These authors contributed equally to this work.
Roles for the Drs2p–Cdc50p Complex in Protein Transport and Phosphatidylserine Asymmetry of the Yeast Plasma Membrane
Article first published online: 1 SEP 2006
Volume 7, Issue 11, pages 1503–1517, November 2006
How to Cite
Chen, S., Wang, J., Muthusamy, B.-P., Liu, K., Zare, S., Andersen, R. J. and Graham, T. R. (2006), Roles for the Drs2p–Cdc50p Complex in Protein Transport and Phosphatidylserine Asymmetry of the Yeast Plasma Membrane. Traffic, 7: 1503–1517. doi: 10.1111/j.1600-0854.2006.00485.x
- Issue published online: 1 SEP 2006
- Article first published online: 1 SEP 2006
- Received 6 January 2006, revised and accepted for publication 7 August 2006
- annexin V;
- Golgi complex;
- papuamide B;
- P-type ATPase
Drs2p, a P-type adenosine triphosphatase required for a phosphatidylserine (PS) flippase activity in the yeast trans Golgi network (TGN), was first implicated in protein trafficking by a screen for mutations synthetically lethal with arf1 (swa). Here, we show that SWA4 is allelic to CDC50, encoding a membrane protein previously shown to chaperone Drs2p from the endoplasmic reticulum to the Golgi complex. We find that cdc50Δ exhibits the same clathrin-deficient phenotypes as drs2Δ, including delayed transport of carboxypeptidase Y to the vacuole, mislocalization of resident TGN enzymes and the accumulation of aberrant membrane structures. These trafficking defects precede appearance of cell polarity defects in cdc50Δ, suggesting that the latter are a secondary consequence of disrupting Golgi function. Involvement of Drs2p–Cdc50p in PS translocation suggests a role in restricting PS to the cytosolic leaflet of the Golgi and plasma membrane. Annexin V binding and papuamide B hypersensitivity indicate that drs2Δ or cdc50Δ causes a loss of plasma membrane PS asymmetry. However, clathrin and other endocytosis null mutants also exhibit a comparable loss of PS asymmetry, and studies with drs2-ts and clathrin (chc1-ts) conditional mutants suggest that loss of plasma membrane asymmetry is a secondary consequence of disrupting protein trafficking.
ADP ribosylation factor (ARF) is a small GTP-binding protein required to bud both coat protein (COP) I-coated and clathrin-coated vesicles from Golgi membranes (1,2), but how ARF determines which coat complex to recruit at a specific time and place is not very well understood. To identify accessory proteins that function with ARF in vesicle-mediated protein transport, we screened for yeast mutants that are synthetically lethal with arf1(3), a genetic interaction that often predicts that the affected proteins work together in the same pathway or in parallel pathways. Seven different complementation groups were identified (swa1–swa7, for synthetically lethal with arf1), and their characterization indicates that this screen was biased for disruption of clathrin function, most obviously by the finding that SWA5 is allelic to the clathrin heavy chain gene (CHC1) (3). In addition, SWA2 codes for yeast auxilin, a co-chaperone with Hsc70 required for uncoating clathrin-coated vesicles (4,5). The SWA3/DRS2 gene encodes a type 4 P-type adenosine triphosphatase (ATPase) (hereafter referred to as a P4-ATPase) and a potential phospholipid translocase (flippase) that has also been linked to clathrin function (6–8).
Strains harboring drs2Δ are viable, and this mutation is synthetically lethal with clathrin heavy chain temperature-sensitive (chc1-ts) alleles but not with ts alleles of COP I or COP II subunits. Moreover, drs2Δ cells exhibit several phenotypes indicating a loss of clathrin function at the Golgi including 1) mislocalization of trans Golgi network (TGN) proteins, 2) accumulation of swollen Golgi cisternae and 3) a substantial decrease in bona fide clathrin-coated vesicles that can be purified from the drs2Δ mutant relative to normal cells. In addition, Drs2p localizes to the TGN where it is positioned to directly contribute to clathrin-coated vesicle (CCV) budding (6). The mislocalization of TGN resident proteins observed in clathrin and drs2Δ mutants appears to reflect a loss of clathrin-dependent transport of these proteins between the TGN and the endosomal system. Surprisingly, Drs2p and clathrin are also required to form one of the two classes of exocytic vesicles required for polarized growth of the bud membrane. Importantly, a drs2-ts mutant exhibits a defect in vesicle biogenesis within 30 min after inactivation, and mutation of the aspartic acid required for ATP hydrolysis (D560) inactivates Drs2p function in vivo(6,7). These data suggest that Drs2p must actively pump some substrate, presumably phospholipid, across the TGN membrane to support the vesicle budding mechanism (9).
Drs2p is nearly 50% identical to mammalian ATPase II (now called ATP8A1), which was purified from chromaffin granules and proposed to catalyze a phosphatidylserine (PS) translocase (flippase) activity found with these exocytic vesicles (10,11). A similar flippase activity specific for PS and phosphatidylethanolamine (PE) is present in the plasma membrane of most eukaryotic cells and appears to be responsible for the fact that the plasma membrane is asymmetric, with most of the PS and PE restricted to the cytosolic leaflet, while the extracellular leaflet is primarily composed of phosphatidylcholine (PC) and sphingolipids (12). Drs2p and ATP8A1 are founding members of a large subfamily of P-type ATPases (type IV) that includes five yeast (Drs2p, Neo1p, Dnf1p, Dnf2p and Dnf3p) and 14 mammalian members (ATP8A1-ATP10D) (8,13). The relationship of P4-ATPases to flippases and membrane asymmetry has been controversial (11,14–16), although these connections are growing stronger. For example, overexpression of ATP8B1 (FIC1) in CHOK1 cells leads to an increase in PS translocase activity in these cells (17). Dnf1p and Dnf2p localize to the yeast plasma membrane and their deletion (dnf1,2Δ) abolishes translocation of PE, PC and PS fluorescent [C6-7-nitro-2-1,3-benzoxadiazol-4-yl (NBD)] derivatives across the plasma membrane and leads to aberrant exposure of endogenous PE on the outer leaflet. These phenotypes are exacerbated by additional deletion of DRS2 (dnf1,2Δ drs2Δ) (18). The dnf1,2Δ mutations also perturb the endocytic pathway (18,19), and it is uncertain whether Dnf1p and Dnf2p directly translocate phospholipid or whether the trafficking defect depletes other proteins from the cell surface that are more directly involved in phospholipid translocation. For Drs2p, we were able to disengage its roles in protein trafficking and lipid translocation by assaying translocase activity with purified Golgi membranes containing a Drs2-ts conditional mutant. These studies indicated that Drs2p is directly coupled to a translocase activity specific for NBD-PS as no activity was found for NBD derivatives of PE or PC (20). In addition to NBD-PS, an NBD-PE translocase activity attributable to Drs2p has also been observed in isolated secretory vesicles, further supporting the proposed aminophospholipid translocase activity for Drs2p (21). However, the influence of this Drs2p-coupled flippase activity on the establishment of PS asymmetry of membranes was not determined in any of these studies.
Here, we describe the cloning of SWA4 and its identification as CDC50(22). The Tanaka laboratory reported that Cdc50p binds Drs2p and is required for export of the Drs2p–Cdc50p complex from the endoplasmic reticulum (ER) to the Golgi (23), which we confirm here. Cdc50p is an integral membrane protein and is a member of a yeast protein family that includes Lem3p (also called Ros3p) and Crf1p (24,25). Deletion of CDC50 causes retention of Drs2p in the ER but has no influence on Dnf1p transport to the plasma membrane. Conversely, deletion of LEM3 causes retention of Dnf1p in the ER but does not perturb Drs2p transport to the Golgi (23). Lem3p is probably also required for plasma membrane localization of Dnf2p because lem3Δ disrupts phospholipid translocation and PE asymmetry at the plasma membrane similarly to dnf1,2Δ double mutants (18,23,26,27). The cdc50Δ and drs2Δ mutants exhibit a defect in polarized cell growth after shifting to the nonpermissive growth temperature of 18°C for 12 h (23,25). Here, we show that cdc50Δ mutants show the same late Golgi defects as drs2Δ and suggest that the cell polarity defects of cdc50Δ are a secondary consequence of disrupting Golgi function. We also examined the influence of drs2Δ, cdc50Δ, lem3Δ and dnfΔ deletions on PS asymmetry of the plasma membrane and found that all these mutants aberrantly exposed PS on the extracellular leaflet. However, clathrin, ARF and endocytosis mutants also expose PS on the extracellular leaflet. This phenotype was associated with chronic loss of these proteins because drs2-ts and chc1-ts mutants did not show a loss of PS asymmetry within 1 h of inactivating these proteins.
Cloning of SWA4
SWA4 was cloned by complementation of the synthetic lethality between swa4-2 and arf1. The original swa mutations were induced in an arf1Δ ARF2 ade2 ade3 parental strain harboring wild-type (WT) copies of ARF1 and ADE3 linked on an episomal plasmid. The plasmid-borne ADE3 gene allows cellular accumulation of a red pigment, providing a visual screen for the presence or absence of the plasmid. Because of redundancy between the two ARF genes, the ARF1-ADE3 plasmid can be lost from the parental strain during cell division without consequence to growth, so colonies exhibit a sectoring (red and white) phenotype. However, the swa4 mutant cannot survive without the ARF1-ADE3 plasmid, so the mutant exhibits a solid red, nonsectoring phenotype (3). The nonsectoring swa4-2 mutant was transformed with a genomic library carried on a centromere-based (single copy) plasmid. Three colonies with a plasmid-linked sectoring phenotype were isolated, the library plasmids were rescued and the ends of the genomic inserts were sequenced. Two of the genomic clones (pCZ01 and pCZ02) carried overlapping fragments from the right arm of chromosome III, with CDC50 being the only intact gene shared between the two clones (Figure 1A). The genomic insert in the third plasmid (pCZ03) contained the ARF2 gene, which suppresses the arf1 mutation rather than complementing the swa4 mutation in the original nonsectoring strain.
The CDC50 gene was subcloned into single-copy centromere (CEN) and multicopy (2μ) vectors and tested for complementation of swa4-2. Both CDC50 subclones complemented the nonsectoring phenotype of swa4-1 (Figure 1B). To determine if swa4 mutations are alleles of CDC50, we crossed a cdc50Δ mutant with swa4-2 and found that the diploid exhibited the same cold-sensitive growth defect as both parents (unpublished data), indicating swa4-2 and cdc50Δ are part of the same complementation group. In addition, we sequenced CDC50 from swa4-1 and swa4-3 and found cdc50 mutations in each strain that interestingly change conserved proline residues to leucine (Figure 1C). These data indicate that the three swa4 mutations isolated in our arf1 synthetic lethal screen are alleles of CDC50.
Cdc50p chaperones Drs2p to the Golgi
It was previously reported that Cdc50p interacts with Drs2p, and this interaction is required for export of C-terminally green fluorescent protein (GFP)-tagged Drs2p from the ER to the Golgi complex (23). We also found that Cdc50p co-immunoprecipitated with Drs2p (Figure 2A) and that an N-terminally GFP-tagged Drs2p required Cdc50p for export from the ER (Figure 2B). In the latter experiment, GFP-Drs2p was modestly overexpressed from the PRC1 promoter and was only seen in the ER of cdc50Δ cells. The ER localization was indicated by the nuclear envelope fluorescence (arrowhead) and by the fluorescence underlying the plasma membrane (cortical ER). Introduction of CDC50 on a single-copy plasmid into the cdc50Δ strain allowed most of GFP–Drs2p to exit the ER and populate the Golgi (puncta), although some fluorescence was still visible in the ER. Overexpression of Cdc50p allowed complete recovery of GFP–Drs2p Golgi localization (Figure 2B). To be certain that GFP–Drs2p was appropriately localized, it was co-expressed with Sec7p fused to red fluorescent protein (RFP), a marker for the yeast TGN (28,29). Nearly complete overlap was observed for the GFP and RFP signals (Figure 2C), so the GFP tag and modest overexpression do not perturb Drs2p localization to the TGN.
To examine the influence of Cdc50p on endogenous (untagged) Drs2p, lysates of WT and cdc50Δ cells were immunoblotted for Drs2p using polyclonal antibodies raised against the large ATPase loop between transmembrane helices 4 and 5 (6). As shown in Figure 3A in triplicate, expression of Drs2p is markedly reduced in cdc50Δ cells relative to the carboxypeptidase Y (CPY) loading control. Surprisingly, deletion of CDC50 did not lead to increased instability of Drs2p as indicated by the pulse–chase experiment shown in Figure 3B. Again, we found that the expression of Drs2p was reduced, but the protein synthesized during the pulse-labeling period was not appreciably degraded during the 120-min chase period.
cdc50Δ exhibits a defect in the transport and modification of CPY
The cdc50 mutant exhibits a defect in the polarized localization of actin patches after shifting cells to the nonpermissive growth temperature for several hours (25). However, the primary defect observed for drs2Δ is a strong perturbation of late Golgi function that is similar to the defects observed in clathrin mutants. For example, drs2Δ causes delayed transport of CPY to the vacuole and mislocalization of late Golgi enzymes responsible for proteolytic processing of pro-α-factor, a high-molecular-weight precursor of the yeast mating pheromone. These phenotypes become more severe immediately after shifting drs2Δ to nonpermissive growth temperatures (20°C or below) but are readily apparent at permissive growth temperatures (e.g. 30°C) (6,7,19).
Pulse–chase experiments were performed to determine if cdc50Δ exhibits protein transport defects similar to drs2Δ. Carboxypeptidase Y is synthesized at the ER as a high-molecular-weight, core-glycosylated precursor called p1 CPY, which is rapidly transported to the Golgi where additional carbohydrate is added to generate the p2 CPY precursor. p2 CPY is sorted from the late Golgi to the late endosome and is proteolytically processed to the mature form (mCPY) on arrival in the vacuole (30). With WT cells, transport of CPY to the vacuole was nearly complete by the 15-min chase point shown in Figure 4A. However, in cdc50Δ or drs2Δ cells, maturation of CPY was delayed with approximately half of CPY remaining in the Golgi p2 form at the 15-min time period. In addition, cdc50Δ and drs2Δ showed a comparable defect in Golgi-dependent glycosylation of CPY. This can be most easily seen by comparing the 5-min time-points for WT, cdc50Δ and drs2Δ samples (Figure 4A). The p1 and p2 CPY precursors from WT cells were readily resolved by SDS–PAGE, while underglycosylation of p2 CPY in cdc50Δ and drs2Δ prevented a clean separation of p1 and p2 CPY. The mCPY from cdc50Δ was also underglycosylated and migrated slightly faster than WT mCPY, which can be more easily seen in the immunoblot shown in Figure 3A. The Cdc50p homolog Lem3p chaperones a Drs2p-related P-type ATPase (Dnf1p) to the plasma membrane (23). As a specificity control, we also labeled lem3Δ and found that this mutant exhibited WT kinetics of CPY transport and modification. In addition, deletion of LEM3 from the cdc50Δ strain (to produce cdc50Δ lem3Δ) did not exacerbate the CPY transport or glycosylation defect of cdc50Δ (Figure 4A). Most of the increase in mass associated with the p1 to p2 transition is caused by addition of α1,3-mannose by Mnn1p, which is a terminal glycosylation event catalyzed in late Golgi compartments (31–33). Mnn1p requires clathrin for Golgi localization (34); therefore, the CPY glycosylation defect of cdc50Δ is a late Golgi, clathrin dysfunction phenotype, while the delayed transport kinetics could reflect either a Golgi or an endosomal perturbation.
cdc50Δ exhibits a defect in α-factor maturation
The cdc50Δ mutant also exhibited the same defect in pro-α-factor processing observed in drs2Δ cells. α-Factor is synthesized as a high-molecular-weight, core-glycosylated precursor (pro-α-factor) in the ER, and this precursor is heavily glycosylated in the Golgi complex. On arrival in the TGN, pro-α-factor is processed by Kex2p, Kex1p and Ste13p to a 13-amino acid peptide that is secreted from the cell and subsequently degraded by extracellular proteases (35). Transport of pro-α-factor through the secretory pathway is very rapid, and all α-factor forms were present at the end of the 5-min labeling period (0 min of chase; Figure 4B). By 5 min of chase in WT or lem3Δ cells, all the precursors were converted to the mature form (m), which was not recovered well because of the extracellular degradation and weaker affinity of the antibody to the mature peptide. In contrast, cdc50Δ, drs2Δ and cdc50Δ lem3Δ all accumulated the high-molecular-weight pro-α-factor form. Note that the ER form disappeared rapidly in all the strains, indicating that ER to Golgi transport was not perturbed. In addition, pro-α-factor was extensively glycosylated in cdc50Δ and drs2Δ cells, indicating that early Golgi function was not perturbed. In contrast to CPY, most of the carbohydrate mass on pro-α-factor is α1,6- and α1,2-linked mannose added in early Golgi compartments (32,33). These observations, combined with the defect in proteolytic processing of pro-α-factor, indicate a Golgi defect at the permissive growth temperature that is specific to the TGN.
To confirm that cdc50Δ and drs2Δ cells are deficient in producing the mature α-factor peptide, these strains were spotted onto a lawn of MATα cells deficient for an extracellular protease (barrier) that degrades α-factor and are thus supersensitive to the pheromone. α-Factor induces a cell cycle arrest in the lawn cells, yielding a zone of growth inhibition, or halo, around patches of cells that secrete the pheromone. Pro-α-factor is not biologically active, so the size of the halo depends on the amount of mature α-factor secreted. As shown in Figure 4C, drs2Δ (patch 2), cdc50Δ (patches 3 and 4) and drs2Δ cdc50Δ (patch 5) all generated a smaller halo than the WT strain (patch 1) at 30°C. In addition, the drs2Δ cdc50Δ double mutant exhibited a defect similar to that exhibited by the two single mutants. This observation indicates that the effects of drs2Δ and cdc50Δ on the Golgi are not additive and support the view that Drs2p and Cdc50p are both required for a single activity (presumably phospholipid translocation) at the Golgi.
Mislocalization of Kex2p in cdc50Δ
The defect in proteolytic processing of pro-α-factor observed in cdc50Δ suggests that the processing enzymes, such as Kex2p, are mislocalized as is observed in drs2Δ and clathrin mutants. To test this possibility, differential centrifugation of WT and cdc50Δ cell lysates was used to obtain a high-speed pellet enriched for Golgi membranes, which was further fractionated in a sucrose gradient to separate early and late Golgi membranes as previously described (6,36). Each fraction was immunoblotted for Drs2p and assayed for Kex2p and guanosine diphosphatase (GDPase) activity, the latter enzyme being a marker for early Golgi compartments. With WT membranes, Kex2p and Drs2p co-fractionated in the dense fractions of the gradient and can be separated from membranes containing GDPase [(6); Figure 5A]. In contrast, the fractions from cdc50Δ were markedly deficient for Kex2p activity, and the small amount of Drs2p in these fractions no longer showed a peak in fractions 7–11. The shift of GDPase to denser fractions is probably not a specific effect of cdc50 because membranes from WT cells often show this type of fractionation profile for GDPase (unpublished data). These studies indicate that Cdc50p is required for Golgi localization of endogenous (untagged) Drs2p and loss of Cdc50p–Drs2p from the Golgi causes mislocalization of Kex2p.
Mutants with a defect in vesicle formation typically exhibit an expansion of the donor organelle. Both drs2Δ and clathrin mutants accumulated enlarged, cup-shaped Golgi cisternae, which in electron micrographs of thin sections appear as rings or half-moon structures (6). As expected, cdc50Δ also accumulated aberrant membrane structures, visualized by electron microscopy (Figure 6, arrowheads and inset), that are similar to the swollen Golgi cisternae observed in drs2Δ cells. Sixty-one enlarged cisternal structures were counted in 87 cdc50Δ cell sections (0.7 structures/cell section), while only 12 similar structures were found in 119 cell sections from the isogenic WT strain (0.1 structures/cell section). Thus, membrane accumulation in cdc50Δ was readily apparent at 30°C (Figure 6).
Loss of plasma membrane PS asymmetry in cdc50Δ and drs2Δ
The Drs2p–Cdc50p complex is required for translocation of NBD-PS from the lumenal leaflet of the TGN to the cytosolic leaflet (20). If endogenous PS is also translocated, then Drs2p should concentrate PS on the cytosolic leaflet of the TGN and the exocytic vesicles that Drs2p helps generate. This activity should contribute to plasma membrane asymmetry of PS because membrane flows from the TGN to the plasma membrane. In addition, deletion of DNF1 and DNF2 causes a defect in NBD-PS translocation across the plasma membrane (18). Thus, the combined activity of the Drs2p and Dnf proteins should be responsible for establishing and maintaining the asymmetric distribution of PS across the plasma membrane. Therefore, we tested whether drs2Δ and dnfΔ cells inappropriately expose PS on the outer leaflet of the plasma membrane by measuring the binding of annexin V to the cell surface of these mutants. Annexin V is a PS-binding protein and preferentially interacts with membranes containing PS (37).
Spheroplasts were prepared from the yeast strains (Figure 7), incubated with annexin V conjugated with Alexa 488, and the fluorescence intensity of intact, living cells was quantified and normalized to the weak fluorescent signal from WT cells. The amount of annexin V bound to dnf1Δ or dnf2Δ single mutants was indistinguishable from WT cells, but the dnf1,2Δ and dnf1,2,3Δ mutants bound significantly more annexin V than WT (1.5- to 2-fold increase). The drs2Δ mutant exhibited a comparable increase in annexin V binding that was not exacerbated by additional deletion of DNF1 (drs2Δ dnf1Δ). However, the drs2Δ dnf1,2Δ triple mutant bound approximately three times more annexin V than WT cells. To determine whether the weak binding of annexin V to WT cells represented a low level of PS exposure on the outer leaflet or background staining, we also measure annexin V binding to cho1Δ (PS synthase) cells, which are completely devoid of PS. The cho1Δ cells bound slightly more annexin V than WT cells, indicating that the fluorescent signal from these cells was simply background staining and the amount of PS in the plasma membrane outer leaflet of WT cells is below the level of detection by this method. To be certain that the increased annexin V binding to drs2Δ cells was due to PS exposure and not an increase in background staining, we examined drs2 cho1 cells and found that the fluorescent signal dropped to WT (background) levels. Therefore, drs2Δ and dnf1,2Δ mutants expose substantially more PS on the outer leaflet of the plasma membrane than WT cells. We also examined the binding of annexin V to cells carrying a deletion of CDC50 family genes, and as expected, we found a significant increase in PS exposure for cdc50Δ and lem3Δ but not for crf1Δ (Figure 7B).
Exposure of PS on the cell surface could be explained by a loss of PS asymmetry of the plasma membrane or a substantial increase in PS levels without loss of asymmetry. To determine if the drs2Δ or dnf1,2,3Δ mutants were producing more PS than WT cells, these strains were labeled to steady state with 32P, and the amounts of the four major glycerophospholipids were quantified (Figure 7C). The relative amount of PS in drs2Δ cells (8%) was actually lower than that in WT cells (18%). The levels of PS were unchanged in dnf1,2,3Δ, although phosphatidylinositol levels were reduced. Although the reason for these differences in phospholipid composition between WT and mutant cells is unclear, the possibility that increased cell surface exposure of PS was caused by increased PS synthesis can be ruled out.
The data in Figure 7 support a model that P-type ATPases in the Drs2p family have overlapping roles in establishing and/or maintaining PS asymmetry of the plasma membrane. An important caveat to this interpretation is that drs2Δ, cdc50Δ, dnf1,2Δ and presumably lem3Δ also perturb protein trafficking in the exocytic and endocytic pathways. Therefore, it was necessary to determine if loss of PS asymmetry at the plasma membrane was a specific consequence of deleting the translocases or whether this phenotype is a secondary consequence of disrupting protein and membrane trafficking. Of particular relevance, clathrin and ARF mutants were tested for PS exposure because Drs2p–Cdc50p is required for ARF- and clathrin-dependent trafficking from the Golgi complex. Strikingly, chc1Δ, clc1Δ and arf1Δ all exposed PS on the cell surface to a level comparable to drs2Δ (Figure 8). We also examined the endocytosis mutants end3Δ and end4Δ (sla2), and these cells also bound more annexin V than WT, with end4Δ showing comparable staining to drs2Δ. It was later found that the end3Δ strain used in this assay had picked up an extragenic suppressor that partially suppressed the endocytosis defect (38), which is why it bound less annexin V than the end4Δ strain. Not all trafficking mutants expose PS as vps mutants that perturb protein transport through the late endosome (vps4, 5 and 21) were not significantly different from WT (39). The class C vps11Δ mutant has more pleiotropic effects on vacuole transport (40) and exposes slightly more PS than the other vps mutants tested (Figure 8A). These data indicate that disruption of protein trafficking in the exocytic or early endocytic pathways causes a loss of plasma membrane PS asymmetry.
The annexin V studies shown in Figures 7 and 8A were performed with null mutants to allow assessment of the complete loss of function phenotype. However, the chronic nature of these null mutations allows ample time for expression of secondary phenotypes. Therefore, we tested if the acute loss of Drs2p or clathrin function in temperature conditional mutants would allow us to assign priority to their roles in trafficking or membrane asymmetry. First, we tested dnf1,2,3Δ strains harboring either WT DRS2 or two different drs2-ts alleles (drs2-12ts and drs2-31ts), expecting to see an increase in PS exposure on inactivation of Drs2p (at least as much annexin V binding as the drs2Δ dnf1,2Δ cells). However, the amount of PS on the cell surface of these cells was no different before and after a 1-h incubation at the nonpermissive temperature (37°C; Figure 8B). This time is sufficient to inactivate Drs2-ts function in vesicle biogenesis in vivo and in NBD-PS translocation in vitro (7,20). The same experiment was performed with a drs2-ts strain carrying WT copies of the DNF genes, and these cells failed to show an increase in annexin V binding after a 1-h shift to the nonpermissive temperature. Moreover, the chc1-ts mutant also failed to expose PS after shifting cells to the nonpermissive temperature for 1 h. Protein trafficking at the TGN is perturbed within 5–30 min of temperature shift for this clathrin-ts mutant (41,42). The temperature shift does not suppress PS exposure on drs2Δ or clc1Δ cells (37°C; Figure 8C), so it is unlikely that the temperature shift somehow suppressed PS exposure on the drs2-ts or chc1-ts cells. We conclude that the onset of protein trafficking defects in drs2-ts and chc1-ts substantially precedes a measurable defect in membrane asymmetry. Therefore, loss of PS asymmetry of the plasma membrane of drs2Δ and chc1Δ cells may be a secondary consequence of disrupting protein trafficking.
During the course of these studies, a cyclic lipopeptide called papuamide B (Pap B), derived from the sea sponge Theonella mirabilis (43), was found to specifically permeabilize membranes containing PS (44). drs2Δ strains are hypersensitive to Pap B, while WT strains are relatively resistant, indicating that PS must be exposed on the outer leaflet of the plasma membrane in order for Pap B to exert its antifungal activity. Deletion of the PS synthase gene (CHO1) eliminates PS and renders yeast strongly resistant to Pap B (44). Consistent with the annexin V binding studies, drs2Δ, cdc50Δ, end3Δ and end4Δ strains are hypersensitive to Pap B, while cho1Δ and drs2Δ cho1Δ strains are resistant (Figure 9A). Even the modest PS exposure of the partially suppressed end3Δ strain used for the annexin V experiments in Figure 8A led to Pap B hypersensitivity (end3Δ-old; Figure 9A). An end3Δ strain derived from backcrossing to remove the extragenic mutation was more sensitive to Pap B, indicating an increase in PS exposure (end3Δ-new; Figure 9A). Using a cell viability assay (see Materials and Methods), we found that treatment of drs2Δ cells with 1 μg/mL Pap B killed 95% of the cells within 1 h. In contrast, the drs2-ts strain (ZHY615M2D-3A) showed no loss of viability during 4 h of incubation with 1 μg/mL Pap B at 37°C (data not shown). Therefore, acute inactivation of Drs2-ts does not lead to exposure of PS on the extracellular leaflet of the plasma membrane.
The cdc50Δ, lem3Δ (ros3Δ), drs2Δ and dnf1,2Δ mutants were previously shown to be hypersensitive to another cyclic antifungal peptide, Ro09-0198 (Ro), that targets PE exposed on the outer leaflet of the plasma membrane (18,27). While the drs2Δ cho1Δ strain is resistant to Pap B, it is more sensitive to Ro than either the single mutant or the WT strain (Figure 9B). This result indicates that cho1Δ confers resistance to Pap B by eliminating its target (PS) rather than restoring plasma membrane integrity to the drs2Δ cho1Δ strain, further demonstrating the PS specificity of Pap B. It is not known if Ro sensitivity is a specific consequence of deleting the potential PE translocases or a secondary consequence of perturbing protein trafficking. Figure 9C shows that endocytosis mutants (end3Δ-new and end4Δ) exhibit a comparable sensitivity to Ro as the dnf1,2Δ strain. The end3Δ-old strain, harboring the extragenic suppressor, is also partially suppressed for the Ro hypersensitivity phenotype. These data indicate that disruption of endocytosis causes a loss of plasma membrane asymmetry and exposure of PS and PE on the outer leaflet.
Our screen for mutations that are synthetically lethal with arf1 recovered three swa4 mutations, which we show here are alleles of CDC50. Thus, cdc50(swa4) and drs2(swa3) share the same synthetic lethal relationship with arf1. This is not surprising as Drs2p and Cdc50p interact to form what appears to be an αβ heterodimer, and Cdc50p is required for transport of Drs2p from the ER to the Golgi complex [this study and Saito et al. (23)]. Therefore, one would expect that cdc50Δ would exhibit the same phenotypes as drs2Δ if Golgi residence is essential for Drs2p function. In fact, we find that cdc50Δ exhibits late Golgi defects that are nearly identical to drs2Δ. These phenotypes include delayed vacuolar transport and reduced glycosylation of CPY, mislocalization of Kex2p from the TGN with the resulting failure to proteolytically process pro-α-factor to its mature form and the accumulation of aberrant membrane structures that are likely expanded TGN cisternae. These are also phenotypes exhibited by the clathrin-ts (chc1-5/swa5-1) and auxilin (swa2) mutants isolated in the swa screen (3), suggesting that perturbed clathrin function is the root cause of these drs2/cdc50 phenotypes. The effects of clathrin mutations on CPY transport are complex (39), so it is not clear if the delayed transport observed for drs2Δ or cdc50Δ reflects a misrouting of CPY to the cell surface followed by endocytosis to the vacuole, as suggested for the chc1Δ strain (45) or simply a kinetic delay in transport from the TGN through the late endosome to the vacuole. Importantly, all these cdc50 phenotypes are observed at a permissive growth temperature of 30°C (Figures 3–6) where the polarized localization of actin cortical patches is normal (25). Similarly, we have not seen a significant perturbation of actin patch localization in drs2Δ cells grown at the permissive temperature (Z. Hua and T.R. Graham, unpublished data).
Deletion of CDC50 can cause depolarization of actin patches, disruption of actin cables and mislocalization of cell polarity determinants such as Bni1p and Gic1p. However, these phenotypes are only observed after shifting cdc50Δ to the nonpermissive growth temperature of 18°C for 12 h. The loss of Drs2p function in cdc50 should have another important consequence after shifting cells to low temperature for several hours, which is a defect in ribosome synthesis (hence the name drs2) and a loss of new protein synthesis (46). This phenotype does not indicate a direct role for Drs2p–Cdc50p in ribosome assembly because many mutants that perturb protein transport in the secretory and endocytic pathways cause the same defect in ribosome synthesis as drs2Δ (47,48). Disrupting plasma membrane and cell wall growth activates the protein kinase C (PKC) cell integrity pathway, which shuts down transcription of ribosomal protein, ribosomal RNA and transfer RNA genes (49,50). In addition, translation initiation is rapidly attenuated through a PKC-independent pathway when secretory function is disrupted (51). Protein synthesis starts to diminish approximately 3 h after shifting drs2Δ cells to low temperature (unpublished data), so it is possible that the cell polarity defect observed in cdc50Δ is caused by the combined loss of Drs2p and short-lived proteins during the long incubation period at 18°C. Moreover, delivery of polarity determinants to the bud membrane by exocytic vesicles reinforces cell polarity cues (52–54), so the defects in exocytic vesicle budding should affect cell polarity. Because the Golgi defects reported here for cdc50Δ are observed at any temperature and precede the appearance of cell polarity defects, it is likely that the cell polarity defects are a secondary consequence of disrupting Golgi function.
Asymmetry of the plasma membrane is a near universal feature of eukaryotic cells. Wild-type yeast cells are no exception, and it appears that most of the PS and PE are restricted to the cytosolic leaflet of the plasma membrane (15,55,56). Phosphatidylserine represents approximately 34% of the plasma membrane glycerophospholipid (57); yet, we could not detect the presence of PS on the plasma membrane outer leaflet of WT cells using annexin V-Alexa 488, suggesting that more than half of the cytosolic leaflet is composed of PS. Marx et al. reported that drs2Δ and end4 cells did not expose more PS on the cell surface than WT cells using annexin V-fluorescein isothiocyanate as a probe (15). We also failed to detect a difference between mutant and WT strains with this probe. However, using the more sensitive annexin V-Alexa 488, we found that drs2Δ, cdc50Δ, dnf1,2Δ and lem3Δ cells expose a significant amount of PS on the outer leaflet, with the drs2Δ dnf1,2Δ triple mutant exposing the highest level of PS. Moreover, mutant cells that bind more annexin V are also hypersensitive to Pap B, which requires PS on the external leaflet of the plasma membrane to kill cells. At face value, these data support the proposed flippase activity for Drs2p–Cdc50p and Dnf1,2p–Lem3p and suggest these enzymes pump endogenous PS to the cytosolic leaflet to generate an asymmetric membrane.
However, drs2Δ and dnf1,2Δ disrupt protein trafficking, so it was important to determine if loss of membrane asymmetry is a specific consequence of deleting P4-ATPases or if other mutants with disrupted protein trafficking pathways also expose PS and PE. In fact, we find that clathrin and endocytosis null mutants exhibit a loss of membrane asymmetry comparable to the drs2Δ or dnf1,2Δ cells. Loss of plasma membrane PS and PE asymmetry in end4Δ was surprising because Drs2p and Dnf1p cycle between the endocytic and the exocytic pathways, and the end mutants exhibit a substantial accumulation of these P4-ATPases at the plasma membrane (23,38). Also surprising was the observation that acute inactivation of drs2-ts did not cause a loss of PS plasma membrane asymmetry within 1–4 h of shifting to the nonpermissive temperature. This time is sufficient to inactivate NBD-PS translocase activity in purified drs2-ts Golgi membranes (20), and the drs2-ts mutant exhibits a loss of transport vesicle biogenesis in vivo within 30 min after shifting to the nonpermissive temperature (7). Likewise, the clathrin-ts mutant fails to expose PS after acute inactivation although protein transport defects are apparent within 5 min of temperature shift (42), and PS is readily detected on the surface of chc1Δ and clc1Δ null cells. Therefore, defects in protein transport precede a measurable loss of plasma membrane asymmetry in drs2-ts and chc1-ts mutants. We conclude that loss of PS plasma membrane asymmetry in these cells is a secondary consequence of a chronic disruption in protein and membrane trafficking to and from the plasma membrane. In fact, a genome-wide screen for Pap B or Ro hypersensitivity identified several hundred gene deletions, many of which perturb protein trafficking, that cause exposure of some PS or PE on the cell surface (44).
These data do not rule out the possibility that Drs2p and Dnf1,2p are flippases that translocate endogenous PS and/or PE to establish membrane asymmetry but rather emphasize the need for caution in interpreting data obtained with null mutants with pleiotropic phenotypes. P4-ATPases remain the best candidates for catalyzing transbilayer movement of phospholipid to the cytosolic leaflet, so how can we reconcile the data presented here with this model for P4-ATPase function? For example, it is possible that membrane asymmetry decays very slowly on inactivation of the PS translocase, particularly in cells containing the Dnf ATPases, so the lack of an immediate effect of drs2-ts inactivation on membrane asymmetry does not disprove a direct role for Drs2p in this process. Unfortunately, conditional alleles of DNF1 and DNF2 have not been generated, so it is not possible to acutely inactivate this entire group of P4-ATPases at the same time. Moreover, it is possible that residual activity of the Drs2-ts against endogenous substrate at 37°C may be insufficient to maintain vesicle biogenesis but sufficient to maintain phospholipid asymmetry. The drs2-ts mutant can grow on Pap B plates at 37°C, while drs2Δ cannot (data not shown), so Drs2-ts does retain residual activity at this temperature. Drs2p is primarily responsible for an NBD-PS translocase activity found in TGN membranes and in low-density secretory vesicles, and this ATPase is trapped at the plasma membrane of endocytosis mutants (20,21,23,38). Yet, disruption of DRS2 in the end4 background does not perturb NBD-PS uptake across the plasma membrane (15). These observations suggest that Drs2p is inactive on the plasma membrane, and therefore, the endocytosis mutants would deplete the Golgi of phospholipid translocase activity without a compensatory increase at the plasma membrane. Appropriate trafficking and/or localization of the phospholipid translocases appears to be critically important for their activity, as many mutants that perturb protein transport exhibit a loss of membrane asymmetry (58). Therefore, it is possible that Drs2p and Dnf1,2p are phospholipid translocases and that loss of membrane asymmetry in end4Δ, arf1Δ and clathrin mutants is a consequence of disrupting their normal trafficking patterns. While available evidence strongly supports an NBD-phospholipid translocase function for P4-ATPases in association with Cdc50p or Lem3p (18,20,23,26,27), their endogenous substrate preferences and role in plasma membrane phospholipid asymmetry remain uncertain.
The pleiotropic effects of mutations in P4-ATPases are medically significant, so it is important to define consequences of both chronic and acute loss of function for these enzymes. Mutations in FIC1 (ATP8B1) cause hereditary intrahepatic cholestasis, an impairment in bile flow through the liver, which in its most severe form causes liver failure (13,59). In addition, deletion of the mouse ATP10C gene (also called ATP10A) causes diet-induced obesity and a defect in insulin-stimulated glucose uptake into muscle and adipocytes (similar to type 2 diabetes in humans) (60,61). Any of the defects described for the yeast P4-ATPase mutants could contribute to these disease phenotypes. For example, defects in the regulated trafficking of the GLUT4 glucose transporter in Atp10c mice could cause the insulin resistance. The bile salt export pump (ABCB11) and PC transporter (ABCB4) require polarized exocytosis to the bile canalicular membrane to function (62), so defects in cell polarity or protein transport processes could lead to cholestasis. Perhaps relevant to familial intrahepatic cholestasis, drs2Δ mutants exhibit a defect in cholesterol uptake at the plasma membrane that can be attributed to mislocalization of ABC transporters more directly involved in cholesterol (or ergosterol) transport (63). Loss of plasma membrane phospholipid asymmetry could potentially perturb the function of any of these integral membrane transporters. Exposure of PS on the cell surface is an early marker of apoptosis and tags the dying cell for recognition and removal by phagocytes (64). Aberrant exposure of PS could lead to tissue damage if viable cells are mistakenly targeted for phagocytosis. In addition, the regulated exposure of PS on activated red blood cells and platelets stimulates the clotting cascade (65). Failure to expose PS on activated platelets causes a bleeding disorder called Scott syndrome and aberrant PS exposure could contribute to a hypercoagulable state and cardiovascular disease (66). Studies in the experimentally tractable yeast system should provide insights into pathologies associated with P4-ATPases.
Materials and methods
Strains and media
Yeast were grown in standard rich medium (yeast extract, peptone and dextrose (YPD)) or synthetic defined minimal medium containing required supplements (67). Strains used are listed in Table 1, and unless otherwise indicated, the Δ specifies Kanr replacements (Δ::KanMX). The cdc50Δ::HIS3 disruptions were generated by polymerase chain reaction (PCR)-mediated gene disruption using pFA6a-His3MX6 as the template as previously described (68). The 13MYC and 3HA tags were chromosomally integrated by the PCR method using pFA6a-13Myc-KanMX6 and pFA6a-3HA-HisMX6 as the templates (68). Tagged forms of Drs2p and Cdc50p were functional based on WT growth of these strains at 20°C. The drs2Δ::LEU2 disruptions were generated using pZH523 as previously described (19).
|BY4742||MATα his3 leu2 ura3 lys2||Research Genetics|
|BY4741||MATα his3 leu2 ura3 met15||Research Genetics|
|BY4742 YCR094W||MATα his3 leu2 ura3 lys2 cdc50Δ||Research Genetics|
|SCY150||MATα his3 leu2 ura3 lys2 cdc50Δ::HIS3||This study|
|SCY118||MATα his3 leu2 ura3 lys2 lem3Δ||This study|
|SCY126||MATα his3 leu2 ura3 cdc50Δ::HIS3 lem3Δ||This study|
|SCY250||MATα his3 leu2 ura3 lys2 drs2Δ cdc50Δ::HIS3||This study|
|KLY2501||MATα his3 leu2 ura3 lys2 drs2Δ::LEU2 CDC50::3XHA||This study|
|KLY2502||MATα his3 leu2 ura3 lys2 DRS2::13MYC CDC50::3XHA||This study|
|BY4742 YER166W||MATα his3 leu2 ura3 lys2 dnf1Δ||Research Genetics|
|BY4742 YDR093W||MATα his3 leu2 ura3 lys2 dnf2Δ||Research Genetics|
|PFY3275F||MATα his3 leu2 ura3 met15 dnf1Δ dnf2Δ||Hua et al. (19)|
|PFY3273A||MATα his3 leu2 ura3 met15 dnf1Δ dnf2Δ dnf3Δ||Hua et al. (19)|
|ZHY615M2D||MATα his3 leu2 ura3 lys2 drs2Δ||Hua et al. (19)|
|ZHY2149D||MATα his3 leu2 ura3 drs2Δ dnf1Δ||Hua et al. (19)|
|ZHY709||MATα his3 leu2 ura3 met15 dnf1Δ dnf2Δ drs2Δ::LEU2||Hua et al. (19)|
|JWY2102||MATα his3 leu2 ura3 lys2 cho1Δ||This study|
|JWY2203||MATα his3 leu2 ura3 lys2 cho1Δ drs2Δ::LEU2||This study|
|ZHY409||MATα leu2 ura3 met15 dnf1Δ dnf2Δ dnf3Δ drs2Δ::LEU2 pRS313-DRS2||Gall et al. (7)|
|ZHY410-1B||MATα leu2 ura3 met15 dnf1Δ dnf2Δ dnf3Δ drs2Δ::LEU2 pRS313-drs2-12ts||Gall et al. (7)|
|ZHY410-3A||MATα leu2 ura3 met15 dnf1Δ dnf2Δ dnf3Δ drs2Δ::LEU2 pRS313-drs2-31ts||Gall et al. (7)|
|ZHY615M2D-3A||MATα his3 leu2 ura3 lys2 drs2Δ pRS313-drs2-31ts||This study|
|CCY2811||MATα swa4-2 leu2 ura3-52 his3 trp1-Δ901 ade2 ade3 arf1Δ::HIS3 pCC8218 (YCp50 ADE3-URA3-ARF1)||Chen and Graham (3)|
|RC634||MATα sst1-3 rme ade1-2 ura1 his6 met1 can1 cyh2 GAL||Chan (73)|
|BY4742 chc1-521||BY4742 chc1-521::URA3||This study|
|BY4742 YPR173C||MATα his3 leu2 ura3 lys2 vps4Δ||Research Genetics|
|BY4742 YOR069W||MATα his3 leu2 ura3 lys2 vps5Δ||Research Genetics|
|BY4742 YMR231W||MATα his3 leu2 ura3 lys2 vps11Δ||Research Genetics|
|BY4742 YOR089C||MATα his3 leu2 ura3 lys2 vps21Δ||Research Genetics|
|BY4742 YNL084C||MATα his3 leu2 ura3 lys2 end3Δ||Research Genetics|
|BY4742 YNL243W||MATα his3 leu2 ura3 lys2 end4Δ||Research Genetics|
|BY4742 YGL206C||MATα his3 leu2 ura3 lys2 chc1Δ||Research Genetics|
|BY4742 YGR176W||MATα his3 leu2 ura3 lys2 clc1Δ||Research Genetics|
|SEY6210||MATα leu2-3,112 ura3-52 his3-Δ200 trp1-Δ900 lys2-801 suc2-Δ9||Robinson et al. (74)|
|LSY93.1-10A||MATα leu2-3,112 ura3-52 trp1 suc2 clc1Δ::HIS3||Chu et al. (75)|
|6210 drs2Δ||SEY6210 drs2Δ::TRP1||Chen et al. (6)|
|6210 chc1-521||SEY6210 chc1-521::URA3||Chen and Graham (3)|
|TGY1912||6210 end4-1||Gall et al. (7)|
|6210 arf1Δ||SEY6210 arf1Δ::HIS3||Gaynor et al. (76)|
|KLY1101||SEY6210 SEC7::DsRed.T4||This study|
|KLY201||MATα his3 leu2 ura3 lys2 end3Δ||Liu et al. (38)|
Cloning of SWA4
The swa4-2 strain CCY2811 was transformed with a centromere-based genomic library (69) and plated on minimal media (Leu−) containing 4 μg/mL adenine. The sectoring phenotype of the transformed colonies was scored after at least 1 week of growth at 30°C. Sectoring transformants were replated to confirm the phenotype, and the library plasmids were rescued from these strains. Three of the rescued library plasmids (pCZ01–pCZ03) were able to confer the sectoring phenotype on retransformation into CCY2811. Partial sequencing mapped the genomic inserts to the segments of chromosome III shown in Figure 1. The CDC50 gene was recovered from pCZ01 on a SnaBI fragment and subcloned into SmaI digested pRS416, pRS426, pRS315 and pRS425 (70). CDC50 was amplified from swa4-1 and swa4-3 using primers 5′-GTCCATGACGCACTGCGAAC-3′ and 5′-CACAAATACCTACAGGCAC-3′. The PCR product was sequenced directly and after subcloning into pGEM-T easy (Promega, Madison, WI, USA). The swa4-1 and swa4-3 mutations are C→T transitions in the second position of the codons 252 and 307, respectively. For expression of the GFP–Drs2p fusion protein, a 1.3 kb SpeI-ClaI fragment from pGOGFP, consisting of the PRC1 promoter and GFP(S65T) (71), was inserted into pRS416 to generate pRS416–GFP. The plasmid pRS315–DRS2 was used with primers SalI-Drs2-F and Drs2-SalI-R as a PCR template to generate a 3.7 kb fragment that placed SalI sites at both the start and the end of the DRS2 coding region. This fragment was subcloned into SalI site of pRS416-GFP, creating the plasmid pGFP–DRS2m. To eliminate mutations introduced by PCR, an AgeI-ClaI fragment from pRS315-DRS2 was used to replace the AgeI-ClaI fragment in pGFP–DRS2m to generate pGFP–DRS2.
Labeling and immunological methods
Cell labeling, immunoprecipitation and immunoblotting were performed as previously described (6,9) using anti-α-factor (33), anti-CPY (Gift of Scott Emr, University of California, San Diego), anti-Drs2p (6), anti-hemagglutinin (HA-7; Sigma, St Louis, MO, USA) and anti-Myc (9E10; Sigma). Co-immunoprecipitation of Drs2p and Cdc50p was performed essentially as described (23), except that 10 μL of rabbit anti-Drs2 serum was used for the primary immunoprecipitation, and these pellets were heated to 50°C for 5 min in 30 μL of SDS/urea sample buffer [40 mm Tris–HCl (pH 6.8), 8 m urea, 0.1 mm ethylenediaminetetraacetic acid, 1% 2-mercaptoethanol, 5% SDS and 0.25% bromophenol blue] before electrophoresis.
Annexin V assay
Cell surface exposure of endogenous PS was determined using annexin V conjugated with Alexa Fluor 488 (Molecular Probes, Eugene, OR, USA). Yeast were grown in YPD, harvested while in log phase (∼0.5 OD600/mL), and then washed once in 50 mm Tris–Cl, pH 7.4, and resuspended at 50 OD600/mL in 50 mm Tris–Cl, pH 7.4, containing 1.4 m sorbitol. They were then converted to spheroplasts during a 45-min incubation with 1 μL of Zymolyase-100T (10 mg/mL in water; Seikagaku Corp. Tokyo, Japan) per 10 OD600 units of cells. Spheroplasts were resuspended at 1 OD600 cells/mL in binding buffer (10 mm Na HEPES, pH 7.4, 140 mm NaCl and 2.5 mm CaCl2) with 1.4 m sorbitol. Annexin V solution (5 μL) was added to 195 μL spheroplast suspension, and the mixture was incubated for 10 min on ice in the dark. Cells were then washed two times and suspended in binding buffer with 1.4 m sorbitol. The spheroplasts were viewed using a fluorescence microscope (Carl Zeiss, Thornwood, NY, USA), and the images (differential interference contrast and fluorescence) were captured with a cooled Correct change device camera (Hamamatsu C4880, Hamamatsu City, Japan) using the same exposure time and gain settings for mutant and WT cells. metamorph software (Universal Imaging, Downingtown, PA, USA) was used to draw circles with defined dimensions around 30–40 living cells per experimental condition and to quantify the total fluorescence intensity for each cell (15). Dead cells, identified by their change in refractive properties in DIC images, bound 8- to 10-fold more annexin V than living WT cells and were excluded from the analysis. An average fluorescent intensity was determined for each mutant and was divided by the average fluorescent intensity of the WT cells in that experiment. The data reported are the mean of 3–9 separate experiments ± standard deviation between experiments and are described as annexin V binding relative to WT cells. By definition, WT cells have a ratio of one for each experiment, so there is no standard deviation for the WT cells used to normalize the data. Similar data were also obtained by flow cytometry, although clumping of spheroplasts made this technique less reliable. Similar data were also obtained using cells depleted for ATP by treatment with sodium azide and fluoride, suggesting that the annexin V is not trapping PS actively pumped to the outer leaflet during the incubation with annexin V on ice but rather measures the steady-state concentration of PS in the outer leaflet.
The viability of cells treated with Pap B was determined by incubating strains BY4742 (WT), ZHY615M2D (drs2Δ) and ZHY615M2D-3A (drs2-ts) at 0.5 OD/mL in YPD with or without 1 μg/mL Pap B at 27°C. Half of each culture was shifted to 37°C, and aliquots were removed 0, 1, 2 and 4 h later, then diluted 1000-fold and plated on rich medium. After 3 days of growth at 27°C, the number of colonies was counted, and the per cent viability was calculated by dividing colony numbers from the Pap B-treated samples by the corresponding untreated samples.
We thank the electron microscopy resource center at Vanderbilt for their assistance, Masato Umeda for providing the Ro peptide and Ben Glick for the providing Sec7-RFP construct. This work was supported by a grant from the National Institutes of Health to T. R. G. (GM62367).
- 9Metabolic labeling and immunoprecipitation of yeast proteins. In: Bonifacino JS, Dasso M, Harford JB, Lippincott-Schwartz J, Yamada K, editors. Current Protocols in Cell Biology. New York: John Wiley & Sons, Inc.; 2000, pp. 7.6.1–7.6.9..
- 38Yeast P4-ATPases Drs2p and Dnf1p are essential cargos of the NPFXD/Sla1p endocytic pathway. Submitted for publication. 2006., , , .
- 39Protein Transport to the Yeast Vacuole. Amsterdam: Academic Press; 2002., .