) The behavior of Pma1-mNG at the Exo84–mSc-I–accumulating bud site. Arrowheads indicate the exclusion of Pma1-mNG from the bud site. Scale bar, 2 μm. (B) The behavior of mNG-PB^Ist2×2 at the Exo84–mSc-I–accumulating bud site. Scale bar, 2 μm. (C) Fluorescence intensity change of PM proteins at the bud sites and non-bud sites during Exo84–mSc-I polarization. n = 16 (Pma1), 15 (Pdr5), 20 (PB^Ist2×2), and 15 (Ras2C). ***P < 0.001 (paired t test). (D) Fluorescence recovery after photobleaching (FRAP) of mNG-tagged PM proteins. Data are represented as means ± SEM. n = 20. (E and F) Halftime (E) and mobile fractions (F) of fluorescence recovery in (D). Data are represented as means + SEM. ***P < 0.001 (Tukey-Kramer test). (G to I) FRAP of Cry2PHR^WT- or Cry2PHR^D387A–mSc-I–PB^Ist2×2 with or without 488-nm light irradiation. Data are represented as means ± SEM for (G) or means + SEM for (H and I). n = 15. ***P < 0.001 (Tukey-Kramer test). (J) The behavior of Cry2PHR^WT–mSc-I–PB^Ist2×2 with or without 488-nm light irradiation. An arrowhead indicates the exclusion of Cry2PHR^WT–mSc-I–PB^Ist2×2 from the bud site. We note that Cry2PHR^WT–mSc-I–PB^Ist2×2 in nucleus-like structures formed foci upon 488-nm light irradiation. Scale bar, 2 μm. (K) Fluorescence intensity change of Cry2PHR^WT- or Cry2PHR^D387A-fused mSc-I–PB^Ist2×2 and mSc-I–Ras2C at the bud sites and non-bud sites during Exo84-mNG polarization. Cells were irradiated with 488-nm light. n = 6 [wild type (WT), PB^Ist2×2], 5 (D387A, PB^Ist2×2), 7 (WT, Ras2C), and 6 (D387A, Ras2C). **P < 0.01 (paired t-test).

Next, we investigated the possibility that mNG-PB^Ist2×2, mNG-Ras2C, and mSc-I–PH^PLCδ×2 were excluded from the bud sites but their exclusions were undetectable due to fast diffusion. Because the ER detaches from the PM at the bud sites (Fig. 1D), ER-PM tethering proteins would not be able to return to the bud sites after the exclusion. In contrast, the exclusion of ER-unbound PM proteins may not be detectable due to lateral diffusion if their diffusion rate is fast enough. We found, by fluorescence recovery after photobleaching (FRAP), that the diffusion of these proteins was faster than that of the excluded PM proteins (Fig. 4, D to F, and fig. S6, A to C). To further elucidate the effects of diffusion on the exclusion, we attempted to slow down the diffusion of PB^Ist2×2 and Ras2C using optogenetic homo-oligomerization. These PM-associating domains were fused with the photolyase homology region (PHR) of Arabidopsis Cry2 (Cry2PHR) that oligomerizes in response to blue light irradiation (50, 51). FRAP experiments confirmed that diffusion of Cry2PHR–mSc-I–PB^Ist2×2 and Cry2PHR–mSc-I–Ras2C was slowed down in response to blue light (Fig. 4, G to I, and fig. S6, D to G). Time-lapse imaging revealed that oligomerized, but not monomeric, Cry2PHR–mSc-I–PB^Ist2×2 and Cry2PHR–mSc-I–Ras2C were excluded from the bud emergence sites (Fig. 4J). We also performed dual-color imaging using Exo84-mNG, and both oligomerized Cry2PHR–mSc-I–PB^Ist2×2 and Cry2PHR–mSc-I–Ras2C were excluded from the Exo84-mNG–accumulating sites (Fig. 4K and fig. S6H). The exclusion was not observed for a light-insensitive D387A mutant (51, 52) of Cry2PHR (Fig. 4K and fig. S6H). These results suggest that PM-localizing proteins including PM-binding motifs of Ist2 are excluded from the bud sites when their lateral diffusion is negligible.

We also assessed whether the exclusion of the PM proteins required exocytosis. Exocytosis inhibition by Sec6 degradation reduced de novo gap sizes of oligomerized Cry2PHR–mSc-I–PB^Ist2×2 (Fig. 5, A and B). Next, we observed Cry2PHR–mSc-I–PB^Ist2×2 gap dynamics at the Spa2-mNG–accumulating polarity sites (Fig. 5, C and D). Sec6-mAID×3 degradation declined populations of cells with single polarity sites from 97 to 42% (n = 30 and 52). The Cry2PHR–mSc-I–PB^Ist2×2 gaps of 5-Ph-IAA–treated cells with single polarity sites were smaller than that of DMSO-treated cells (Fig. 5, C and D). Together, these data support our notion that exocytic PM flows exclude ER-PM tethering proteins from the bud sites.

Fig. 5. Involvement of exocytosis for exclusion of the PM-associating proteins from the bud sites.

(A) 3D reconstructed time-lapse images of Cry2PHR^WT–mSc-I–PB^Ist2×2 under the repression of exocytosis. Cell cycle was synchronized with α factor, and exocytosis was inhibited by degrading Sec6-mAID×3 with 5 μM 5-Ph-IAA. The cells were irradiated by 488-nm laser for the oligomerization of Cry2PHR. Scale bar, 2 μm. (B) Maximum diameter of de novo Cry2PHR^WT–mSc-I–PB^Ist2×2 gap formed in each cell during 1 hour of observation. Data are represented as means + SEM. n = 43 (DMSO), 43 (DMSO, short sync), and 59 (5-Ph-IAA). ***P < 0.001 (Tukey-Kramer test). (C) 3D reconstructed time-lapse images of Cry2PHR^WT–mSc-I–PB^Ist2×2 and Spa2-NG under the repression of exocytosis. Cell cycle was synchronized with α factor, and exocytosis was inhibited by degrading Sec6-mAID×3 with 5 μM 5-Ph-IAA. Scale bar, 2 μm. (D) Maximum diameter of Cry2PHR^WT–mSc-I–PB^Ist2×2 gaps at the Spa2-mNG–polarized sites. The cells with single Spa2 polarization sites were examined. n = 29 (DMSO) and 22 (5-Ph-IAA). ***P < 0.001 (Welch’s t test).

Exocytosis-dependent reconciliation of ER-septin competition on the PM

To examine whether the ER remodeling by exocytosis is required for proper septin patch localization, we observed mNG-Ist2 and septins under exocytosis inhibition. Because strong inhibition of exocytosis abolishes septin polarization to the bud sites (43), we mildly induced the degradation of Sec6-mAID×3 with a low concentration of 5-Ph-IAA (fig. S4, C and D). Spa2–mSc-I–positive polarity sites were established at single sites in 67% of the cells under the mild Sec6-mAID×3 degradation condition, compared to 97% under DMSO treatment (n = 30 and 30) (fig. S7). In this condition, Cdc10–mSc-I restrictively distributed at the area with weak mNG-Ist2 signals (Fig. 6A and fig. S8). We note that Cdc10–mSc-I accumulated at the multiple sites of the cell periphery in 45% (n = 20) of the cells with single Spa2-mNG–polarized sites under the Sec6-mAID×3 degradation (figs. S7 and S8). These results support our model that the remodeling of ER-PM tethers by active exocytosis is involved in large septin collar formation when the PM at the selected bud sites is covered with the ER.

Fig. 6. Resolution of competition between septins and PM-associated ER by exocytosis.

(A) 3D reconstructed time-lapse images of mNG-Ist2 and septins under exocytosis inhibition. Cell cycle was synchronized with α factor, and exocytosis was inhibited by Sec6-mAID×3 degradation with 100 nM 5-Ph-IAA. Scale bar, 2 μm. (B) A model for the resolution of ER-septin competition at the bud sites. PM flows by exocytosis exclude ER-PM tethering proteins from the bud sites, resulting in the local ER detachment from the PM. Then, septins form a collar at the ER-detached PM.

DISCUSSION

In this study, we propose a reconciliation mechanism of ER-septin competition at the yeast bud sites. We demonstrated that PM remodeling by polarized exocytosis excluded ER-PM tethering proteins from the bud sites. Our results also suggest that the exclusion enables septins to assemble collar structures without interference from the PM-associated ER.

ER-PM tethering is dynamically reorganized by changes in PM lipid composition and posttranslational modification–or Ca²⁺-dependent alteration of lipid-binding affinity of the tethering proteins (4, 6, 53). We revealed that PM-localized proteins, including the oligomerized PM-binding motifs of Ist2, were excluded from the bud sites, indicating that PM binding is crucial for the exclusion of ER-PM tethering proteins. The PM proteins that were confirmed to be excluded had several PM-association properties: PM spanning, phosphoinositide binding (28), and PM insertion of the palmitoyl group, suggesting the non-selectivity of the exclusion. Therefore, PM remodeling rather than changes in lipid-binding affinities of the tethering proteins is likely to be responsible for the exclusion. Furthermore, we showed that oligomeric forms, but not monomers, of Cry2PHR–mSc-I–PB^Ist2×2 and Cry2PHR–mSc-I–Ras2C were excluded from the bud sites. Because both monomers and oligomers of these proteins have the same PM association domains, changes in lipid composition by exocytosis do not explain the exclusion of ER-PM tethers.

Alternatively, we here propose that exocytic PM flows exclude ER-PM tethering proteins from the yeast bud sites (Fig. 6B). In the PM of fungal and plant cells, turgor pressure creates equilibrant membrane tension (45). Upon exocytosis, membrane of PM-fused secretory vesicles is stretched by the PM tension, and the lipids and proteins present at the vesicle-binding sites of the PM are pushed away (45, 47). As exocytic PM flows non-selectively extrude mobile lipids and proteins, the bulk flows are expected to extrude ER-PM tethers as well, thereby detaching the ER from the PM.

The PM-associated ER also blocks the PM binding of secretory vesicles (7). The diameter of secretory vesicle is around 80 to 100 nm (54), which is smaller than the diameter of septin ring and hourglass, ~1 μm (18). The ER forms highly curved tubular networks (24, 55), and preexisting small ER gaps may enable the association between the secretory vesicles and the PM. It is also possible that unknown mechanisms make small ER gaps at the polarity sites before polarized exocytosis. Intriguingly, in a mutant defective in actin cable formation, septins bind to the bud site PM before secretory vesicle accumulation (56). Because septins form uneven ring-like structures in the mutant, septins might bind to the preexisting ER gaps before the exocytosis-mediated exclusion of ER-PM tethers.

After polarity establishment, a portion of the ER is transported to the bud tips by myosin, and the ER at the bud tips is tethered to the polarisome by the interactions between ER-membrane–spanning Scs2 and a polarisome component, Epo1 (57, 58). Exocytosis is active at the bud tips during tip growth (37), although ER-tethering polarisome is concentrated at the bud tips by myosin despite of PM flows (59). One possible explanation is that Epo1 tethers the ER to the bud tips but not in proximity to the PM. This is consistent with the previous research showing that small area of the PM in small buds is covered by the ER (24). This may enable active exocytosis at the bud tips without interference from the ER after budding.

We showed that exocytosis repression inhibited the exclusion of ER-PM tethers even in cells with Spa2 polarized in single sites. This supports our model that PM flows rather than polarity establishments detach the ER from the PM. However, we note that some cells have multiple polarity sites under exocytosis inhibition, and the polarity might be partly disturbed even in cells with single polarity sites. Thus, it is still possible that disturbed polarity or other downstream pathways of polarized exocytosis altered the ER detachments from the PM.

While PM tension is often not equilibrated in animal cells, PM flows by exocytosis were reported in specific subcellular domain, such as front of migrating cells and growth cone of neuronal axon (45, 48, 49, 60). Thus, we speculate that exocytic PM flows remodel the ER-PM tethering in a wide range of eukaryotic species.

Our findings suggest that the exocytosis-dependent remodeling of ER-PM tethering is crucial to reconcile the ER-septin competition at the yeast bud sites. Septins bend the membrane and work as scaffolds for polarity factors and cytokinesis-related proteins, which are crucial for cellular morphogenesis, polarity establishment and maintenance, and cytokinesis (13–19). Furthermore, we previously reported that the PM-associated ER was excluded from the yeast bud necks when septin collars were intact (22). This suggests that after exocytosis-dependent ER detachments, the septin collars prevent the ER attachments to the PM at the bud necks. Upon cytokinesis, septins split into double rings, and contractile actomyosin rings are formed between the double rings (13, 34). As actomyosins exclusively bind to the ER-unbound PM at least in fission yeasts (8, 36), the ER-free PM maintained by septins would help efficient contractile ring assembly, although we note that the requirement of intact septin double rings for cytokinesis is argued (61, 62).

More broadly, the PM-associated ER restricts the PM binding and assembly on the PM of various organelles, cytoskeletons, and PM invaginations (7–10). Moreover, ER-PM tethering is crucial for the non-vesicular transport of lipids between the PM and the ER membrane and the regulation of intracellular ion homeostasis (4, 6). Therefore, exocytic PM flows could also regulate the location of these processes by reorganizing ER-PM tethering.

MATERIALS AND METHODS

Yeast strains and plasmids

Yeast strains and plasmids used in this study are listed in tables S1 and S2, respectively. Haploid strains are derived from BY4741 (BY23849), and diploid strains were constructed by mating BY4741-derived and BY4742-derived strains. Deletion and knock-in mutants were constructed by the standard homology recombination method.

Growth condition

Unless otherwise noted, yeast cells were grown at 25°C in synthetic complete (SC) medium [yeast nitrogen base (6.7 g/liter) without amino acids, l-adenine (550 mg/liter), l-arginine (280 mg/liter), l-alanine (280 mg/liter), l-asparagine (280 mg/liter), l-aspartic acids (280 mg/liter), l-cysteine (280 mg/liter), glycine (280 mg/liter), l–glutamic acids (280 mg/liter), l-glutamine (280 mg/liter), l-isoleucine (280 mg/liter), l-lysine (280 mg/liter), l-phenylalanine (280 mg/liter), l-proline (280 mg/liter), l-serine (280 mg/liter), l-threonine (280 mg/liter), l-tyrosine (280 mg/liter), l-valine (280 mg/liter), leucine (530 mg/liter), methionine (86 mg/liter), histidine (86 mg/liter), uracil (22 mg/liter), myo-inositol (100 mg/liter), and p-aminobenzoic acid (3 mg/liter) (pH 5.5)] with 2% glucose. Leucin was removed for the selection of plasmid-harboring yeasts. For the induction of genes under the GAL1 promoter, cells were precultured in SC medium with 2% raffinose, and 1/10 volume of 20% galactose was added to the medium. YPD media (1% yeast extract, 2% bacto peptone, and 2% glucose) were used for cell cycle synchronization with α factor.

Microscopy

For microscopic observation, cells were mounted on concanavalin A (ConA) (Nacalai Tesque)–coated glass-bottom dish and observed with A1R (Nikon). A1R was equipped with a CFI Apochromat TIRF 60×/1.49 oil objective lens (Nikon). mNG and super folder green fluorescent protein were excited by a 488-nm laser, and the fluorescence that passed a 525/50-nm band-pass filter was detected with a GaAsP detector. mSc-I was excited by a 561-nm laser, and the fluorescence that passed a 595/50 nm band-pass filter was detected with a GaAsP detector. For Cry2PHR excitation, cells were irradiated with 488-nm laser. Acquired images were processed with Fiji (63) for two-dimensional (2D) images or NIS Elements (Nikon) for 3D reconstructed images. 3D images were reconstructed from Z-stack images with 0.2- or 0.225-μm steps by α blending and cropped to remove the opposite side of the observation plane.

For observation of the protein exclusion from the bud sites, Z-stack images were taken with 0.2-μm steps every 3 min. Region-of-interest (ROI) circles of 0.6-μm diameter were set at a bud site and five non-bud sites. After the subtraction of the background signal, the fluorescence intensity of each protein at the time point when Exo84 signal at the bud sites first reached the threshold (300 arbitrary units (AU) for Exo84-mNG or 500 AU for Exo84–mSc-I) was divided by the signal intensity 3 min before the time point. For non-bud sites, averages of relative intensities at the 5 ROIs were quantified. The gap diameter was quantified with 2D binarized images. Cells that moved during observation were excluded from the analyses. For shmoo formation, cells mounted on the ConA-coated glass-bottom dish were washed with SC glucose medium twice. Then, α factor (10 μg/ml) was added to induce shmoo formation. For cell peripheral damage assay, 405-nm laser was irradiated to a circle of 0.5-μm diameter in a cell. The laser power and irradiation time were set to 20% and 0.125 s, respectively.

Overexpression of ER-PM tethering proteins

Cells were grown in SC-Leu liquid medium containing 2% raffinose overnight. For spot assay, cell culture at the point of optical density at 600 nm (OD₆₀₀) = 0.1 to 0.3 was diluted to OD₆₀₀ = 0.1, and 3 μl of the fivefold serial dilutions were spotted on SC-Leu plates containing 2% raffinose or 2% raffinose and 2% galactose. The plates were incubated at 25°C for 3 or 7 days.

For microscopic observation and Western blot, cells grown in SC-Leu medium containing 2% raffinose were diluted to OD₆₀₀ = 0.1, and the 1/10 times volume of 20% galactose was added. After 4 to 8 hours of incubation at 25°C, cells were subjected to microscopic observation or Western blot. Cdc10-mNG localization was observed every 0.2-μm step. The number of cells that have cytoplasmic Cdc10-mNG foci was counted. The cells that did not show mSc-I fluorescence were excluded from the analyses.

Western blot

Cell pellets were resuspended in cold lysis buffer (0.25 M NaOH and 1% β-mercaptoethanol) and incubated on ice for 15 min. Trichloroacetic acid (50%) was added to the lysates to achieve a final concentration of 20%. After 15 min of incubation on ice, the lysates were spun down, and the supernatant was discarded. The precipitates were washed with ice-cold acetone and resuspended in SDS–polyacrylamide gel electrophoresis (PAGE) sample buffer [63 mM tris-Cl (pH 6.8), 2% SDS, 1% β-mercaptoethanol, 0.01% Bromophenol Blue, and 10% glycerol]. Western blot was performed with anti-mini AID (MBL, M214-3), anti-mCherry (Abcam, ab125096), or anti–tubulin α (Bio-Rad, MCA78G) antibodies.

Fluorescence recovery after photobleaching

The fluorescence in a square of 0.9 μm by 0.5 μm was photobleached with 488- or 561-nm laser. After photobleaching, images were acquired every 2 s for 91 times. The fluorescence intensity was quantified with NIS Elements (Nikon) and processed with easyFRAP-web (64). The fluorescence recovery curve was normalized by full-scale normalization. Halftime and mobile fractions were calculated by fitting the normalized fluorescence intensity to double exponential curves.

Exocytosis inhibition

For spot assay, cells were precultured in SC glucose medium. At the point of OD₆₀₀ = 0.1 to 0.3, cell culture was diluted to OD₆₀₀ = 0.1, and 3 μl of the fivefold serial dilutions were spotted on SC glucose plates containing 5 μM 5-Ph-IAA (BioAcademia) or 0.005% DMSO. The plates were incubated at 25°C for 2 days.

For Western blot, cells grown in SC glucose medium were treated with 100 nM 5-Ph-IAA, 5 μM 5-Ph-IAA, or 0.005% DMSO for 30, 60, or 90 min. The cells were collected by centrifugation and subjected to SDS-PAGE.

For microscopic observation, cells were pre-cultured in YPD medium. The cells were washed with YPD medium twice and treated with α factor (10 μg/ml) at 25°C for 2 or 3 hours. Then, 100 nM 5-Ph-IAA, 5 μM 5-Ph-IAA, or 0.005% DMSO was added to the cell culture, and the cells were incubated at 25°C for 1 hour. The cells were mounted on ConA-coated glass-bottom dish and washed six times with SC glucose medium containing same concentration of 5-Ph-IAA or DMSO as pretreatment. After the addition of pronase E (50 μg/ml; Sigma-Aldrich), the cells were observed with A1R. The gap diameter was quantified with 2D binarized images. The diameter of largest gaps formed in each cell during 1 hour of observation was presented. The number of Spa2-mNG or Spa2–mSc-I polarization sites was counted with 2D binarized images. Cells that moved during observation were excluded from the analyses.

Statistical analyses

Paired t tests and Welch’s t tests (single comparison) were performed with Excel 2019 (Microsoft). Tukey-Kramer tests (multiple comparisons) were done with Prism 9 (GraphPad software).

Supplementary Materials

This PDF file includes:

Figs. S1 to S8

Tables S1 and S2