Suppression of Rft1 Expression Does Not Impair the Transbilayer Movement of Man₅GlcNAc₂-P-P-Dolichol in Sealed Microsomes from Yeast

Abstract

To further evaluate the role of Rft1 in the transbilayer movement of Man₅GlcNAc₂-P-P-dolichol (M5-DLO), a series of experiments was conducted with intact cells and sealed microsomal vesicles. First, an unexpectedly large accumulation (37-fold) of M5-DLO was observed in Rft1-depleted cells (YG1137) relative to Glc₃Man₉GlcNAc₂-P-P-Dol in wild type (SS328) cells when glycolipid levels were compared by fluorophore-assisted carbohydrate electrophoresis analysis. When sealed microsomes from wild type cells and cells depleted of Rft1 were incubated with GDP-[³H]mannose or UDP-[³H]GlcNAc in the presence of unlabeled GDP-Man, no difference was observed in the rate of synthesis of [³H]Man₉GlcNAc₂-P-P-dolichol or Man₉[³H]GlcNAc₂-P-P-dolichol, respectively. In addition, no difference was seen in the level of M5-DLO flippase activity in sealed wild type and Rft1-depleted microsomal vesicles when the activity was assessed by the transport of GlcNAc₂-P-P-Dol₁₅, a water-soluble analogue. The entry of the analogue into the lumenal compartment was confirmed by demonstrating that [³H]chitobiosyl units were transferred to endogenous peptide acceptors via the yeast oligosaccharyltransferase when sealed vesicles were incubated with [³H]GlcNAc₂-P-P-Dol₁₅ in the presence of an exogenously supplied acceptor peptide. In addition, several enzymes involved in Dol-P and lipid intermediate biosynthesis were found to be up-regulated in Rft1-depleted cells. All of these results indicate that although Rft1 may play a critical role in vivo, depletion of this protein does not impair the transbilayer movement of M5-DLO in sealed microsomal fractions prepared from disrupted cells.

The lipid-linked oligosaccharyl donor, Glc₃Man₉GlcNAc₂-P-P-dolichol (mature DLO²), in protein N-glycosylation is formed in two stages in the endoplasmic reticulum (ER) (1–4). In the first stage the lipid intermediates Man-P-dolichol (Man-P-Dol), Glc-P-dolichol (Glc-P-Dol), and Man₅GlcNAc₂-P-P-dolichol (M5-DLO) are formed on the cytoplasmic leaflet of the ER with GDP-Man, UDP-Glc, and UDP-GlcNAc, serving as the glycosyl donors. The biosynthesis of the mature DLO is completed with the addition of four more mannosyl units and the formation of the triglucosyl cap in the second stage after the transbilayer movement of Man-P-Dol, Glc-P-Dol, and M5-DLO to the lumenal monolayer. Although many details about the genetics, enzymology, and regulation of these 14 glycosylation reactions are known, there is virtually nothing known about the ER proteins that are presumably required to allow the lipid-bound hydrophilic glycosyl groups to traverse the hydrophobic core of the ER bilayer.

The PER5/RFT1 gene was originally identified by Walter and coworkers (5) as a gene that was up-regulated by the unfolded protein response and required for efficient protein N-glycosylation in yeast. In a related study (6), the rft1 mutation was shown to be inscrutably suppressed by p53, a soluble protein that has not been found in yeast.

Helenius et al. (7) have reported evidence from metabolic labeling experiments indicating that the RFT1 gene in Saccharomyces cerevisiae encodes a protein that is involved in the flipping of M5-DLO in vivo. More recently, a point mutation in the human orthologue of the RFT1 gene has been shown to result in the accumulation of M5-DLO in fibroblasts from a patient containing an R67C amino acid substitution (8). Although these results implicate Rft1 in the transverse diffusion of M5-DLO, the topological orientation of the accumulated intermediate in the mutant cells and the precise function of the protein in the transbilayer movement of the glycolipid intermediate remain to be defined.

Two reports (9, 10) have demonstrated that Rft1 is not required for the “flipping” of M5-DLO in a reconstituted proteoliposomal system, raising questions about the precise relationship between Rft1 and the M5-DLO flippase. A more recent corroborative study further characterizing the reconstituted flippase activity indicates that the in vitro assay exhibits an impressive specificity for M5-DLO (11).

The current study was conducted to further explore the possible role of Rft1 in the transbilayer movement of M5-DLO in the ER. Our results establish the accumulation of chemical amounts of M5-DLO in the Rft1-depleted cells by FACE analysis, supporting the results obtained by metabolic labeling in the yeast (7) and human (8) mutant cells. However, a series of experiments conducted with sealed microsomal vesicles indicate that, although Rft1 may be required to overcome a biophysical constraint for the flipping of M5-DLO in vivo, its depletion does not hinder the flipping of M5-DLO in sealed microsomal preparations in vitro. The resemblance of these results to the loss of the requirement for the Lec35 gene (12) in the transverse diffusion and/or utilization of Man-P-Dol and Glc-P-Dol for lipid intermediate biosynthesis during disruption of intact cells is discussed.

EXPERIMENTAL PROCEDURES

Materials

GDP-Man was obtained from Sigma-Aldrich. UDP-[1-³H]GlcNAc (20 Ci/mmol), UDP-[1-³H]Glc (20 Ci/mmol), [2-³H]mannose (15 Ci/mmol), [¹⁴C]isopentenyl pyrophosphate (50 mCi/mmol), and [γ-³²P]ATP (100 mCi/mmol) were obtained from American Radiolabeled Chemicals (St. Louis, MO). GDP-[2-³H]Mannose was synthesized enzymatically from [2-³H]mannose (13). [γ-³²P]CTP was synthesized from [γ-³²P]ATP by phosphorylation of CDP using nucleoside diphosphate kinase (Sigma) and purified by ion-exchange chromatography on benzyl-DEAE cellulose (Sigma) as described (14). Dol₁₅, a short-chain analogue of dolichol containing a reduced α-isoprene and one internal cis-(Z)-isoprene, was synthesized chemically as described previously (15). Dol₉₅-P was a generous gift from Dr. M. Mizuno (Kuraray Chemical Co., Okayama, Japan). Rabbit, anti-carboxypeptidase Y (CPY) antibody was a generous gift from Dr. Neta Dean (SUNY, Stonybrook, NY) or was purchased from Molecular Probes (Eugene, OR). All other chemicals and reagents were obtained from standard commercial sources.

Yeast Culture

Yeast strains SS328 and YG1137 (MAT α ade2-201 ura3-52 his3Δ200 lys2-801 Pgal-rft1::KanMX), a haploid strain with RFT1 under control of the Gal1-10 promoter (7), were generously provided by Dr. J. Helenius and Dr. M. Aebi. The strains were cultured at 30 °C in 1% yeast extract (BD Biosciences), 2% Bacto-Peptone (BD Biosciences), and 2% glucose. Strain YG1137 was maintained in the presence of 2% galactose, and Rft1-depleted cells were prepared by switching the carbon source to 2% glucose for 20–30 generations. During this time the growth rate was monitored by measuring the absorbance at 600 nm (A₆₀₀). As the cells became depleted of Rft1, the growth rate decreased from a doubling time of ∼2 h to a doubling time of 6–10 h. Depletion of Rft1 was confirmed by the absence of the fully glycosylated CPY glycoform on SDS-PAGE by Western blot analysis (16).

Preparation of Sealed Yeast Microsomal Vesicles

Intact yeast microsomal vesicles (>94%) were prepared from logarithmically growing cultures by homogenization (tight-fitting Dounce homogenizer) after lyticase (Sigma-Aldrich) treatment, as described by Fernandez et al. (17).

Determination of Yeast Microsomal Integrity by Glucosidase I/II Latency

The integrity of yeast microsomal vesicles was determined by measuring glucosidase I/II activity using [³H]Glc_1–3Man₉GlcNAc₂ oligosaccharide as substrate before and after complete disruption of the vesicles with Triton X-100 (2 mg/ml) as described by Rush and Waechter (18).

Preparation of GlcNAc₂-P-P-Dol₁₅

Synthetic reactions for the preparation of GlcNAc₂-P-P-Dol₁₅ contained 25 mm Tris-HCl, pH 8.5, 5 mm MgCl₂, 0.5 mm sodium orthovanadate, 20 milliunits of snake venom 5′-nucleotidase (Sigma-Aldrich), 1 mm 5′-AMP, 200 μm Dol₁₅-P, a microsomal fraction from CHO TN10 cells (19) (1 mg of membrane protein prepared from CHO cells overexpressing UDP-GlcNAc:Dol-P GlcNAc phosphotransferase), and 0.4 mm UDP-[³H]GlcNAc (350 dpm/pmol) in a total volume of 0.4 ml. After incubation at room temperature overnight, the reaction was stopped by the addition of methanol and chloroform to give a final composition of CHCl₃/CH₃OH/H₂O (65:35:6). Insoluble material was removed by centrifugation, and the supernatant was poured over a Sep-Pak Vac RC silica column (Waters Corp., Milford, MA), equilibrated in CHCl₃/CH₃OH/H₂O (65:35:6), and washed with 50 ml of solvent. The column eluate was dried by rotary evaporation under vacuum at less that 30 °C and deacylated by incubation in 4 ml of CH₃OH containing 0.1 m KOH on ice for 30 min. The reaction was neutralized with acetic acid, dried by rotary evaporation, and dissolved in phosphate-buffered saline.

The sample was then desalted by adsorption chromatography over a column of C18 reverse phase Sep-Pak Vac RC (Waters Corp.). The reverse phase column was eluted with two column volumes of phosphate-buffered saline followed by four column volumes of H₂O. GlcNAc_1–2-P-P-Dol₁₅ was eluted with 5 ml of CH₃OH. Methanol was evaporated under a stream of nitrogen, and GlcNAc_1–2-P-P-Dol₁₅ was dissolved in H₂O and purified further by ion-exchange chromatography on a Whatman DE51 cellulose column (5 ml) equilibrated in H₂O. After elution with 4 column volumes of H₂O, GlcNAc_1–2-P-P-Dol₁₅ was eluted with a linear gradient (0–0.5 m) of NH₄HCO₃. Two ml fractions were collected and monitored for radioactivity. The fractions containing GlcNAc_1–2-P-P-Dol₁₅ were combined, and the analogue was recovered by solid-phase extraction using a C18 RP VAC Sep-Pak column, as described above.

GlcNAc₁-P-P-Dol₁₅ was separated from GlcNAc₂-P-P-Dol₁₅ by lectin affinity chromatography on a WGA-agarose (Vector Laboratories, Burlingame, CA) column (7 ml) eluted with phosphate-buffered saline. The fractions (1 ml) containing the individual glycolipids were combined, recovered by solid phase extraction chromatography as described above, dried by rotary evaporation, and stored at −20 °C until use. The glycan composition of the enzymatic products was verified after mild acid hydrolysis by co-chromatography with authentic GlcNAc and di-N-acetylchitobiose by descending paper chromatography on Whatman 1 paper by developing with butanol/pyridine/water (6:4:3).

Transport of Water-soluble Analogs into Sealed Yeast ER Vesicles

Transport of GlcNAc₂-P-P-Dol₁₅ into sealed yeast ER vesicles was determined by a filtration assay using Millipore HA filter disks (Baxter Scientific Products, Obetz, OH) as described elsewhere (20, 21).

Synthesis of Man₉[³H]GlcNAc₂-P-P-Dol in Sealed Yeast Microsomal Vesicles

Reaction mixtures for the preparation of [³H]GlcNAc-labeled DLOs contained yeast microsomes (0.3 mg of membrane protein) from either wild type or Rft1-depleted cells, 50 mm Tris-HCl, pH 8, 0.25 m sucrose, 5 mm sodium orthovanadate, 5 mm AMP, 5 mm MgCl₂, and 12.5 μm UDP-[³H]GlcNAc (2,000 cpm/pmol). After 10 min at 30 °C, unlabeled GDP-Man (1 mm) was added for the indicated periods of time (1–10 min). After the second incubation period, the reactions were stopped by the addition of 20 volumes of CHCl₃/CH₃OH (2:1), and Man_5–9[³H]GlcNAc₂-P-P-Dol was recovered by the multiple-extraction method described previously (22).

Transfer of [³H]Mannose from GDP-[³H]Man into [³H]Man-P-Dol

Reaction mixtures for Man-P-Dol synthase contained 50 mm Tris-HCl, pH 8, 5 mm AMP, 2 mm sodium orthovanadate, 10 mm MgCl₂, yeast microsomal fraction (0.25 mg of membrane protein) and 10 μm GDP-[³H]Man (1000 dpm/pmol) in a total volume of 0.05 ml. Reactions that were supplemented with exogenous Dol-P contained, in addition, 0.1% CHAPS and 0.1 mm Dol-P (prepared as an ultrasonic dispersion in 1% CHAPS). Samples were incubated with GDP-[³H]Man for 1 min after a 5-min preincubation at 30 °C, and the reaction was terminated by the addition of 20 volumes of CHCl₃/CH₃OH (2:1). The amount of [³H]Man-P-Dol formed was assayed as described previously (22).

Transfer of [³H]Glucose from UDP-[³H]Glc into [³H]Glc- P-Dol

Reaction mixtures for Glc-P-Dol synthase contained 50 mm Tris-HCl, pH 7.4, 5 mm MgCl₂, a yeast microsomal fraction (0.25 mg of membrane protein), and 10 μm UDP-[³H]Glc (2500–5000 dpm/pmol) in a total volume of 0.05 ml. Reactions that were prepared with exogenous Dol-P contained 0.1% CHAPS and 0.1 mm Dol-P (prepared as an ultrasonic dispersion in 1% CHAPS). Samples were incubated with UDP-[³H]Glc for 1 min after a 5-min preincubation at 30 °C, and the reaction was terminated by the addition of 20 volumes of CHCl₃/CH₃OH (2:1). Because yeast microsomes synthesize a glucosylsterol (23–25) in addition to Glc-P-Dol, the two labeled glucolipids recovered in the lipid extract were resolved by thin layer chromatography on silica gel G by developing with CHCl₃/CH₃OH/H₂O (65:25:4). The plates were dried, and the individual [³H]glucolipids were detected using a Bioscan radiochromatogram scanner (Bioscan AR-2000, Bioscan, Washington, D. C.) Because of the extreme lability of yeast microsomal Glc-P-Dol synthase, yeast microsomes were analyzed for Glc-P-Dol synthase immediately after preparation. All data are the result of duplicate analyses and are representative of at least four individual experiments.

Transfer of N-Acetyl-[³H]glucosamine from UDP-[³H]GlcNAc into [³H]GlcNAc-P-P-Dol

Reaction mixtures for GlcNAc phosphate transferase contained 50 mm Tris-HCl, pH 8, 5 mm AMP, 10 mm 2-mercaptoethanol, 10 mm MgCl₂, yeast microsomal fraction (0.25 mg of membrane protein), and 10 μm UDP-[³H]GlcNAc (2500–5000 dpm/pmol) in a total volume of 0.05 ml. Where indicated, exogenous Dol-P was dispersed in 1% Nonidet P-40 and added to a final concentration of 0.2% Nonidet P-40. Samples were incubated with radiolabeled substrate for 5–10 min after a 5-min preincubation, all at 30 °C and stopped by the addition of 20 volumes of CHCl₃/CH₃OH (2:1). Reactions were partitioned as described previously (22) and dried under a stream of nitrogen, and [³H]GlcNAc-labeled phosphatidylinositol was deacylated by treating with toluene/MeOH (1:3) containing 0.1 n KOH at 0 °C for 60 min. After deacylation, the reactions were partitioned, and the lower phase was dried under a stream of nitrogen. The amount of labeled glycolipid formed was determined by liquid scintillation spectrometry.

Assay for Dolichol Kinase Activity

Assay mixtures for dolichol kinase contained 50 mm Tris-HCl, pH 8, 30 mm CaCl₂, 10 mm 2-mercaptoethanol, 20 mm UTP, yeast microsomes (0.25 mg of membrane protein), and 40 μm CTP (100–1000 dpm/pmol) in a total volume of 0.05 ml. Reactions that were prepared with exogenous dolichol contained 0.1 mm dolichol (dispersed in 1% Triton X-100, final concentration = 0.1%). After a 1-min incubation, the reaction was terminated by the addition of 20 volumes of CHCl₃/CH₃OH (2:1). Lipid extracts were partitioned with 1/5 volume of 0.9% NaCl, and the lower (organic) phase was dried under a stream of nitrogen. [³²P]Phosphatidic acid was deacylated in toluene/methanol (1:3) containing 0.1 n KOH at 0 °C for 60 min. The reactions were partitioned and dried under a stream of nitrogen, and the amount of Dol-[³²P]P formed was determined by liquid scintillation spectrometry.

Assay for Yeast cis-Isoprenyltransferase Activity

Reaction mixtures for the analysis of cis-isoprenyltransferase contained 25 mm HEPES-NaOH, pH 8.5, 5 mm MgCl₂, 0.5 mm sodium orthovanadate, 0.1 mm farnesyl diphosphate, yeast microsomal fraction (0.1–0.5 mg of membrane protein), and 20 μm [¹⁴C]isopentenyl pyrophosphate (121 dpm/pmol) in a total volume of 0.05 ml. Samples were incubated with radiolabeled substrate for 15 min at 30 °C and stopped by the addition of 20 volumes of CHCl₃/CH₃OH (2:1). Reactions were partitioned as described previously (22), dried under a stream of nitrogen, and analyzed for radioactivity by liquid scintillation spectrometry as described above.

Determination of Yeast Oligosaccharyltransferase (OTase) Activity Using [³H]GlcNAc₂-P-P-Dol₁₅ as Glycosyl Donor

Reaction mixtures for the determination of yeast OST activity contained 50 mm Tris-HCl, pH 7.5, 5 mm MnCl₂, 0.25 m sucrose, 3.5 μm [³H]GlcNAc₂-P-P-Dol₁₅ (425 dpm/pmol), yeast microsomal fraction (0.1 mg of membrane protein), and either 100 μm acceptor peptide (acetyl-Asn-Tyr-Thr-NH₂) or a control peptide (acetyl-Gln-Tyr-Thr-NH₂) in a total volume of 0.02 ml. After incubation at 37 °C for 0–60 min, the reactions were stopped by the addition of 0.5 ml methanol, and the mixtures were applied to an AG 1X8 column (0.35 ml) and eluted with 1.5 ml of methanol. The column eluate was collected in vials, dried under a stream of nitrogen, and analyzed for radioactivity by liquid scintillation spectrometry. OTase activity was dependent upon the addition of acceptor peptide and linear with respect to time for at least 60 min.

Analysis of Oligosaccharides Derived from Radiolabeled DLOs

To characterize the DLOs synthesized in vitro after incubation of yeast microsomal fractions with GDP-[³H]Man or UDP-[³H]GlcNAc, the oligosaccharides were released by mild acid hydrolysis by incubation in CHCl₃/CH₃OH/H₂O (10:10:3) containing 0.1 n HCl at 50 °C for 60 min. The reactions were dried several times out of water under a stream of nitrogen, dissolved in a small portion of 10% glycerol, and chromatographed on a 1.5 × 40-cm column of Bio-Gel P-6 column (200–400 mesh) equilibrated in 0.2 n acetic acid. Fractions (2.5 ml) were collected and analyzed for radioactivity by scintillation spectrometry.

Analytical Methods

Protein concentrations were determined using the BCA protein assay (Pierce) after precipitation of membrane proteins with deoxycholate and trichloroacetic acid according to the Pierce Biotechnology bulletin “Eliminate Interfering Substances from Samples for BCA Protein Assay.” Samples were analyzed for radioactivity by scintillation spectrometry using a Packard-Tri Carb 2100TR scintillation spectrometer and Econosafe Economical Biodegradable Counting Mixture (Research Products International Corp., Mount Prospect, IL).

RESULTS

Compositional Analyses of M5-DLO Levels in Wild Type (SS328) and Rft1-depleted Cells (YG1137) by FACE

Metabolic labeling studies have indicated that [³H]M5-DLO accumulates in Rft1-depleted yeast cells relative to wild type strains (7). Because assessing DLO and N-linked glycoprotein levels by metabolic labeling with [³H]mannose can occasionally be misleading (26), the chemical amounts of M5-DLO in Rft1-depleted cells were compared with wild type cells by FACE analysis (27). The depletion of Rft1 was verified by showing that repression of the RFT1 gene resulted in hypoglycosylation of carboxypeptidase Y (Fig. 1, panel A).

FIGURE 1.

Analysis of wild type (SS328) and Rft1-depleted (YG1137) yeast cells for carboxypeptidase Y N-glycosylation (panel A) and DLO synthesis (panel B). Cellular extracts were prepared from wild type (lane 1) and Rft1-depleted (lane 2) yeast cells, and proteins were resolved by SDS-PAGE on an 8% polyacrylamide gel and analyzed by Western blot with anti-CPY antibody (16) (panel A). The positions of mature CPY (mCPY) and hypoglycosylated glycoforms lacking one to four N-linked oligosaccharide chains are indicated (−1 to −4). FACE analysis of DLOs from wild type and Rft1-depleted cells is shown. DLOs were extracted from wild type (panel B, lane 1) and Rft1-depleted (panel B, lane 2) yeast cells as described under “Experimental Procedures.” Oligosaccharides from normalized amounts of cell material were released by mild acid hydrolysis, reductively aminated by reaction with 7-amino-1,3-naphthalenedisulfonic acid, and analyzed by FACE (27). The positions of glucose oligomers and DLO standards are shown. The chemical amount of M5-DLO in Rft1-depleted cells was 37 ± 13-fold (n = 3) greater than that of Glc₃Man₉GlcNAc₂-P-P-Dol in wild type cells. The experimental variability of this measurement is likely because of the extent of Rft1 depletion being highly dependent upon time in glucose culture. M5-DLO was detectable on FACE gels of wild type DLOs but not readily visible on the image submitted for publication and was ∼3-fold less than Glc₃Man₉GlcNAc₂-P-P-Dol.

As seen in Fig. 1 (panel B), a band corresponding to mature-DLO is the major species found in control cells (lane 1). However, in the Rft1-depleted cells (panel B, lane 2), there is a more intense band, ∼37 times (n = 3) the amount of mature-DLO seen in control cells, corresponding to the mobility of M5-DLO. This result provides the first direct analysis documenting that a depletion of Rft1 produces a substantial accumulation of chemical quantities of M5-DLO in intact cells.

Comparison of Enzymatic Labeling of M5/M9-DLO in Vitro with either GDP-[³H]Man or UDP-[³H]GlcNAc in Sealed Microsomal Vesicles from Rft1-depleted and Wild Type Cells

Because the FACE compositional analyses established that large amounts of M5-DLO accumulated in vivo, a series of in vitro enzymatic labeling experiments was performed to see if an accumulation of M5-DLO could be demonstrated when sealed microsomal vesicles were incubated with GDP-[³H]Man and UDP-[³H]GlcNAc as radiolabeled substrates.

When sealed vesicles from wild type and Rft1-depleted cells were incubated with GDP-[³H]Man, the amount of [³H]Man incorporated into Man-P-Dol and DLOs was higher in vesicles from Rft1-depleted cells relative to the wild type control (Fig. 2). Surprisingly, when the enzymatically labeled DLOs were released by mild acid hydrolysis and analyzed by gel filtration, labeled M9-DLO was the major DLO seen for both wild type and Rft1-depleted vesicles (Fig. 3). These results (Figs. 2 and 3) suggest that M5-DLO was able to diffuse transversely at the same rate in the sealed vesicles in the presence of normal or depleted levels of Rft1.

FIGURE 2.

Time course for the transfer of [³H]mannose from GDP-[³H]Man into [³H]Man-P-Dol and [³H]Man-DLOs in wild type and Rft1-depleted yeast microsomes. Reaction mixtures contained 50 mm Tris-Cl, pH 8, 0.25 m sucrose, 2 mm CaCl₂, 5 mm MgCl₂, 5 mm AMP, 10 μm [³H]GDP-Man (977 cpm/pmol), and microsomal fractions from either wild type (○) or Rft1-depleted (●) cells (0.2 mg of membrane protein) in a total volume of 0.1 ml. After incubation for the indicated times at 37 °C, incorporation of [³H]mannose into [³H]Man-P-Dol (panel A) and [³H]Man-DLO (panel B) was determined by the multiple extraction method described under “Experimental Procedures.” The data are the average values of duplicate analyses and are representative of several independent comparisons.

FIGURE 3.

Gel filtration on Bio Gel P-6 of [³H-Man]oligosaccharides released from DLOs. Enzymatically labeled DLOs from microsomes prepared from wild type (panel A) and Rft1-depleted cells (panel B) were hydrolyzed in CHCl₃/CH₃OH/H₂O (10:10:3) containing 0.1 m HCl, 50 °C, 30 min, neutralized, dried, and analyzed by gel filtration on a Bio-Gel P-6 column (1.5 × 30 cm) equilibrated in 0.2 n acetic acid and collecting 3.2-ml fractions. This analysis is representative of at least three separate experiments. The elution positions of standard oligosaccharides are indicated by the arrows.

Because a significant amount of labeled M9-DLO could plausibly have been formed from small amounts of endogenous M5/8-DLO present in the lumenal leaflet of the Rft1-depleted vesicles, a similar assay was done by enzymatically pre-labeling endogenous [³H]GlcNAc_1–2-P-P-Dol with UDP-[³H]GlcNAc in the absence of unlabeled GDP-Man. The conversion of pre-labeled endogenous [³H]GlcNAc_1–2-P-P-Dol to [³H]GlcNAc-labeled M9-DLO was then assayed after the addition of unlabeled GDP-Man. Under these incubation conditions any radiolabeled M9-DLO formed would have been synthesized de novo from endogenous, pre-labeled [³H]GlcNAc_1/2-P-P-Dol on the cytoplasmic leaflet, requiring the transbilayer movement of newly synthesized [³H]GlcNAc-labeled M5-DLO from the cytoplasmic monolayer to the lumenal leaflet.

As seen in Fig. 4, [³H]GlcNAc-labeled DLOs are not formed in sealed microsomes from either wild type or Rft1-depleted cells during the initial 10-min labeling period. However, after the addition of nonradioactive mannose, a time-dependent formation of [³H]GlcNAc-DLO was observed. Gel filtration analyses revealed that M8/9-DLO were the major chain lengths formed by microsomes derived from both wild type (Fig. 5, left-hand panels) and Rft1-depleted cells (Fig. 5, right-hand panels) at all times after the addition of nonradioactive GDP-Man. The time course of [³H]GlcNAc-DLO formation (Fig. 4) shows that DLO is elongated as robustly in Rft1-depleted microsomes, relative to wild type, under apparent initial rate conditions, arguing strongly against any factor specifically constraining DLO elongation in the Rft1-depleted reactions. Furthermore, in this analysis the amount of [³H]GlcNAc-labeled M8/9-DLO was actually higher in the Rft1-depleted microsomes. If Rft1 were required for the transbilayer movement of M5-DLO in vitro, as reported in vivo (7), an accumulation of [³H]GlcNAc-labeled M5-DLO would have been expected.

FIGURE 4.

Rate of conversion of endogenous, pre-labeled [³H]GlcNAc_1–2-P-P-Dol to [³H]GlcNAc-labeled DLOs in the presence of unlabeled GDP-Man. Yeast microsomes from either wild type (○) or Rft1-depleted cells (●) were preincubated with UDP-[³H]GlcNAc as described under “Experimental Procedures.” After 10 min at 30 °C, the reactions were incubated for the indicated periods of time in the presence of unlabeled GDP-Man (1 mm). After the second incubation [³H]GlcNAc-labeled DLO was recovered by the multiple-extraction method described previously (22). The data are the average values from four separate experiments.

FIGURE 5.

Gel filtration on Bio Gel P-6 of [³H-GlcNAc]oligosaccharides released from DLOs. The DLO fractions from Fig. 4 were hydrolyzed by incubation in CHCl₃/CH₃OH/H₂O (10:10:3), 0.1 m HCl, 50 °C, 30 min, neutralized, dried, and analyzed by gel filtration chromatography on a Bio-Gel P-6 column (1.5 × 30 cm) equilibrated in 0.2 n acetic acid and collecting 2.5-ml fractions. The elution positions of standard oligosaccharides are indicated by the arrows.

Importantly, at no time point (Fig. 5) was an accumulation of M5-DLO in Rft1-depleted microsomes observed. Based on the time course for M5-DLO extension in Rft1-depleted microsomes, we conclude that the transbilayer movement of the lipid intermediate was mediated by an Rft1-independent process rather than by any traces of Rft1 remaining after suppression of expression.

Comparison of M5-DLO Flippase Activity in Sealed Microsomes from Wild Type and Rft1-depleted Cells by Assaying the Transport of the Water-soluble Analogue, GlcNAc₂-P-P-Dol₁₅

The relative levels of M5-DLO flippase activity were also assessed by comparing the rates of transport of the water-soluble analogue, GlcNAc₂-P-P-Dol₁₅. Structurally similar analogues have been used previously to assay the transbilayer movement of Glc-P-Dol and Man-P-Dol (18, 20, 21) and an undecaprenol-P-P-linked trisaccharide (28). Moreover, several studies have indicated that the natural intermediate, GlcNAc₂-P-P-dolichol, can be translocated to the lumen of microsomal vesicles from yeast (29, 30), hen oviduct (31, 32), and calf brain (33) and utilized as a substrate by OTase. Similar results have been obtained with alg1 mutant cells (34).

This comparison revealed that the water-soluble analogue was transported in a time-dependent process and that the rate and extent of transport was actually higher in the sealed vesicles from the Rft1-depleted cells (Fig. 6). To confirm that GlcNAc₂-P-P-Dol₁₅ entered the lumenal compartment and was not simply adsorbed to the exterior face of the vesicle, its ability to serve as a substrate for the OTase-mediated glycosylation of a membrane-permeant acceptor peptide was tested. As seen in Table 1, substantial and very similar amounts of radiolabeled glycopeptide were formed in sealed vesicles from wild type and Rft1-depleted cells.

FIGURE 6.

Time course of transport of [³H]GlcNAc₂-P-P-Dol₁₅ by sealed yeast ER vesicles. Reaction mixtures for the transport of GlcNAc₂-P-P-Dol₁₅ contained 10 mm HEPES-NaOH, pH 7.4, 0.25 m sucrose, 10 mm MgCl₂, 5 mm EDTA, 1.7 μm GlcNAc₂-P-P-Dol₁₅ (704 dpm/pmol), and 0.125 mg of microsomal protein in a total volume of 0.02 ml. After incubation for the indicated periods of time, transport of the water-soluble analogue was determined by the filtration assay described under “Experimental Procedures.” The data are average values from three separate transport experiments.

TABLE 1.

OTase activity in sealed yeast microsomal vesicles using [³H]GlcNAc₂-P-P-Dol₁₅ as glycosyl donor

Vesicular integrity was determined by measuring glucosidase I and II latency (18), and OTase activity was determined after incubation for 60 min at 37 °C as described under “Experimental Procedures.” (1) and (2) indicate two separate experiments.

Yeast strain	Vesicular integrity	OTase + control peptide	OTase + acceptor peptide
	%	pmol/mg	pmol/mg
Rft1-depleted (1)	97	0.08	35.5
Rft1-depleted (2)	100	0.08	26.7
Wild type (SS328)	100	1.03	35.3

Comparison of Man-P-Dol Synthase, Glc-P-Dol Synthase, GlcNAc Phosphate Transferase, Dolichol Kinase, and cis-Isoprenyltransferase Activities in Yeast Microsomal Fractions

Because the results of the experiment in Fig. 2 indicated that the rate of Man-P-Dol and DLO synthesis was higher in the mutant cells, the description of the Rft1-depleted cells was extended by a comparison of Man-P-Dol synthase, Glc-P-Dol synthase, GlcNAc phosphate transferase, dolichol kinase, and cis-isoprenyltransferase activities in microsomal fractions from wild type (SS328) and Rft1-depleted yeast cells (YG1137), both grown in the presence of 2% glucose. The results in Table 2 show that all of these enzyme activities are, indeed, higher in the Rft1-depleted cells relative to wild type control cells, possibly as a compensatory response to the loss of Rft1 function.

TABLE 2.

Several enzyme activities involved in Dol-P and lipid intermediate biosynthesis were up-regulated in microsomes from Rft1-depleted cells relative to wild type control cells

Wild type (SS328) and Rft1-depleted cells (YG1137) were grown in the presence of 2% glucose, microsomal fractions were prepared, and the indicated enzyme activities were assayed as described under “Experimental Procedures.” Man-P-Dol synthase (MPDS), Glc-P-Dol synthase (GPDS), and GlcNAc phosphate transferase (GPT) were assayed in the presence and absence of exogenous Dol-P, whereas dolichol kinase assays were conducted in the presence and absence of exogenous dolichol. The rates are the average values obtained from at least four separate microsomal preparations and are representative of numerous comparisons. cis-IPTase, cis-isoprenylransferase. Numbers in parentheses represent the -fold increase in the rates in the presence or absence of exogenous Dol₉₅-P or dolichol.

Enzyme	Wild type (SS328)		Rft1-depleted (YG1137)
	No addition	Plus Dol-(P)	No addition	Plus Dol-(P)
	pmol/mg/min		pmol/mg/min
MPDS	17.4	127.5	74.8 (4.3)	324.7 (2.5)
GPDS	3.6	5	7.8 (2.2)	12.3 (2.5)
GPT	0.65	22	3.6 (5.5)	39.3 (1.8)
Dolichol kinase	6.7	9	9.58 (1.4)	11 (1.2)
cis-IPTase	93.3		200 (2.1)

DISCUSSION

It has been recognized for many years that flippases play crucial roles in membrane biology in membrane bilayer biogenesis, glycerophospholipid asymmetry, and the translocation of polar glycolipid intermediates in glycoprotein and glycosylphosphatidylinositol anchor biosynthesis (35–38). Of the many mechanistic and regulatory aspects of the dolichol pathway, progress has been extremely slow in devising assays for, and identifying, ER proteins involved in the flip-flopping of Man-P-Dol, Glc-P-Dol, and M5-DLO. Thus, the genetic evidence (7) that implicated Rft1 in the transbilayer movement of M5-DLO aroused considerable interest in this field. The objective of the in vitro studies reported here was to obtain more biochemical evidence to support a role of Rft1 in the transbilayer movement of M5-DLO using sealed microsomal vesicles from yeast as a detergent-free model system.

Although FACE analyses provided solid evidence that large chemical amounts of M5-DLO accumulated in Rft1-depleted yeast cells, several types of experiments indicate that Rft1 is not required for the transverse diffusion of M5-DLO from the cytoplasmic leaflet to the lumenal monolayer in sealed microsomal preparations.

Although the precise extent of depletion of Rft1 in the vesicles used as a model system is uncertain, based on a dilution of endogenous Rft1 into the daughter cells after the switch from galactose to glucose as the primary carbon source, glucose-grown cells would contain less than 0.003% of the initial content of Rft1 after only 15 generations. This number declines to 0.00095% after 20 generations. This estimate is based on the assumptions that Rft1 synthesis ceases immediately upon the switch to glucose as carbon source and that Rft1 is not being actively degraded.

In one approach the formation of M9-DLO was assayed by incubating sealed microsomes with GDP-[³H]Man. Under these in vitro conditions sealed vesicles from wild type and Rft1-depleted cells both formed [³H]M9-DLO, and no accumulation of [³H]M5-DLO was observed. The incorporation of [³H]mannose into Man-P-Dol and DLOs was actually higher in sealed vesicles from Rft1-depleted cells compared with wild type cells, suggesting that Man-P-Dol synthase was up-regulated in response to loss of Rft1 function. It should be noted that the enzymes that require Dol-P as acceptor substrate, Man-P-Dol synthase and GlcNAc phosphate transferase, show an even greater stimulation in the absence of exogenously added Dol-P, suggesting that microsomal fractions from Rft1-depleted cells contain higher levels of endogenous Dol-P available for lipid intermediate biosynthesis relative to wild type cells. Thus, it is possible that this global elevation of biosynthetic enzyme activities contributed to the marked increase of M5-DLO after Rft1 depletion (Fig. 1).

To demonstrate directly that newly synthesized M5-DLO was flipped in sealed microsomes, the same comparison was done by pre-labeling endogenous GlcNAc_1–2-P-P-Dol by incubating the microsomal preparations with UDP-[³H]GlcNAc followed by the addition of unlabeled GDP-Man. The observation that [³H]GlcNAc-labeled M9-DLO was synthesized by control and Rft1-depleted microsomes provides convincing evidence that newly formed M5-DLO was flipped in sealed microsomes from wild type and Rft1-depleted cells at approximately the same rate. It is interesting that the amount of M9-DLO formed by Rft1-depleted microsomes was actually higher than the control microsomes. It is also curious that despite the large amount of M5-DLO detected by FACE analysis, there was no evidence that the newly synthesized intermediate, [³H]GlcNAc-labeled M5-DLO, was isotopically diluted by the relatively large endogenous pool. It is possible that at least a portion of the M5-DLO pool accumulates in a metabolically inert compartment as observed in neuronal ceroid lipofuscinosis mouse models (39).

To obtain additional evidence for M5-DLO flippase activity, another approach was taken assaying the transport of GlcNAc₂-P-P-Dol₁₅, a water-soluble analogue of GlcNAc₂-P-P-Dol₉₅ that has been shown to be flipped, presumably by the same flippase mediating the transbilayer movement of M5-DLO, in microsomes from a variety of tissues and yeast (29–34). The results from this experimental approach also indicated that the analogue entered the sealed microsomal vesicles from wild type and Rft1-depleted cells at essentially the same rate. The observation that transport of the water-soluble analogue appears to reach a higher equilibrium value in Rft1-depleted vesicles suggests that the Rft1-depleted microsomal vesicles are either more highly enriched in ER or may have a slightly larger volume, possibly as a consequence of the unfolded protein response. Entry of the analogue to the lumenal compartment was verified by showing that the labeled disaccharide was transferred to a permeant N-glycosylatable peptide acceptor.

The results presented here using a detergent-free model system are in accord with recent related studies (9–11) which have demonstrated that the flipping of M5-DLO in reconstituted proteoliposomes, assessed by a concanavalin A-capture assay, occurs similarly in the presence or absence of Rft1. The reconstituted flippase activity could be clearly resolved from Rft1 and a glycerophospholipid translocase by Cibacron Blue dye resin chromatography (10), and very significantly, the flippase activity discriminated between the natural intermediate that is flipped, M5-DLO, and M7-DLO (11).

This study also extends the description of the mutant cells by showing that several enzymes involved in Dol-P and lipid intermediate biosynthesis are elevated as a result of the loss of Rft1 function. In addition, the rate of transport of Man-P-Dol₁₀ and Glc-P-Dol₁₀ was slightly elevated in vesicles from Rft1-depleted cells relative to wild type cells (data not included), suggesting that Man-P-Dol and Glc-P-Dol flippases as well as the enzymes involved in Dol-P and lipid intermediate synthesis assayed in Table 2 could also be up-regulated as a compensatory response to the impaired transbilayer movement or accumulation of M5-DLO in vivo. In related studies, per5/rft1 mutants were identified on the basis of their requirement for an intact unfolded protein response for survival (5), and Travers et al. (40) have reported that genes encoding pertinent enzymes in the dolichol pathway, including DPM1, ALG7, and SEC59, are up-regulated during the unfolded protein response in S. cerevisiae based on mRNA levels.

Finally, the results described in this report do not entirely exclude the possibility that Rft1 plays a role in the flipping of M5-DLO in intact living cells. However, it appears that whatever cellular or biophysical constraints that require Rft1 for flipping in vivo apparently no longer exist after cells are disrupted. This observation is reminiscent of the function of Lec35, which is essential for the utilization of Man-P-Dol and Glc-P-Dol in intact cells but not in ER vesicles (12). It is quite possible that some structural aspects of the organization of the ER are lost when cells are disrupted during the preparation of microsomal fractions. In any case, more genetic and biochemical studies will be required to elucidate the precise function of Rft1 in vivo and to identify other ER proteins that are involved in the transbilayer movement of M5-DLO as well as Man-P-Dol and Glc-P-Dol.