Arv1 regulates PM and ER membrane structure and homeostasis but is dispensable for intracellular sterol transport

Abstract

The pan-eukaryotic endoplasmic reticulum (ER) membrane protein Arv1 has been suggested to play a role in intracellular sterol transport. We tested this proposal by comparing sterol traffic in wild-type and Arv1-deficient Saccharomyces cerevisiae. We used fluorescence microscopy to track the retrograde movement of exogenously supplied dehydroergosterol (DHE) from the plasma membrane (PM) to the ER and lipid droplets and high performance liquid chromatography to quantify, in parallel, the transport-coupled formation of DHE esters. Metabolic labeling and subcellular fractionation were used to assay anterograde transport of ergosterol from the ER to the PM. We report that sterol transport between the ER and PM is unaffected by Arv1 deficiency. Instead, our results indicate differences in ER morphology and the organization of the PM lipid bilayer between wild-type and arv1Δ cells suggesting a distinct role for Arv1 in membrane homeostasis. In arv1Δ cells, specific defects affecting single C-terminal transmembrane domain proteins suggest that Arv1 might regulate membrane insertion of tail-anchored proteins involved in membrane homoeostasis.

Sterols are highly enriched in the plasma membrane (PM)¹ of eukaryotic cells (1, 2). For example, ~70% of total cellular ergosterol is found in the PM of the budding yeast Saccharomyces cerevisiae, comprising ~40 mole percent of PM lipids (3). Ergosterol is synthesized in the endoplasmic reticulum (ER) and transported to the PM primarily by an ATP-requiring, non-vesicular mechanism (2–4). Retrograde transport also occurs, because exogenously supplied sterols move from the PM to the ER, where they are esterified by the sterol acyl transferase enzymes Are1 and Are2 for encapsulation in lipid droplets (LDs)(4). As sterols are largely insoluble in water, they must be shielded from the aqueous environment for transport to occur efficiently. In analogy with other lipid transport processes, sterol transfer might require a cytoplasmic lipid transport protein, similar to the ceramide transport protein (CERT) (5) and the steroidogenic acute regulatory protein (StAR)(6, 7). Oxysterol binding protein homologs (termed Osh proteins in yeast) are cytoplasmic, sterol-binding proteins that could function as ergosterol transporters (8). We recently eliminated these proteins from consideration by showing that sterol transport between the ER and PM is unaffected in yeast cells lacking functional Osh proteins (9, 10). Thus the mechanism of intracellular sterol transport remains enigmatic.

The ER membrane protein Arv1 (11, 12)(Fig. 1A) has been proposed to be a key element of the sterol homeostatic machinery in all eukaryotic cells, with an unspecified role in sterol transport between the ER and PM. Arv1 was originally discovered through its essential function in maintaining the viability of S. cerevisiae cells that lack Are1 and Are2. It was subsequently discovered that in the absence of Arv1, yeast cells upregulate the unfolded protein response (13, 14) and display a variety of lipid-related phenotypes. These include altered intracellular sterol distribution (11, 15), sphingolipid metabolism (16), GPI-protein biosynthesis (17) and PI(4,5)P₂ polarization in response to pheromone treatment (18). It has been suggested that defective sterol transport in arv1Δ cells gives rise to these pleiotropic ‘lipid phenotypes’, leading eventually to induction of the unfolded protein response (13). Here we examine the proposal that Arv1 is involved in sterol transport between the ER and PM. Our data indicate that whereas sterols traffic normally in arv1Δ cells, Arv1 deficiency results in dramatic changes in the organization of the PM lipid bilayer and in ER morphology.

Figure 1. Arv1 – membrane topology and subcellular localization.

(A) Experimentally determined membrane topology of Arv1 (12). S. cerevisiae Arv1 is a 321 amino acid protein. The Arv1 homology domain (AHD; amino acids 2–63 of yeast Arv1) is indicated as a thick grey line.

(B) Perinuclear localization of over-expressed GFP-Arv1. Deconvolved widefield fluorescence microscopy of GFP-Arv1 expressed from a high-copy 2µ plasmid (pCB1033) in wild-type cells (BY4741) in which cortical (arrowheads) and perinuclear ER (arrows) localization was detected. (C) Perinuclear fluorescence was largely absent in wild-type cells expressing GFP-Arv1 from a low-copy plasmid (pCB1031). (E) Quantification of GFP-Arv1 perinuclear localization in wild-type cells expressing GFP-Arv1 from “low” (CEN) or “high” (2µ) copy-number plasmids (N = 100 cells). (E) Wt cells co-expressing both GFP-Arv1 from a high-copy plasmid and the ER marker RFP-ER (pCB1024) visualized by confocal fluorescence microscopy. Scale bar, 5 µm.

Results and Discussion

Arv1 is localized to the cortical ER

To determine the sub-cellular localization of Arv1 in budding yeast we expressed GFP-Arv1 (previously shown to be functional (14)) on a high-copy plasmid under the control of its own promoter. Fluorescence microscopy revealed a perinuclear and cortical fluorescence pattern (Fig. 1B) as previously reported (14,15), consistent with localization of the protein at the nuclear and cortical ER, respectively. As Arv1 over-expression might result in its accumulation at sites where it is not normally found (14), we expressed GFP-Arv1 at physiological levels under control of its own promoter on a low-copy number plasmid. Using long exposure times and reiterative image deconvolution, we could detect GFP-Arv1 even when expressed at these low levels. Under these conditions, the perinuclear fluorescence observed with over-expressed GFP-Arv1 was not seen but only cortical fluorescence remained (Fig. 1C). Whereas perinuclear fluorescence was seen in >80% of cells with over-expressed GFP-Arv1, it could be detected in fewer than 5% of cells in which GFP-Arv1 was expressed at low levels (bar chart, Fig. 1D). Over-expressed GFP-Arv1 always co-localized with the nuclear and cortical ER marker RFP-ER (RFP fused to the transmembrane domain of Scs2)(Fig. 1E); we were unable to perform similar co-localization analyses with GFP-Arv1 expressed at low levels on account of low signal intensity. Nevertheless, our data suggest that Arv1 is normally restricted to the cell periphery in a distribution consistent with cortical ER.

Characterization of S. cerevisiae arv1Δ cells

We examined key characteristics of arv1Δ cells as a prelude to studies of intracellular sterol transport in these cells. We confirmed that arv1Δ cells have a temperature-sensitive growth phenotype (Fig. 2A)(11) and established that they grow slowly when incubated under hypoxic conditions in media containing ergosterol and fatty acids (Fig. 2B). It has been suggested (11) that the slow growth of arv1Δ cells under hypoxic conditions may be due to a reduced ability to incorporate sterols from the growth medium into the plasma membrane (PM). However, our experiments reveal that the fluorescent sterol DHE (Fig. 3A) is incorporated into the PM of arv1Δ cells to an extent similar to that seen in wild-type cells (Fig. 3B, left panels), indicating that the hypoxic growth phenotype of arv1Δ cells may require a more complex explanation.

Figure 2. Characterization of arv1Δ cells.

(A) 10-fold serial dilutions of wild-type (wt) and arv1Δ yeast cultures growing in YPD were placed on YPD plates and incubated for two days at 25°C or 37°C.

(B) Serial dilutions of wt and arv1Δ cultures were placed on YPD plates containing 20 mM ergosterol, 0.5% (w/v) Tween 80, 0.5% (v/v) ethanol and incubated either aerobically or in a hypoxic chamber for 2 days at 30°C.

(C) Non-polar lipid extracts from 0.1 OD₆₀₀ equivalents of logarithmically growing wt and arv1Δ cultures were separated by HPLC. Eluted sterols (top) and steryl esters (bottom) were detected by absorbance at 280 nm. erg, ergosterol; A, ergostatrienol (MW=396); B, Ergosta-5,7-dien-3β-ol (22-dihydroergosterol; MW=398). The identities of sterols A and B were confirmed by liquid chromatography-mass spectrometry. The likely molecular species of the steryl esters is indicated in the bottom panel.

(D) Structure of ergosterol and 22,23-dihydroergosterol (sterol ‘B’ in panel C).

(E) Ergosterol and 22,23-dihydroergosterol content of sucrose-gradient separated cellular membranes from wt and arv1Δ cell homogenates. Fractions were collected from the top, and aliquots were either extracted with hexane for ergosterol analysis or washed by centrifugation in 100 mM Tris-HCl pH 7.5 and analyzed by SDS-PAGE and immunoblotting using antibodies directed against organelle-specific marker proteins (ER, Sec61; vacuole, Vph1; PM, Gas1). Immunoblots were quantified by densitometry using the ImageJ program (http://rsb.info.nih.gov/ij/).

Figure 3. DHE transport from PM to ER.

(A) Structure of dehydroergosterol (DHE).

(B) Epifluorescence micrographs of wild-type (wt) and arv1Δ cells grown under hypoxic conditions for 36h at 30°C in yeast extract-peptone-dextrose (YPD) supplemented with 20 µg/mL DHE in 0.5% ethanol, 0.5% Tween-80, and chased under aerobic conditions for 0, 1 or 2h. Scale bar = 20 µm.

(C) Quantification of images such as those recorded in panel B. The images were scored manually by assigning individual cells to one of the three categories (PM, intermediate, puncta) illustrated schematically above the graphs. Data are from three independent experiments; at least 200 cells were counted for each time-point in each experiment.

(D) DHE-loaded cells were processed as in panels B and C, except that at each time-point an aliquot of the cells was taken for lipid extraction and HPLC analysis to resolve and quantify DHE and DHE esters. The graph shows the percentage of DHE recovered in the form of DHE-ester.

Previous reports (16, 17) indicated that Arv1 is required for normal glycerophospholipid, sphingolipid and sterol metabolism. We determined that arv1Δ cells have a higher sterol, steryl ester and phospholipid content than wild-type cells, but that the molar ratio of sterol to phospholipid is unaffected (Table 3). HPLC analyses revealed that whereas the ergosterol content of arv1Δ was similar to that of wild-type cells, arv1Δ cells accumulated the ergosterol biosynthetic intermediate 22,23-dihydroergosterol (Fig. 2C,D and Table 3). This increase may be caused by deficiency of the C-22 sterol desaturase (Erg5) (19) or because the biosynthetic intermediate is mis-localized and unavailable to the enzyme. The high level of 22,23-dihydroergosterol results in ergosterol accounting for only ~60% of total sterol in arv1Δ cells compared with 77% in wild-type cells (Table 3). The decreased fraction of ergosterol relative to total sterol in arv1Δ vs wild-type cells is similar to that reported previously by Swain et al. (16). However, whereas we find that arv1Δ cells accumulate 22,23-dihydroergosterol, Swain et al. (16) reported that the cells accumulate lanosterol and other uncharacterized sterol intermediates. It is possible that 22,23-dihydroergosterol was the main uncharacterized intermediate reported by these authors.

Table 3.

Sterol, steryl ester and phospholipid content of wild-type and arv1Δ cells

	ergosterola	22,23-dihydroergosterola	steryl estera	PLa,b	Sterol/PLb,c
BY4742	1.33 ± 0.09	0.40 ± 0.01	0.056 ± 0.005	18.1 ± 1.3	~0.096
arv1Δ	1.53 ± 0.04	1.00 ± 0.14	0.138 ± 0.013	27.5 ± 5.1	~0.092

^anmol per OD unit, i.e., per 2 × 10⁷ cells

^bPL, phospholipid

^cmolar ratio

Overnight cultures were diluted to OD₆₀₀=0.5 in YPD and grown for 4h at 30°C before harvesting. Non-polar lipids (including sterols and steryl esters) were extracted from glass-bead disrupted yeast cells with hexane. Phospholipids were extracted from an aliquot of the same lysate using methanol:chloroform (1:2, by volume). The organic solvent was evaporated in a speed-vac and an aliquot of the non-polar extract corresponding to 0.1 OD units of cells was analysed by HPLC, while 2.0 OD equivalents of the polar extract was analysed by measuring the phosphate content of the hydrolyzed lipids. Standard curves were generated with pure ergosterol (Sigma) or sodium phosphate. The numbers are from 3 (sterol) or 2 (phospholipid) independent measurements for each strain.

We next examined the intracellular distribution of ergosterol in arv1Δ cells by using a well-vetted sucrose gradient centrifugation procedure (9, 20, 21) to separate the PM from intracellular membranes (Fig. 2E, top panels). As Sec61 is distributed throughout the ER, including the cortical ER region that lies adjacent to the PM, it is clear from Fig. 2E (top panels) that the PM-enriched fractions are not contaminated by ER membranes. Fractionation profiles of arv1Δ cells and wild-type cells were similar (Fig. 2E, top panels). We found that the majority (~70%) of ergosterol is located in the PM in both arv1Δ and wild-type cells, whereas a large fraction (~50%) of 22,23-dihydroergosterol is recovered in intracellular fractions (Fig. 2E, bottom panels).

Previous reports used metabolic radiolabeling (in conjunction with thin layer chromatography) (11) and filipin staining (17) to conclude that sterols accumulate in intracellular membranes of arv1Δ cells. Our data suggest that this is primarily due to the increased level of 22,23-dihydroergosterol.

We conclude that arv1Δ cells display aberrant sterol metabolism exemplified by a ~2.5-fold accumulation of 22,23-dihydroergosterol, but that the distribution of sterols amongst cellular membranes (PM versus internal membranes) in arv1Δ is similar to that found in wild-type cells.

Dehydroergosterol (DHE) is transported from the PM to the ER and lipid droplets in arv1Δ cells

We recently described a fluorescence-based assay to visualize and quantify intracellular sterol transport in yeast (9). The assay exploits the fluorescence of DHE (Fig. 3A), a naturally occurring sterol. DHE is an excellent analog of ergosterol as it supports the growth of hem1Δ cells that require exogenous sterol for growth (9). In a typical experiment, cells are incubated with DHE under hypoxic conditions to enable the fluorescent sterol to enter the PM. Hypoxic incubation overcomes ‘aerobic sterol exclusion’ and is required for sterol loading (9). On chasing the cells under aerobic conditions, DHE is seen to move out of the PM and appear in LDs. DHE redistribution, scored by cell counting and the conversion of DHE to DHE-ester, occurs in an ATP-dependent manner that does not require vesicle-mediated transport (9).

We applied the DHE redistribution assay to assess intracellular sterol transport in arv1Δ cells. As shown in Fig. 3B, both wild-type and arv1Δ cells could be efficiently loaded with DHE and showed PM-localized fluorescence at the start of the chase period. After a 2h chase under aerobic conditions, fluorescence appeared in the form of discrete puncta corresponding to LDs. Redistribution was evident in both wild-type and arv1Δ cells, with the principal difference being in the size and number of puncta per cell. DHE-stained LDs were small and numerous in wild-type cells, whereas arv1Δ cells had fewer and larger LDs, consistent with previous observations (13). Thus DHE is able to redistribute from the PM to intracellular compartments in arv1Δ cells.

We quantified DHE redistribution by scoring cells according to the categories depicted schematically in Fig. 3C: PM staining, intermediate staining and puncta. The data (Fig. 3C) indicate that DHE moves out of the PM at a similar rate in wild-type and arv1Δ cells such that after 2 h there are essentially no cells with PM-localized DHE (Fig. 3C, left panel). Wild-type and arv1Δ cells differed somewhat in the rate at which DHE appeared in puncta (Fig. 3C, right panel). Puncta were discernible in both cell types after a ~1 h lag period, but were more evident in wild-type than arv1Δ cells after 3 h. To provide an independent read-out of transport we extracted lipids from the cells at different time points and determined the percentage of DHE that had been converted to DHE ester. As shown in Fig. 3D, DHE esterification could be detected after a lag period of ~1.5 h and increased monotonically in both cell types over the 3 h period of the experiment. We conclude that sterol transport from the PM to the ER is unaffected by Arv1 deficiency as DHE moves out of the PM (visual scoring) and arrives at the ER (conversion to DHE ester) at a similar rate in both wild-type and arv1Δ cells. We note that DHE internalization is unlikely to involve endocytosis as a parallel redundant route as the deletion of ARV1 results in endocytosis defects (15), consistent with synthetic growth defects observed when combining arv1Δ with the abp1Δ, sla1Δ, or sla2Δ endocytosis mutations in a haploid cell (data not shown).

Ergosterol is transported from the ER to the PM in arv1Δ cells

To examine anterograde sterol transport in arv1Δ cells we pulse-labeled the cells for 4 min at 30°C with [³H-methyl]methionine to incorporate [³H] into newly synthesized sterols and chased for 90 min at the same temperature. Cell samples taken at the end of the pulse labeling period and after the chase were rapidly cooled, treated with energy poisons and homogenized by bead-beating. The resulting homogenate was fractionated by sucrose gradient centrifugation to resolve the PM from intracellular membranes as shown in Fig. 2E.

The specific radioactivity (SR = cpm ÷ absorbance units) of ergosterol was determined for subcellular fractions as well as for whole cells, and the relative specific radioactivity (RSR_frac = SR_frac ÷ SR_cell) of fractions pooled pair-wise from the top of the gradient was calculated. We first analyzed wild-type cells. At the end of the pulse period, intracellular membranes (fractions 1–8) had an RSR_frac > 1 whereas the PM (fractions 9–14) had an RSR_frac < 1 (Fig. 4A). At the end of the chase period all gradient fractions had an RSR_frac ~1 (Fig. 4B). These data indicate that at the end of the labeling pulse [³H]ergosterol is mainly located in the ER-enriched intracellular fractions as expected, but at the conclusion of the chase period it equilibrated completely with the PM and other cellular membranes.

Figure 4. Ergosterol transport from ER to PM.

Wild-type (wt) and arv1Δ cells were pulse-labeled for 4 min and chased for 90 min at 30°C. Samples taken at the end of the pulse-labeling period (chase = 0 min) and at the end of the chase (chase = 90 min) were fractionated on sucrose gradients. The specific radioactivity (SR = cpm ÷ absorbance units) of ergosterol was determined for subcellular fractions as well as for whole cells by HPLC analysis, and the relative specific radioactivity (RSR_frac = SR_frac ÷ SR_cell) of fractions pooled pairwise from the top of the gradient was calculated. (A) RSR of fractions analyzed after the pulse-labeling period; (B) RSR of fractions analyzed after a 90 min chase; (C) Average RSR of the PM (fractions 9–14) after the pulse and chase periods. The fractionation protocol is the same as that characterized in Fig. 2E. Fractions containing ER (Sec61), vacuole (Vph1) and PM (Gas1) markers are indicated.

The spontaneous movement of ergosterol between membranes during the fractionation procedure is not expected to be a factor in our analyses as it is predicted to be very slow (9). Our data support this assertion. If ergosterol moved spontaneously between membranes at an appreciable rate during the fractionation procedure, then RSR_frac would equal 1 for all fractions in the end-of-pulse sample but we note that RSR_frac was ≠ 1 for almost all these samples. Also relevant is our observation that ergosterol and 22,23-dihydroergosterol fractionate differently (Fig. 2E, lower panels), indicating that sterol exchange during cell homogenization and fractionation is not a complicating factor in our analysis.

The RSR_frac profiles for Arv1-deficient and wild-type cells at the end of the pulse period and after a 90 min chase are essentially the same (Fig. 4A,B) indicating that Arv1-deficiency does not affect movement of ergosterol from the ER to the PM. Of note, RSR_frac of the PM was ≥ 0.5 at the end of the pulse period for both strains (Fig. 4C) indicating that the t_1/2 for ergosterol movement between the ER and PM is < 4 min, irrespective of Arv1 function.

Altered PM lipid organization in arv1Δ cells

Cells lacking Arv1 are hyper-sensitive to nystatin (11), a polyene anti-fungal compound that manifests toxicity by binding ergosterol at the PM surface. We confirmed this phenotype in our strain background (Fig. 5C). In addition, whereas only ~1.5 ± 0.4% of PM ergosterol was extracted by methyl-β-cyclodextrin from wild-type cells under our standard conditions, similar to data previously reported (9, 22), the efficiency of extraction from arv1Δ cells was 4.5-fold higher (6.8 ± 3.4%). The increased availability of PM ergosterol to methyl-β-cyclodextrin extraction and the nystatin-sensitivity of arv1Δ cells indicate that Arv1 influences the organization of sterols at the PM.

Figure 5. Drug sensitivity tests.

(A) Zone of inhibition assay comparing the growth of wt (BY4741), drs2Δ (CBY2317), osh4Δ (CBY5150) and arv1Δ (CBY994) cells in the presence of papuamide B. At the concentrations indicated, papuamide B was added to each disc and the diameters of the zones where cell growth was inhibited were measured. Average diameters of inhibition zones were determined from triplicate trials and tabulated for each strain as a ratio relative to wild type. (B) Ten-fold serial dilutions of wt, arv1Δ, drs2Δ and osh4Δ cultures spotted onto SC medium with and without the inhibitory PC analogue, edelfosine (60 µM) and (C) spotted onto SC or YPD medium with or without 10 µM nystatin. Plates were incubated at 30°C for 48 hrs.

To determine whether the change in PM lipid organization was restricted to sterols, we tested whether arv1Δ cells were sensitive to papuamide B, a marine cyclodepsipeptide that is toxic to cells that expose phosphatidylserine (PS) at their surface (23). As PS is normally not exposed at the cell surface, wild-type cells are relatively resistant to papuamide B. However, drs2Δ and osh4Δ strains are papuamide B-sensitive as reported previously (24), and so we used these strains as controls. We found that arv1Δ cells were hypersensitive to papuamide B (Fig. 5A), indicating a loss of PM lipid asymmetry.

PS exposure at the surface of arv1Δ cells could be due to compromised phospholipid flippase activity. Preliminary experiments in which cells were labeled with NBD-PC, a fluorescent derivative of phosphatidylcholine (PC), indicated that this was not the case: NBD-PC was taken up by arv1Δ cells and distributed amongst internal membranes exactly as seen in wild-type cells (not shown). To examine this result more closely we tested the sensitivity of the cells to edelfosine, a toxic analog of PC. Edelfosine toxicity requires its import across the PM by lipid flippases of the P4-type ATPase class (25, 26). Thus, wild-type cells are sensitive to edelfosine but drs2Δ cells that lack a major contributor to flippase activity (27) are predicted to be resistant. As shown in Fig. 5B, Drs2-deficient cells are relatively resistant to the drug but arv1Δ and wild-type cells are similarly sensitive to edelfosine indicating that PS exposure in arv1Δ is not due to compromised PM lipid flippase activity. These results indicate that Arv1 has little impact on PC trafficking across the PM and between the PM and internal organelles. We conclude that the organization of ergosterol and PS within the PM is specifically affected by Arv1 deficiency.

Cortical and cytoplasmic ER defects in arv1Δ cells

Considering that GFP-Arv1 resides in the cortical ER (Fig. 1C) and arv1Δ cells exhibit PM defects (Fig. 5), we examined whether Arv1 deficiency affects the organization of cortical ER at its interface with the PM. In medial optical sections through wild-type cells, the ER-specific marker Sec61-GFP displays ring-like perinuclear fluorescence as well as discontinuous fluorescence along the cell cortex (Fig. 6A). However, the web-like lattice of the cortical ER organization is best observed through optical sections focused at the cell surface. Although perinuclear and cortical Sec61-GFP fluorescence was observed in arv1Δ cells, extraneous cytoplasmic fluorescence corresponding to ER tendrils was also seen and the cortical ER lattice was disorganized (Fig. 6B, C). After shifting arv1Δ cells cultured at 30°C to 37°C, which inhibits cell growth (Fig. 2A), the percentage of arv1Δ cells with ER defects increased (Fig. 6C). In arv1Δ cells expressing RFP-ER (a fusion of RFP with the C-terminal transmembrane domain of tail-anchored Scs2, an ER membrane protein (28)), perinuclear and cortical fluorescence was also detected, though 5-fold longer exposures were required for visualization relative to wild-type cells (Fig. 6D, E). In medial sections, 86% of arv1Δ cells had extraneous cytoplasmic ER compared to only 7% in wild-type cells (at equal exposures) (N = 100) (Fig. 6D). At the cell surface, the regular lattice of cortical ER observed in 96% of wild-type cells was present in only 9% of arv1Δ cells (N = 100) (Fig. 6E). These results suggest that Arv1 affects ER structural organization at the cell cortex.

Figure 6. Cortical ER organization is disrupted in arv1Δ cells.

(A) Medial optical sections of wt (BY4742) and arv1Δ (CBY994) cells expressing the ER-localized marker Sec61-GFP. (B) Sec61-GFP localization shown in optical sections focused at the cell surface. To visualize cortical ER in arv1Δ cells, due to background fluorescence caused by the proliferation of cytoplasmic ER, contrast adjustments between wt and arv1Δ images were not equivalent. (C) Bar graphs quantify extraneous cytoplasmic ER and organization defects in cortical ER (N > 100 cells). (D) At equivalent exposures, RFP-ER (pCB1024) fluorescence cannot be visualized in arv1Δ cells (middle panel) unless relative to wt cells 5-fold longer exposures are performed (right panel). (E) Optical sections focused at the cell surface also show that the RFP-ER-marked cortical ER is disorganized in arv1Δ cells compared to wt. Arrowheads in all panels indicate examples of disorganized and extraneous peripheral/cortical ER observed in arv1Δ cells. Scale bars = 5 µm.

As the tail-anchored RFP-ER was expressed at reduced levels in arv1Δ cells (Fig. 6D, E; Fig. 7B), we tested whether the expression of another intact tail-anchored ER protein was affected. Sbh1 is a tail-anchored integral membrane protein inserted into the ER membrane by the GET complex (28). In 85.5% wild-type cells, RFP-Sbh1 expressed from a low-copy plasmid localizes to the cortical and perinuclear ER. In contrast, this localization pattern is observed in only 8.5% arv1Δ cells, and overall fluorescence is diminished (N = 200 cells; Fig. 7A). Accordingly, a comparison of protein levels by immunoblot analysis indicated reduced expression of both RFP-ER and RFP-Sbh1 in arv1Δ cell extracts relative to wild type extracts, using actin as the loading control (Fig. 7B). Due to its low expression, RFP-Sbh1 was detected in wild-type extracts only when expressed from a high-copy plasmid. However, regardless of plasmid dosage RFP-Sbh1 could not be detected in arv1Δ cell extracts. These results implicate Arv1 in the targeting and expression of tail-anchored single C-terminal transmembrane domain proteins.

Figure 7. Tail-anchored ER integral membrane proteins in arv1Δ cells.

(A) Wt (BY4742) and arv1Δ (CBY994) cells expressing the ER tail-anchored protein RFP-Sbh1 from a low-copy plasmid (pCB1080). Scale bar, 5 µm. (B) In extracts isolated from wt and arv1Δ cells, expression of RFP-ER (pCB1024), and RFP-Sbh1 expressed from a low CEN (pCB1080) and high 2µ (pCB1081) copy-number plasmids, were compared on immunoblots probed with an anti-RFP antibody. To show equivalent loading, equal amounts of sample were also probed on immunoblots with an anti-actin antibody that detects yeast Act1.

Summary

We examined intracellular sterol transport in arv1Δ cells using both a visual fluorescence-based assay to assess retrograde sterol transport from the PM to the ER and a subcellular fractionation approach to monitor anterograde transport of newly synthesized ergosterol from the ER to the PM. The assays revealed that transport between the PM and ER occurs normally in arv1Δ cells, eliminating a direct role for Arv1 in sterol movement between these membrane compartments. We note, however, that exogenously supplied DHE was esterified at the same rate in arv1Δ and wild-type cells but it took longer to appear in LDs in arv1Δ cells. The reason for this is unclear but as the morphology of LDs is affected by Arv1 deficiency (arv1Δ cells have fewer and larger LDs than wild-type cells, e.g. Fig. 3B and reference (13)), it is possible that the absence of Arv1 impacts the recruitment of steryl esters from the ER to LDs. Further work will be needed to investigate this phenomenon.

Our studies revealed that Arv1 provides an unusual functional link between the ER and PM. Arv1-deficient cells show defects in cortical and cytoplasmic ER, and changes in the lipid organization of the PM such that sterol ‘availability’ increases and the transbilayer asymmetry of the PM phospholipid bilayer is altered. It is possible that these effects are mediated by interactions between the cytoplasmically oriented AHD (Arv1 Homology Domain) of Arv1 (12)(Fig. 1A) and one or more PM components. The nystatin sensitivity of arv1Δ cells is partially reversed by expression of the AHD consistent with the possibility that the AHD is a key element in the ability of Arv1 to regulate PM lipid organization in trans. Future work will determine if this possibility can be generalized to account for other observed changes in PM lipid organization in Arv1-deficient cells, as well as changes in ER morphology.

Distended cytoplasmic ER tendrils are observed in some mutants that perturb the retrieval of ER-resident proteins from the Golgi (14). In fact, arv1Δ cells also fail to retrieve the ER Hsp70 Kar2, which is secreted by default out of the cell (14). Defects in the ER retrieval pathway do not affect cell growth unless unfolded protein response (UPR)-mediated transcription is also inhibited. In cells where both ER retrieval and UPR transcription are inhibited, the loss of Kar2 and other ER proteins from the cell cannot be compensated by increased UPR-dependent expression, resulting in cell lethality (29). Likewise combining deletion of ARV1 and the UPR-transcriptional activator HAC1 is lethal in a haploid cell (14). ARV1 exhibits interactions with some other genes that affect Kar2 retention, including all those representing the GET complex (14), which inserts tail-anchored proteins into the ER membrane (28, 30). In arv1Δ cells, the polytopic transmembrane protein Sec61-GFP is unaffected but fluorescence and levels of the C-terminal transmembrane domain proteins RFP-ER and RFP-Sbh1 are markedly reduced. This result suggests that Arv1 is involved in GET-mediated tail-end insertion into the ER. The role of Arv1 in ER structure and/or lipid droplet biogenesis might then simply reflect functional requirements for proper insertion of tail-end membrane proteins.

Materials and Methods

General

Yeast strains and plasmids are listed in Tables 1 and 2, respectively. All other materials were obtained as described previously (9). To generate N-terminal GFP-ARV1 fusions, a stitch PCR approach was used in which 320 bp ARV1 promoter was joined with the sequence encoding eGFP and in turn fused in-frame to the ARV1 coding sequence with 97 bp downstream of the stop codon. The amplified fusion product was inserted into the BamHI site of pRS315 and pRS325 and confirmed by sequencing. Procedures for hypoxic incubation of yeast cells, DHE fluorescence microscopy and quantification of DHE redistribution by analysis of fluorescence images, lipid extraction, HPLC analysis of lipids, ACAT assays and sucrose gradient fractionation to measure transport of ergosterol from the ER to the PM were carried out exactly as described previously (9).

Table 1.

Yeast strains used in this study.

Strain	Genotype	Source
BY4741	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0	(33)
BY4742	MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0	(33)
CBY994	BY4741 arv1Δ::kan-MX4	(33)
CBY2317	BY4741 drs2Δ::kan-MX4	(33)
CBY5150	BY4741 osh4Δ::kan-MX4	(33)
CBY5194	BY4742 lem3Δ::kan-MX4	(33)
CBY5303	BY4742 SEC61-GFP::HIS3	This study
CBY5307	CBY994 SEC61-GFP::HIS3	This study

Table 2.

Plasmids used in this study.

Plasmid	Markers	Source
pCB1024	PPHO5-dsRED-SCS2220–244 CEN URA3	(34)
pCB1031	GFP-ARV1 CEN LEU2	This study
pCB1033	GFP-ARV1 2µ LEU2	This study
pCB1080	mRFP-SBH1 CEN URA3	This study
pCB1081	mRFP-SBH1 2µ URA3	This study
pRS315	CEN LEU2	(35)
pRS325	2µ LEU2	(35)

Drug sensitivity tests

Papuamide B zone of inhibition assay: For each yeast strain tested, 2 ml of a saturated overnight culture was suspended in 3 ml of 1% molten agar and spread over the surface of solid medium plates. After agar hardening, a 6-mm disc of filter paper containing 12 µl of different concentrations of Papuamide B (Lynsey Huxham, University of British Columbia, Canada) in DMSO was placed on the surface. Plates were incubated at 30°C for 36 hr, and drug sensitivities were measured by comparing the diameter surrounding each disc within which growth was inhibited. Values reported represent an average of three independent trials.

Edelfosine and nystatin growth inhibition: Ten-fold serial dilutions of yeast cultures were spotted onto solid SC (Synthetic complete) or YPD (Yeast Peptone Dextrose) media containing 60 µM edelfosine (Cayman Chemicals, Ann Arbor, MI) or 10 µM nystatin (Sigma-Aldrich Chemicals, Oakville, ON) and incubated at 30°C for 48 hrs.

Fluorescence microscopy

Fluorescent cell images were observed on a Leica DMXRA2 microscope equipped with a X100/1.40 Plan-Apo objective using a GFP filter set. Images were captured with an Orca100ER digital camera (Hamamatsu Photonics, Hamamatsu-City, Japan). For GFP-Arv1 localization studies, log phase cells were imaged in filter-sterilized synthetic medium using 2 sec and 250 msec exposures at 55% arc lamp intensity for CEN (pCB1031) and 2µ (pCB1033) GFP-ARV1 transformants, respectively. Deconvolution was performed using the deconvolution module in Volocity 6.0 (Perkin-Elmer, Waltham, MA) based on a theoretical point spread function (widefield, NA 1.40, emission 507 nm) with a confidence limit of 99%. All experiments were conducted on log phase cells with a 30% arc lamp intensity with a 500 msec exposure to visualize RFP-Sbh1, and 200 or 500 msec for RFP-ER as indicated. Images from RFP-Sbh1 expressing cells were deconvolved using a 95% confidence limit and a theroretical point spread function (widefield, 1.40 NA, emission 607 nm). Cells expressing GFP and RFP fusions were visualized using GFP and Texas-Red filters, respectively (Chroma Technology, Bellows Falls, WT). GFP-Arv1/RFP-ER co-localization studies were performed on a Zeiss Axio Observer microscope modified with a Yokagawa CSU-10 spinning disk using a Hamamatsu 9100 EMCCD camera. A 491 nm laser was set for 80% power and 1 s exposures using a 230 analog gain for GFP fluorescence. Identical settings were used for RFP, but with a 561 nm laser. Confocal images were deconvolved using the Volocity deconvolution module based on theoretical point spread functions and a confidence of 95%. All image analyses were performed using Improvision software (Lexington, MA). Exposure times and contrast enhancement were constant for each series of images unless otherwise noted.

Immunoblot analyses

For analysis of protein expression, 10 OD of RFP-ER expressing cells were prepared as described by Ohashi et al. (31). Pellets were resuspended in SDS sample buffer and boiled for 5 min before SDS-PAGE. Protein transfer to nitrocellulose membranes and immunoblot conditions were as previously described (32). To detect RFP-ER, immunblots were incubated with a 1:1000 titre of anti-RFP antibody (Thermo Fisher Scientific Inc., Rockford, IL) followed by a 1:10000 titre of anti-rabbit-HRP secondary antibody (Bio-Rad Laboratories, Hercules, CA). Actin was detected using a 1:1000 titre of anti-actin antibody (Cedarlane, Burlington, ON) followed with a 1:10000 titre of anti-mouse-HRP secondary antibody (Life Technologies Inc., Burlington, ON). Detection of RFP-Sbh1 by immunoblot required protein transfer to a polyvinylidene fluoride (PVDF) membrane, using conditions recommended by the manufacturer (Bio-Rad Laboratories, Hercules, CA). The PVDF immunoblot was probed with the same antibodies and titres as used for RFP-ER detection.