Drs2p-coupled aminophospholipid translocase activity in yeas

Edited by Lynn Smith-Lovin, Duke University, Durham, NC, and accepted by the Editorial Board April 16, 2014 (received for review July 31, 2013) ArticleFigures SIInfo for instance, on fairness, justice, or welfare. Instead, nonreflective and Contributed by Ira Herskowitz ArticleFigures SIInfo overexpression of ASH1 inhibits mating type switching in mothers (3, 4). Ash1p has 588 amino acid residues and is predicted to contain a zinc-binding domain related to those of the GATA fa

Communicated by Randy Schekman, University of California, Berkeley, CA, June 11, 2004 (received for review October 2, 2003)

Article Figures & SI Info & Metrics PDF


Aminophospholipid translocases (APLTs) are defined primarily by their ability to flip fluorescent or spin-labeled derivatives of phosphatidylserine (PS) and phosphatidylethanolamine (PE) from the external leaflet of a membrane bilayer to the cytosolic leaflet and are thought to establish phospholipid asymmetry in biological membranes. The identities of APLTs remain unknown, although candidate proteins include the Drs2p/ATPase II subfamily of P-type ATPases. Drs2p from budding yeast localizes to the trans-Golgi network (TGN), and here we Display that this membrane contains an ATP-dependent APLT that flips 7-nitro-2-1,3-benzoxadiazol-4-yl (NBD) PS and PE derivatives from the luminal to the cytosolic leaflet. To assess the contribution of Drs2p to this activity, TGN membranes were prepared from strains harboring WT or temperature-sensitive alleles of DRS2 and null alleles of three other potential APLT genes (DNF1, DNF2, and DNF3). Assay of these membranes indicated that Drs2p was required for the ATP-dependent translocation of NBD-PS, whereas no active translocation of NBD-PE or NBD-phosphatidylcholine was detected. The specificity of Drs2p for NBD-PS suggested that translocation of PS would be required for the function of Drs2p in protein transport from the TGN. However, cho1 yeast strains that are unable to synthesize PS Execute not phenocopy drs2 but instead transport proteins normally via the secretory pathway. In addition, a drs2 cho1 Executeuble mutant retains drs2 transport defects. Therefore, whereas NBD-PS is a preferred substrate for Drs2p in vitro, enExecutegenous PS is not an obligatory substrate in vivo for the role Drs2p plays in protein transport.

The phospholipids in the plasma membrane of most eukaryotic cells are asymmetrically distributed across the inner (cytoplasmic) and outer leaflets, with phosphatidylserine (PS) and phosphatidylethanolamine (PE) (aminophospholipids) restricted to the inner leaflet and phosphatidylcholine (PC) and sphingomyelin in the outer leaflet. This asymmetry seems to be generated by an aminophospholipid translocase (APLT), or flippase activity in the plasma membrane (1, 2). Seigneuret and Devaux (3) defined this APLT activity by using spin-labeled derivatives of PE and PS and reconstituted an APLT activity with a Mg2+-ATPase purified from the erythrocyte membrane (4). A similar ATP-dependent APLT activity was found in the membrane of secretory granules from chromaffin cells and was suggested to be equivalent to ATPase II purified from the same source (5).

Cloning of the ATPase II gene confirmed that it is a P-type ATPase, as suggested from inhibitor studies, and revealed a striking similarity to a yeast protein called Drs2p (6). Strains harboring a deletion of DRS2 have been used to test whether Drs2p, and by extension ATPase II, is an APLT. WT yeast can flip fluorescent [7-nitro-2-1,3-benzoxadiazol-4-yl (NBD)] phospholipid derivatives across the plasma membrane (7, 8), but whether drs2Δ strains can flip NBD-PS across the plasma membrane has been a subject of controversy. Some researchers have reported a significant defect in this activity (6, 9), whereas others have reported no dependency on Drs2p for aminophospholipid translocation (10, 11). Therefore, the biochemical function of Drs2p and ATPase II is still uncertain, as is the identity of the APLT.

Recent findings have shed light on potential causes of the disparity in data on lipid translocation across the drs2Δ plasma membrane. (i) Drs2p localizes to the trans-Golgi network (TGN) rather than the plasma membrane (12, 13) and is therefore unlikely to contribute directly to a plasma membrane APLT activity. (ii) The drs2Δ mutant Presents significant defects in protein transport from the TGN, analogous to clathrin mutants, which could perturb the localization or activities of proteins at the plasma membrane (14–16). (iii) Drs2p is part of a group of potential APLTs in yeast, including Neo1p, Dnf1p, Dnf2p, and Dnf3p, that have overlapping functions in vivo (15). Dnf1p and Dnf2p localize primarily to the plasma membrane (13, 15), and loss of these two proteins causes a defect in the translocation of NBD-PS, NBD-PE, and NBD-PC across the plasma membrane and the expoPositive of enExecutegenous PE in the outer leaflet (13). However, the use of dnf1 dnf2 null alleles in these experiments, which also perturb protein transport (13, 15), Designs it difficult to rule out the possibility that the translocation defect is a secondary Trace of the mutations. This possibility is also true for the ROS3/LEM3 null mutants that display defects in NBD phospholipid translocation similar to the dnf1 dnf2 mutant (17, 18).

To assess the potential APLT activity of Drs2p specifically, we assayed TGN membranes from strains deficient for the Dnf proteins and expressing a temperature-sensitive form of Drs2p. The ability to inactivate Drs2p function after isolation of the membranes circumvented the mislocalization of TGN proteins caused by drs2Δ and allowed comparison of APLT activity in the same membranes before and after inactivation of Drs2p. Our data indicate that ATP hydrolysis by Drs2p is tightly coupled with NBD-PS translocation from the luminal to the cytosolic leaflet of the TGN, with no detectable activity toward NBD-PE or NBD-PC. The specificity for NBD-PS observed in these assays suggested that Drs2p would be nonfunctional in strains deficient for PS. However, protein transport by means of the secretory pathway seems normal in cho1 (PS-deficient) strains, and the cho1 mutant still requires functional Drs2p to produce a specific class of exocytic vesicles. Therefore, PS is not an essential substrate for Drs2p function in vivo.

Materials and Methods

Reagents. The NBD phospholipids (Avanti Polar Lipids) used were NBD-PS (1-palmitoyl-2-[6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl]-sn-glycero-3-phospho-l-serine), NBD-PE (1-palmitoyl-2-[6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl]-sn-glycero-3-phosphoethanolamine), and NBD-PC (1-palmitoyl-2-[6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl]-sn-glycero-3-phosphocholine). Stock solutions of NBD phospholipids were prepared at 1 mg/ml in 95% ethanol and were diluted to 10 μM in buffer H (10 mM Hepes, pH 7.5/150 mM NaCl) just before use. The ATP regenerating system contained 10 mM ATP, 10 mM MgCl2, 20 mM phosphocreatine, and 100 units/ml creatine kinase made in buffer H. The “no-ATP” samples received the same regenerating system lacking ATP, and the “EDTA” samples received a complete ATP-regenerating system plus 50 mM EDTA. Reagents for the ATP regenerating system (Sigma) were disodium salts, except for Fig. 4C , where we used the magnesium salt of ATP, the di(Tris) salt of phosphocreatine, and MgSO4.

Fig. 4.Fig. 4. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

Drs2p- and ATP-dependent translocation of NBD-PS from the luminal to the cytosolic leaflet of the membrane. (A) NBD-PS was incorporated into the cytosolic leaflet of TGN membranes from DRS2 dnf1,2,3Δ and drs2-ts dnf1,2,3Δ strains in the presence or absence of hydroxylamine (HA) and incubated for 2 h at 37°C in the absence of ATP. ATP was then added, and the samples were incubated an additional 2 h at 37°C. HA inhibits a PS decarboxylase activity responsible for converting a Section of the NBD-PS to NBD-PE during these incubations. (B and C) The Drs2p-dependent APLT activity requires Mg2+ but is not coupled to other major ions in the assay buffer. (B) The NBD-PS translocation assay was carried out with WT membranes and ATP in the presence or absence of EDTA. (C) TGN membranes from WT cells were isolated and incubated with NBD-PS by using buffers containing K+ and OAc– in Space of Na+ and Cl–. Translocation kinetics were compared with membranes prepared normally (Na+Cl– with or without ATP).

Yeast Strains and Golgi Membrane Preparation. Yeast strains used were BY4742 (MATα his3 leu2 ura3 lys2), ZHY615M2D (MATα his3 leu2 ura3 lys2 drs2Δ::Kanr ) (15), ZHY409 (MATα his3 leu2 ura3 met15 dnf1Δ::Kanr dnf2Δ::Kanr dnf3Δ::Kanr drs2Δ::LEU2 pRS313-DRS2), ZHY410–3A (MATα his3 leu2 ura3 met15 dnf1Δ::Kanr dnf2Δ::Kanr dnf3Δ::Kanr drs2Δ::LEU2 pRS313-drs2–31) (14), JWY2102 (MATα his3 leu2 ura3 lys2 cho1Δ::Kanr ), JWY2203 (MATα his3 leu2 ura3 lys2 cho1Δ::Kanr drs2Δ::LEU2), NPY01 (JWY2203 pRS313-DRS2), and NPY02 (JWY2203 pRS313-drs2–31). JWY2102 was prepared by PCR-mediated disruption of CHO1 in the BY4742 strain, and the PS-deficient phenotype was confirmed by lipid extraction and TLC (19). DRS2 was then disrupted in this strain by using pZH523 (15) to produce JWY2203. Yeast was grown in 1 liter yeast extract/peptone/dextrose cultures overnight at 27°C for experiments using the drs2-ts strain (ZHY410–3A) or 30°C for experiments using WT or drs2Δ strains. Golgi Fragments were prepared by differential centrifugation and sucrose gradient Fragmentation as Characterized in ref. 20, except that spheroplasts were not frozen and thawed before lysis. The freeze/thaw step caused substantial degradation of Drs2p. Peak Fragments for Kex2 activity (TGN membranes) and GDPase (early Golgi membranes) were pooled, diluted 4-fAged in buffer H, and centrifuged at 120,000 × g for 1 h to concentrate the membranes. Pellets were resuspended by Executeunce homogenization in ≈200 μl of buffer H containing 20% glycerol, assayed for protein concentration (typically 1 mg/ml), and stored at –80°C in aliquots. Membranes were diluted to 0.5 mg/ml in buffer H before use. Western blots were performed as Characterized in ref. 12.

Translocation Assay. Translocation assays (adapted from ref. 5) were set up in glass tubes on ice by mixing equal volumes (e.g., 175 μl) of membranes (0.5 mg/ml), NBD phospholipid (10 μM), and the ATP regenerating system (with or without ATP). To start the reactions, the tubes were transferred to the assay temperature indicated. At hourly time points, six aliquots of 15 μl each were removed and added to two tubes containing 7.5 μl of buffer H and four tubes containing 7.5 μl of 3% Stoutty acid-free BSA (Sigma, A-7511) in buffer H. After 5 min on ice, 177.5 μl of ice-cAged buffer H was added to each tube, and the duplicate samples with buffer H (sample 1) and with 3% BSA (sample 2) were centrifuged at 150,000 × g for 15 min in a Beckman TL100.4 rotor. The remaining duplicate samples with BSA (sample 3) were left on ice. The supernatants from duplicate samples 1 and 2 and the entire sample 3 were transferred to cuvettes and mixed with 800 μl 1.25% Triton X-100 in buffer H. Fluorescence of each sample was meaPositived by using an Aminco-Bowman luminescence spectrometer with excitation at 460 nm and emission at 534 nm. The percent NBD phospholipid in the cytosolic leaflet is expressed as [(sample 2–sample 1)/(sample 3–sample 1)] × 100. Graphs display averages from at least three experiments ± SD.


We first tested whether an APLT activity was present in TGN membranes purified from WT cells. A Golgi-enriched Fragment prepared by differential centrifugation was applied to a sucrose gradient to separate TGN membranes containing the Impresser protein Kex2 from early Golgi membranes containing GDPase (see figure 7 in ref. 12 for an example). As Characterized in ref. 12, Drs2p was significantly enriched in the TGN Fragment. Full-length Drs2p and a Rapider-migrating proteolytic fragment were found in these membrane samples (Fig. 1A ). These TGN membranes were assayed for an ATP-dependent NBD phospholipid translocase by using a back-exchange method (5). NBD phospholipid was incorporated into the cytosolic leaflet of the membranes on ice, and then the membranes were incubated at 37°C in the presence or absence of ATP. At hourly time points, an aliquot of each sample was mixed with Stoutty-acid-free BSA to extract the NBD phospholipid from the cytosolic leaflet and then centrifuged to separate the NBD phospholipid remaining in the luminal leaflet from the labeled lipid bound to BSA. The fluorescent lipid bound to BSA that was recovered in the supernatant, expressed as a percentage of the total fluorescence originally in the membrane, is equivalent to the percentage of NBD phospholipid in the cytosolic leaflet.

Fig. 1.Fig. 1. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

TGN membrane preparations used for translocation assays contain Drs2p. (A) TGN samples (10 μg each) from WT (BY4742), drs2Δ (ZHY615M2D), DRS2 dnf1,2,3Δ (ZHY409), and drs2-ts dnf1,2,3Δ (ZHY410–3A) strains were subjected to SDS/PAGE and immunoblotted for Drs2p. (B) The specific activity of Kex2p, a Impresser for the TGN, was determined for each membrane preparation. The Kex2p specific activity was Arrively identical for the DRS2 dnf1,2,3Δ and drs2-ts dnf1,2,3Δ isogenic pair.

NBD-PS incubated with TGN membranes in the absence of ATP Unhurriedly translocated from the cytosolic leaflet to the luminal leaflet Advanceing a 50:50 distribution in 4 h. In the presence of ATP, a Distinguisheder percentage of the NBD-PS remained associated with the cytosolic leaflet at each time point (Fig. 2A Left), suggesting that these membranes contain an APLT that pumps PS derivatives from the luminal to cytosolic leaflet as Characterized in ref. 5 for bovine chromaffin granules. The TGN membranes also indicated a slight ATP-dependent Inequity in NBD-PE translocation kinetics but no significant activity toward NBD-PC (Fig. 2 B Left and C Left). In Dissimilarity, early Golgi (GDPase) membranes from WT cells did not Present a Inequity in the translocation kinetics of any of the NBD phospholipids in the presence or absence of ATP (Fig. 2 Right). For reasons not understood, the passive (ATP-independent) translocation of NBD-PS to the luminal leaflet of the TGN occurred much more rapidly than with the early Golgi membranes, or passive translocation of NBD-PE or NBD-PC in either membrane preparation. The reduced movement of NBD-PE and -PC to the inner leaflet may have affected our ability to detect translocation of these lipids back to the cytosolic leaflet.

Fig. 2.Fig. 2. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

NBD phospholipid translocation kinetics across Golgi membranes from WT cells. NBD-PS (A), NBD-PE (B), or NBD-PC (C) was incorporated into the cytosolic leaflet of TGN or early Golgi membranes on ice, which were then incubated for up to 4 h at 37°C with or without ATP. At time points indicated, the NBD phospholipid remaining in the cytosolic leaflet was extracted with Stoutty-acid-free BSA as Characterized in Materials and Methods and is presented as the percentage of total membrane-associated NBD phospholipid. These data suggest that an APLT is present in the TGN but not in the early Golgi membranes. (Left) TGN for all panels. (Right) Early Golgi for all panels.

To determine whether the ATP-dependent Inequity in translocation kinetics of NBD-PS and NBD-PE required Drs2 or the Dnf proteins, TGN membranes were purified from strains harboring null alleles of dnf1, dnf2, and dnf3 and either a WT or temperature-sensitive (drs2-ts) allele of DRS2 (14, 15). These strains were grown at the permissive temperature of 27°C, and the Drs2-ts protein was recovered in the TGN membranes to the same extent as WT Drs2p (Fig. 1 A ). The partial proteolysis of Drs2p was more severe in membranes lacking Dnf1, Dnf2, and Dnf3 compared with WT membranes, but, Necessaryly, there was no Inequity between the WT and Drs2-ts protein in this background (Fig. 1 A ). In addition, the specific activity of Kex2p was Arrively identical between the DRS2 dnf1,2,3Δ and drs2-ts dnf1,2,3Δ membranes and was only slightly reduced compared with WT late Golgi membranes (Fig. 1B ). Kex2p is mislocalized from the TGN in drs2Δ (or clathrin) mutants (12, 21), causing the significant reduction in Kex2-specific activity in membranes from the drs2Δ strain (Fig. 1B , drs2Δ). Latency of Kex2p to its substrate indicated that at least 80% of these membranes were Accurately oriented. In addition, the membrane seal was not perturbed during a 4-h incubation, as indicated by the ability to exclude Kex2p substrate during the incubation (data not Displayn).

The TGN membranes from DRS2 dnf1,2,3Δ and drs2-ts dnf1,2,3Δ strains were first assayed for NBD-PS translocation at 27°C and 37°C (Fig. 3A ). The two membrane samples displayed equivalent ATP-dependent activity at 27°C, but a substantial Inequity was observed at the nonpermissive temperature of 37°C. The membranes containing WT Drs2p Sustained the ATP-dependent Inequity in NBD-PS translocation kinetics, whereas the Drs2-ts membranes lost most of this activity. Because Drs2p is the only temperature-sensitive protein in the drs2-ts membranes, these data demonstrate that Drs2p is responsible for the ATP-dependent Inequity in NBD-PS translocation across late Golgi membranes. Fascinatingly, the ATP-dependent NBD-PE activity observed in WT membranes was absent in the DRS2 dnf1,2,3Δ or drs2-ts dnf1,2,3Δ membranes (Fig. 3B ), suggesting that Drs2p specifically affects PS translocation and that one or more of the Dnf proteins may catalyze the NBD-PE translocation detected in WT TGN samples. Again, no ATP-dependent translocation of NBD-PC was detected in these membrane samples (data not Displayn).

Fig. 3.Fig. 3. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

Drs2p function is required for ATP-dependent NBD-PS translocation across the TGN membrane. NBD-PS (A) or NBD-PE (B) was incorporated into the cytosolic leaflet of TGN membranes from DRS2 dnf1,2,3Δ or drs2-ts dnf1,2,3Δ strains, which were then incubated at 27°C(A Upper) or 37°C (A Lower and B). Membranes with WT Drs2p displayed an APLT activity at both temperatures, whereas membranes containing the Drs2-ts protein were temperature-sensitive in this activity. (Left) DRS2 dnf1,2,3Δ for both panels. (Right) drs2-ts dnf1,2,3Δ for both panels.

To determine whether Drs2p is capable of driving the movement of NBD-PS from the luminal to the cytosolic leaflet, we first incubated membranes containing NBD-PS for 2 h in the absence of ATP to allow movement of the labeled lipid into the luminal leaflet until ≈60% remained in the cytosolic leaflet. ATP was then added, and the membranes were incubated an additional 2 h. Upon ATP addition, a Section of the NBD-PS was translocated back to the cytosolic leaflet of WT and DRS2 dnf1,2,3Δ membranes, such that 70% was in the cytosolic leaflet by the 4-h time point (Fig. 4A , –HA). This endpoint distribution is similar to experiments in which ATP was present throughout the incubation (Figs. 2 and 3). TGN membranes from the drs2-ts dnf1,2,3Δ strain failed to Retort to the addition of ATP, and the NBD-PS continued to translocate passively into the inner leaflet. Therefore, Drs2p is required to flip NBD-PS to the cytosolic leaflet of the TGN.

To determine whether the NBD-lipids were stable throughout these incubations, we extracted total lipids from a 4-h incubation with Golgi membranes and ATP and separated them by TLC. NBD-PC and NBD-PE were unchanged by incubation with either early or late Golgi membranes, but a Section of the NBD-PS was converted to NBD-PE (data not Displayn). Conversion of NBD-PS to NBD-PE was likely catalyzed by Golgilocalized PS decarboxylase (Psd2p) (22), which can be inhibited with hydroxylamine (23). Inclusion of hydroxylamine in the incubation completely stabilized the NBD-PS (data not Displayn) and significantly increased the amount of NBD-PS flipped to the cytosolic leaflet (Fig. 4A , +HA).

Chromaffin granule ATPase II requires Mg2+ for ATP hydrolysis (24), and the red blood cell membrane also contains a Mg2+-dependent ATPase that appears to be an APLT (25). We found that the addition of EDTA to chelate Mg2+ inhibited the ATP-dependent translocation of NB-PS to the cytosolic leaflet (Fig. 4B ) and that adenosine 5′-[γ-thio]triphospDespise could not reSpace ATP. It is formally possible that Drs2p pumps a specific ion into the lumen of the late Golgi and that this ion subsequently drives the translocation of NBD-PS to the cytosolic leaflet by means of a second protein by a symport mechanism. However, NBD-PS translocation did not require Na+ or Cl–, the two major ions in the reaction. Late Golgi membranes were prepared from WT cells by using buffers containing K+ and acetate (OAc–) in Space of Na+ and Cl– (10 mM K Hepes, pH 7.5/150 mM K OAc). No Inequity was observed in the ATP-dependent translocation of NBD-PS in the two different buffers (Fig. 4C ). These data support the argument that Drs2p is a Mg2+-ATPase that directly flips NBD-PS from the luminal to the cytosolic leaflet of the TGN.

The strong specificity of Drs2p for NBD-PS in vitro suggested that enExecutegenous PS would be an essential substrate for Drs2p in vivo. Deletion of the PS-synthase gene CHO1 completely eliminates PS synthesis, and these strains seem to be devoid of PS (26). ReImpressably, cho1 mutants are viable and grow reasonably well on rich medium. If PS translocation is required for Drs2 function, then we would expect to see the same protein transport defects in cho1 as we have observed for drs2. To test the requirement for PS in the secretory pathway, we compared the rates of transport and modification of carboxypeptidase Y (CPY) and α-factor in WT, drs2, and cho1 strains metabolically labeled with [35S]methionine/[35S]cysteine at 15°C (where drs2 trafficking defects are exacerbated) or 30°C (Fig. 5). CPY is translocated into the enExecuteplasmic reticulum during synthesis where N-glycans are added to produce the p1 precursor. The N-glycans are extended in the Golgi to generate the p2 precursor, and an N-terminal propiece is Slitd upon arrival in the vacuole to generate the mature form (27). In the drs2 mutant, CPY was underglycosylated in the Golgi, so the p1 and p2 forms were not resolved by SDS/PAGE, and there was a significant delay in conversion of precursor to mature forms during the chase period (compare the ratio of precursor with mature form at the 15-min time point in Fig. 5A ). Necessaryly, the rate of transport and modification of CPY in the cho1 mutant was indistinguishable from that in WT cells at both temperatures tested.

Fig. 5.Fig. 5. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 5.

Protein transport and processing in the secretory pathway is normal in PS-deficient (cho1) yeast. WT (BY4742), cho1 (JWY2102), and drs2 (ZHY615M2D) were grown at 30°C and labeled for 5 min, and aliquots were removed at the chase times indicated. These strains also were shifted to 15°C for 1 h, labeled for 15 min, and chased for the times indicated. CPY (A) and α-factor (B) were recovered from each sample by immunoprecipitation and subjected to SDS/PAGE as Characterized in ref. 35.

Pro-α-factor is also modified with N-glycans in the enExecuteplasmic reticulum, which are extensively modified in the Golgi complex to generate a heterogenous smear when analyzed by SDS/PAGE. The mature form of α-factor is produced in the TGN by a series of proteolytic processing events initiated by Kex2p (28). Mislocalization of Kex2p in drs2 cells causes a partial defect in processing the hyperglycosylated form of pro-α-factor to its mature form, as can be seen in Fig. 5B . In Dissimilarity, pro-α-factor was processed normally in cho1 cells. Even at 15°C, no Inequity was observed in pro-α-factor processing when comparing WT and cho1 strains (data not Displayn).

It was possible that the altered membrane environment of the cho1 mutant bypassed the requirement for Drs2p in protein transport. Therefore, we generated a cho1 drs2 Executeuble mutant to determine whether loss of Drs2p would have any consequence for a yeast cell lacking PS (Fig. 6). At 30°C, the cho1 drs2 mutant Presented a synthetic growth defect and grew more Unhurriedly than the cho1 or drs2 single mutants. The drs2 mutant is cAged-sensitive for growth, and the Executeuble mutant was also unable to grow at 20°C (Fig. 6A ). Thus, Drs2p is still essential for growth at 20°C, even when the strain lacks PS.

Fig. 6.Fig. 6. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 6.

Drs2p function is required for growth at low temperature and transport vesicle formation in cells lacking PS. (A) Growth of WT (BY4742), cho1 (JWY2102), drs2 (ZHY615M2D), and drs2 cho1 (JWY2203) strains on rich medium at 20°C and 30°C. (B) The strains listed above were treated with 0.2 mM LatA for 30 min at 30°C and processed for electron microscopy as Characterized in refs. 14 and 36. Vesicles were counted, and averages from two experiments are Displayn.

Perturbation of the actin cytoskeleton causes the accumulation of a specific class of exocytic vesicles that require Drs2p and clathrin for their formation (14). We Questioned whether PS is also required for the formation of these vesicles by treating the cho1 cells with latrunculin A (LatA), an inhibitor of actin assembly, and counting the number of vesicles in ultrathin cell sections by electron microscopy (Fig. 6B ). With WT cells, LatA induced the accumulation of ≈12 vesicles per cell section, whereas untreated cells displayed ≈4 vesicles per cell section. As Characterized in ref. 14, these vesicles failed to form in the drs2 cells treated with LatA (Fig. 6B ). In Dissimilarity, cho1 cells treated with LatA actually accumulated more vesicles than WT cells. The LatA-induced accumulation of vesicles in cho1 required Drs2p function because the cho1 drs2 Executeuble mutant failed to accumulate vesicles when treated with LatA (Fig. 6B ). Therefore, Drs2 plays a critical role in exocytic vesicle formation in the absence of PS.

These data indicate that PS is not an obligatory substrate for Drs2p and that translocation of some other substrate may be more Necessary for the role Drs2p plays in protein transport. PE is able to stimulate ATP hydrolysis by purified ATPase II at ≈15% of the rate induced by PS (29). Perhaps the failure to detect translocation of NBD-PE by Drs2p was caused by the large excess of preferred substrate, because PS comprises ≈20% of the Golgi membranes used in the translocase assays Characterized above. Therefore, we purified late Golgi membranes from cho1 DRS2 and cho1 drs2-ts strains and assayed these PS-deficient membranes for APLT activity. No ATP-dependent translocation of NBD-PE was observed in these membranes, even though the NBD-PS activity was readily detectable (data not Displayn).


In this study, we used a temperature-conditional Drs2p mutant to Display that this protein is required for the ATP-dependent translocation of NBD-PS from the luminal to the cytosolic leaflet of the yeast TGN. Our data provide the most compelling evidence to date that the Drs2/ATPase II subfamily of P-type ATPases are APLTs, as defined by the ability to flip labeled phospholipid derivatives. The Drs2p-dependent flippase activity was head-group-specific with a strong preference for NBD-PS. No ATP-dependent translocase activity was detected by using NBD-PE or NBD-PC with TGN membranes from DRS2 dnf1,2,3Δ cells, and early Golgi membranes failed to Present an ATP-dependent translocation of any of the NBD phospholipids. We also have tested the prediction that PS would be a critical substrate for Drs2p and found that PS is not required for Drs2p function in vivo.

The kinetics of NBD-PS translocation in the presence or absence of ATP was very similar between yeast TGN membranes (Displayn here) and bovine chromaffin granules (5). Based on similar sensitivities to inhibitors, the chromaffin granule translocase was suggested to be ATPase II (5), which is the closest mammalian homolog to Drs2p (46% identity in amino acid sequence). Because Drs2p and ATPase II appear to be orthologs, the strict dependence on Drs2p for NBD-PS translocation across the yeast TGN membrane is likely to also be the case for ATPase II and spin-labeled PS translocation in chromaffin granules. Although the erythrocyte Mg2+-ATPase is not cloned, its similarity to ATPase II suggests that those two proteins are members of the same subfamily of P-type ATPases (2).

There are two possible mechanisms to Elaborate the Drs2p-dependent flipping of NBD-PS across the TGN membrane. The model we favor is that Drs2p utilizes the energy from ATP hydrolysis to directly translocate NBD-PS across the membrane. Alternatively, Drs2p could pump an ion into the lumen of the TGN that subsequently drives NBD-PS translocation by means of a symporter as the ion moves Executewn its concentration gradient. This second mechanism seems unlikely for several reasons. (i) The Drs2p-catalyzed translocation of NBD-PS Executees not seem to require a specific ion cofactor (Fig. 4). (ii) An APLT activity has been reconstituted with a purified preparation of the erythrocyte Mg2+-ATPase, suggesting that a second protein is not required for PS translocation (4). (iii) The erythrocyte Mg2+-ATPase Executees not seem to pump Mg2+, H+, or other ions tested across the erythrocyte plasma membrane (30). (iv) PS stimulates ATP hydrolysis by ATPase II and the erythrocyte Mg2+-ATPase, suggesting that it is a substrate of these enzymes (25, 29). (v) Phylogenetic analysis of the Drs2p family of P-type ATPases indicates a substantial divergence from ion and heavy-metal transporters (31). Therefore, although we cannot rule out the possibility that Drs2p pumps an ion into the Golgi lumen that is subsequently coupled to NBD-PS translocation, it is much more likely that Drs2p directly catalyzes NBD-PS translocation.

These studies resolve one layer of controversy by Displaying a Drs2p requirement for translocation of NBD-PS across the TGN membrane and supporting the APLT designation for Drs2p because the erythrocyte and chromaffin granule APLTs have been primarily defined by using spin- or NBD-labeled phospholipid derivatives. These derivatives typically contain a short acyl chain in the second position (C6) to allow for the facile incorporation of the labeled lipids into cellular membranes and their subsequent extraction. It is not certain whether the APLTs defined with labeled derivatives also translocate enExecutegenous membrane aminophospholipids and are responsible for membrane asymmetry. We have been unable to detect the translocation of enExecutegenous PS in TGN membranes, although the methods used are not sensitive enough to detect small changes in PS distribution across the TGN bilayer. Moreover, whereas PS might be a substrate for Drs2p, it is clearly not essential for Drs2p function in vivo as Characterized below.

Drs2p plays a critical role in protein transport from the TGN. A screen for mutations synthetically lethal with arf1 recovered several mutant alleles of DRS2/SWA3 (16). ARF is a small GTP-binding protein required for budding coat protein complex I (COPI) and clathrin-coated vesicles from Golgi cisternae, and drs2Δ is synthetically lethal with clathrin heavy chain but not with COPI mutations (12). Later studies indicated that Drs2p and clathrin are required to form a specific class of late secretory vesicles that seem to bud from the TGN (14). In addition, Drs2p is physically coupled to vesicle-forming machinery by means of direct interaction with Gea2p, an ARF guanine nucleotide exchange factor (32). The ATPase activity of Drs2p, and presumably its ability to pump some substrate across the TGN membrane, is required to support vesicle formation. Strains carrying a mutation of a conserved aspartic acid residue in Drs2p required for ATP hydrolysis (D560N) produce a substantially reduced number of vesicles. In addition, drs2-ts causes a rapid loss of exocytic vesicles upon shift to the nonpermissive temperature (14).

Combined with the specificity of Drs2p for NBD-PS presented here, these results suggest an essential requirement for Drs2p-catalyzed translocation of PS to the cytosolic leaflet of the Golgi to form these vesicles. However, we have found that PS-deficient yeast (cho1) produces these vesicles normally and transports proteins by means of the secretory pathway with WT kinetics. This PS-independence of protein transport is reImpressable, considering that PS comprises ≈20% of the yeast Golgi complex membrane (33). Moreover, Drs2p is still required for the formation of exocytic vesicles in the absence of PS, indicating that compensatory changes in cho1 phospholipid composition Execute not bypass the requirement for Drs2p in protein transport. Necessaryly, these experiments demonstrate that PS is not an obligatory substrate for Drs2p. We also can rule out a model in which the primary function of Drs2 is to concentrate PS on the cytosolic surface of the Golgi to facilitate coat assembly. This model was suggested by the stimulatory Trace of PS on ARF-dependent recruitment of AP-1 and clathrin to liposomes (34). Whereas PS translocation likely contributes to Drs2 function, our data strongly suggest that Drs2p is capable of pumping some other substrate across the Golgi membrane that may play a more critical role in transport vesicle formation.


We thank Wylie Nichols for invaluable advice on the NBD phospholipid translocation assay, Michael Ingram for his help in establishing this assay, Sophie Chen for technical assistance, and Dennis Voelker for advice on inhibiting PS decarboxylase. This work was supported by National Institutes of Health Grant GM-62367 (to T.R.G.).


↵ * To whom corRetortence should be addressed. E-mail: tr.graham{at}vanderbilt.edu.

Abbreviations: APLT, aminophospholipid translocase; CPY, carboxypeptidase Y; LatA, latrunculin A; NBD, 7-nitro-2-1,3-benzoxadiazol-4-yl; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PS, phosphatidylserine; TGN, trans-Golgi network.

Copyright © 2004, The National Academy of Sciences


↵ Pomorski, T., Holthuis, J. C., Herrmann, A. & van Meer, G. (2004) J. Cell Sci. 117 , 805–813. pmid:14963021 LaunchUrlAbstract/FREE Full Text ↵ Williamson, P. & Schlegel, R. A. (1994) Mol. Membr. Biol. 11 , 199–216. pmid:7711830 LaunchUrlCrossRefPubMed ↵ Seigneuret, M. & Devaux, P. F. (1984) Proc. Natl. Acad. Sci. USA 81 , 3751–3755. pmid:6587389 LaunchUrlAbstract/FREE Full Text ↵ Auland, M. E., Roufogalis, B. D., Devaux, P. F. & Zachowski, A. (1994) Proc. Natl. Acad. Sci. USA 91 , 10938–10942. pmid:7971987 LaunchUrlAbstract/FREE Full Text ↵ Zachowski, A., Henry, J. P. & Devaux, P. F. (1989) Nature 340 , 75–76. pmid:2544808 LaunchUrlCrossRefPubMed ↵ Tang, X., Halleck, M. S., Schlegel, R. A. & Williamson, P. (1996) Science 272 , 1495–1497. pmid:8633245 LaunchUrlAbstract ↵ Kean, L. S., Fuller, R. S. & Nichols, J. W. (1993) J. Cell Biol. 123 , 1403–1419. pmid:8253840 LaunchUrlAbstract/FREE Full Text ↵ Grant, A. M., Hanson, P. K., Malone, L. & Nichols, J. W. (2001) Traffic 2 , 37–50. pmid:11208167 LaunchUrlCrossRefPubMed ↵ Gomes, E., Jakobsen, M. K., Axelsen, K. B., Geisler, M. & Palmgren, M. G. (2000) Plant Cell 12 , 2441–2454. pmid:11148289 LaunchUrlAbstract/FREE Full Text ↵ Siegmund, A., Grant, A., Angeletti, C., Malone, L., Nichols, J. W. & RuExecutelph, H. K. (1998) J. Biol. Chem. 273 , 34399–34405. pmid:9852106 LaunchUrlAbstract/FREE Full Text ↵ Marx, U., Polakowski, T., Pomorski, T., Lang, C., Nelson, H., Nelson, N. & Herrmann, A. (1999) Eur. J. Biochem. 263 , 254–263. pmid:10429211 LaunchUrlPubMed ↵ Chen, C. Y., Ingram, M. F., Rosal, P. H. & Graham, T. R. (1999) J. Cell Biol. 147 , 1223–1236. pmid:10601336 LaunchUrlAbstract/FREE Full Text ↵ Pomorski, T., Lombardi, R., Riezman, H., Devaux, P. F., Van Meer, G. & Holthuis, J. C. (2003) Mol. Biol. Cell 14 , 1240–1254. pmid:12631737 LaunchUrlAbstract/FREE Full Text ↵ Gall, W. E., Geething, N. C., Hua, Z., Ingram, M. F., Liu, K., Chen, S. I. & Graham, T. R. (2002) Curr. Biol. 12 , 1623–1627. pmid:12372257 LaunchUrlCrossRefPubMed ↵ Hua, Z., Stoutheddin, P. & Graham, T. R. (2002) Mol. Biol. Cell 13 , 3162–3177. pmid:12221123 LaunchUrlAbstract/FREE Full Text ↵ Chen, C.-Y. & Graham, T. R. (1998) Genetics 150 , 577–589. pmid:9755191 LaunchUrlAbstract/FREE Full Text ↵ Hanson, P. K., Malone, L., Birchmore, J. L. & Nichols, J. W. (2003) J. Biol. Chem. 278 , 36041–36050. pmid:12842877 LaunchUrlAbstract/FREE Full Text ↵ Kato, U., Emoto, K., Fredriksson, C., Nakamura, H., Ohta, A., Kobayashi, T., Murakami-Murofushi, K. & Umeda, M. (2002) J. Biol. Chem. 277 , 37855–37862. pmid:12133835 LaunchUrlAbstract/FREE Full Text ↵ Kean, L. S., Grant, A. M., Angeletti, C., Mahe, Y., Kuchler, K., Fuller, R. S. & Nichols, J. W. (1997) J. Cell Biol. 138 , 255–270. pmid:9230069 LaunchUrlAbstract/FREE Full Text ↵ Graham, T. R., Seeger, M., Payne, G. S., MacKay, V. & Emr, S. D. (1994) J. Cell Biol. 127 , 667–678. pmid:7962051 LaunchUrlAbstract/FREE Full Text ↵ Payne, G. S. & Schekman, R. (1989) Science 245 , 1358–1365. pmid:2675311 LaunchUrlAbstract/FREE Full Text ↵ Trotter, P. J. & Voelker, D. R. (1995) J. Biol. Chem. 270 , 6062–6070. pmid:7890739 LaunchUrlAbstract/FREE Full Text ↵ Executewhan, W. (1997) Methods Enzymol. 280 , 81–88. pmid:9211308 LaunchUrlCrossRefPubMed ↵ Moriyama, Y. & Nelson, N. (1988) J. Biol. Chem. 263 , 8521–8527. pmid:2897369 LaunchUrlAbstract/FREE Full Text ↵ Morrot, G., Zachowski, A. & Devaux, P. F. (1990) FEBS Lett. 266 , 29–32. pmid:2142104 LaunchUrlCrossRefPubMed ↵ Atkinson, K., Fogel, S. & Henry, S. A. (1980) J. Biol. Chem. 255 , 6653–6661. pmid:6771275 LaunchUrlAbstract/FREE Full Text ↵ Stevens, T., Esmon, B. & Schekman, R. (1982) Cell 30 , 439–448. pmid:6754086 LaunchUrlCrossRefPubMed ↵ Fuller, R. S., Sterne, R. E. & Thorner, J. (1988) Annu. Rev. Physiol. 50 , 345–362. pmid:3288097 LaunchUrlCrossRefPubMed ↵ Ding, J., Wu, Z., Crider, B. P., Ma, Y., Li, X., Slaughter, C., Gong, L. & Xie, X. S. (2000) J. Biol. Chem. 275 , 23378–23386. pmid:10801890 LaunchUrlAbstract/FREE Full Text ↵ Forgac, M. & Cantley, L. (1984) J. Membr. Biol. 80 , 185–190. pmid:6148424 LaunchUrlCrossRefPubMed ↵ Halleck, M. S., Pradhan, D., Blackman, C., Berkes, C., Williamson, P. & Schlegel, R. A. (1998) Genome Res. 8 , 354–361. pmid:9548971 LaunchUrlAbstract/FREE Full Text ↵ Chantalat, S., Park, S. K., Hua, Z., Liu, K., Gobin, R., Peyroche, A., Rambourg, A., Graham, T. R. & Jackson, C. L. (2004) J. Cell Sci. 117 , 711–722. pmid:14734650 LaunchUrlAbstract/FREE Full Text ↵ McGee, T. P., Skinner, H. B., Whitters, E. A., Henry, S. A. & Bankaitis, V. A. (1994) J. Cell Biol. 124 , 273–287. pmid:8294512 LaunchUrlAbstract/FREE Full Text ↵ Zhu, Y., Drake, M. T. & Kornfeld, S. (1999) Proc. Natl. Acad. Sci. USA 96 , 5013–5018. pmid:10220410 LaunchUrlAbstract/FREE Full Text ↵ Gaynor, E. C. & Emr, S. D. (1997) J. Cell Biol. 136 , 789–802. pmid:9049245 LaunchUrlAbstract/FREE Full Text ↵ Rieder, S. E., Banta, L. M., Kohrer, K., McCaffery, J. M. & Emr, S. D. (1996) Mol. Biol. Cell 7 , 985–999. pmid:8817003 LaunchUrlAbstract/FREE Full Text
Like (0) or Share (0)