The C2 domain of phosphatidylserine decarboxylase 2 is not required for catalysis but is essential for in vivo function.

Phosphatidylserine decarboxylase 2 (Psd2p) is currently being used to study lipid trafficking processes in intact and permeabilized yeast cells. The Psd2p contains a C2 homology domain and a putative Golgi retention/localization (GR) domain. C2 domains play important functions in membrane binding and docking reactions involving phospholipids and proteins. We constructed a C2 domain deletion variant (C2Delta) and a GR deletion variant (GRDelta) of Psd2p and examined their effects on in vivo function and catalysis. Immunoblotting confirmed that the predicted immature and mature forms of Psd2(C2Delta)p, Psd2(GRDelta)p, and wild type Psd2p were produced in vivo and that the proteins localized normally. Enzymology revealed that the Psd2(C2Delta)p and Psd2(GRDelta)p were catalytically active and could readily be expressed at levels 10-fold higher than endogenous Psd2p. Both Psd2p and Psd2(GRDelta)p expression complemented the growth defect of psd1Deltapsd2Delta strains and resulted in normal aminoglycerophospholipid metabolism. In contrast, the Psd2(C2Delta)p failed to complement psd1Deltapsd2Delta strains, and [(3)H]serine labeling revealed a severe defect in the formation of PtdEtn in both intact and permeabilized cells, indicative of disruption of lipid trafficking. These findings identify an essential, non-catalytic function of the C2 domain of Psd2p and raise the possibility that it plays a direct role in membrane docking and/or PtdSer transport to the enzyme.

The aminoglycerophospholipids, phosphatidylserine (Ptd-Ser), 1 phosphatidylethanolamine (PtdEtn), and phosphatidylcholine (PtdCho) are important components of membranes in both prokaryotes and eukaryotes that collectively can comprise more than 70% of the phospholipids in many cells (1,2). In many species PtdSer is decarboxylated to form PtdEtn that undergoes three methyl additions to its primary amine to form PtdCho. However, in mammalian systems the methyltransferase reactions that give rise to PtdCho are restricted to the liver (3,4). Eukaryotes are also able to use choline and ethanolamine for the synthesis of PtdCho and PtdEtn primarily via pathways elucidated by Kennedy and co-workers (5,6). The yeast Saccharomyces cerevisiae can make the majority of its aminoglycerophospholipids from either a PtdSer, an Etn, or a Cho precursor (7). An intriguing aspect of aminoglycerophospholipid metabolism in eukaryotes is the segregation of the synthetic enzymes among different organelles (8). In yeast PtdSer is synthesized in the endoplasmic reticulum (ER) or a related membrane associated with the mitochondria (MAM) (9,10). This nascent PtdSer must be transported to the mitochondria or the Golgi/vacuole to be converted to PtdEtn (11,12). The mitochondrial PtdSer decarboxylase is denoted Psd1p, and the Golgi/vacuole enzyme is denoted Psd2p.
Both Psd1p and Psd2p have been extremely useful as tools for monitoring the interorganelle traffic of nascent PtdSer (13). Recent work from this laboratory has used Psd2p to characterize some of the parameters required for interorganelle transfer of PtdSer and to identify additional genes/proteins required for the process. One protein that is required for movement of nascent PtdSer to Psd2p is a phosphatidylinositol (PtdIns)binding protein we named PstB2p (phosphatidylserine transport B2) whose encoding gene (PSTB2) was previously identified in a pleiotropic drug resistance screen and is named PDR17 (14). This same gene has also been identified as a homolog of the yeast PtdIns/PtdCho transfer protein, Sec14p, and is alternatively named SFH4 (15). A second protein that has been identified in genetic screens for strains exhibiting aminoglycerophospholipid transport defects is a PtdIns 4-kinase, Stt4p (16).
Our recent work has led us to consider the possibility that a protein complex may form on the donor and/or acceptor membranes involved in interorganelle lipid transfer processes. Some support for this general idea comes from experiments demonstrating that PstB2p must be present on acceptor membranes for PtdSer to be transferred to Psd2p (17).
In this report we addressed the role of structural domains of Psd2p in the localization, catalytic activity, and in vivo function of the protein. The Psd2p has a domain with homology to the trans-Golgi protein Kex2p, which has been implicated in the localization of this protein (18). However, the function of this domain in Psd2p is not clear. The Psd2p also contains a C2 domain (12). The archetypical C2 domains were first identified in protein kinase C isoforms and constitute Ca 2ϩ and phospholipid binding motifs (19). Numerous other proteins involved in signal transduction and membrane trafficking possess C2 domains (20 -22). In some cases C2 domains bind lipids or membranes in Ca 2ϩ independent reactions. C2 domains are also known to promote both protein-protein and protein-phosphoinositide interactions. The role of the C2 domain in Psd2p function has not previously been examined and has remained unclear.
The purpose of this study was to perform structure/function analyses of Psd2p using biochemistry and genetics. In this report we sought to do the following: 1) define the processing of Psd2p required for its maturation, 2) determine the role of the GR domain in processing, catalysis, targeting, and in vivo function of Psd2p, and 3) determine the role of the C2 domain in the processing, catalysis, targeting, and in vivo function of Psd2p. Our findings indicate that the GR domain is dispensable for all functions examined. In contrast, the C2 domain is not required for processing, catalysis, or targeting but is essential for in vivo function.
Construction of Vector to Express Psd2(GR⌬)p or Psd2(C2⌬)p in Yeast Cells-YEp352-PSD2/GR⌬ and YEP352-PSD2/C2⌬, yeast expression vectors containing a 3Ј-linked HA-epitope tag sequence, were constructed by the overlap extension polymerase chain reaction (PCR) method (25). Briefly, genomic DNA from a SEY6210 yeast strain and PCR primers specifically designed to delete the GR or C2 region in the PSD2 gene were mixed and amplified with a thermal cycler by using Pfu DNA polymerase (Stratagene). The PSD2/GR⌬ or PSD2/C2⌬ DNA PCR products containing PstI restriction enzyme sites located 5Ј and 3Ј to the sequence of interest were purified by agarose gel electrophoresis after PstI digestion. The YEp352-PSD2 yeast expression vector constructed in a previous study (12) was also digested with PstI, and the resultant DNA fragments were purified by electrophoresis. Next, the appropriate pieces of DNA were ligated to yield YEP352-PSD2/GR⌬ and YEp352-PSD2/C2⌬. The desired plasmids were cloned in Escherichia coli (XL-1 Blue, Stratagene) and purified. The sequences of the final constructs were verified by automated DNA sequencing.
Western Blot Analysis for Psd2p Variants-The PTY44 (psd1⌬psd2⌬) strain was the recipient for the YEp352-PSD2, YEp352PSD2/GR⌬ and YEp352-PSD2/C2⌬ constructs. The PTY44 cells were transformed with the vectors using the LiAc method described in the Yeastmaker TM Transformation System (CLONTECH, CA). After the selection on SCE minus uracil (SCE -U) plates, a positive clone for each variant was incubated with SCE -U liquid media at 30°C overnight to give an A 600 of ϳ0.6. The cells were collected by centrifugation, washed with distilled water, and homogenized with glass beads using a Beadbeater (Biospec Products) in a buffer containing 50 mM Tris, 0.25 M sucrose, 0.5 mM EDTA, 10 mM 2-mercaptoethanol, and 0.5 mM phenylmethylsulfonyl fluoride, pH 7.0. Total cell-free extracts were then prepared by centrifugation at 3,000 rpm for 5 min. Each extract, containing 75 g of protein, was used for Western blot analysis to confirm the protein expression level. Electrophoresis was performed with Tris-glycine gels (8 -16%, 1 mm ϫ12 wells, Invitrogen), and the proteins were transferred to nitrocellulose membrane (Bio-Rad). The proteins were detected with a rat anti-HA monoclonal antibody (3F10, Roche), and horseradish peroxidase-conjugated goat anti-rat IgG polyclonal antibodies (Jackson ImmunoResearch Laboratories) in conjunction with ECL reagents (Sigma).
Determination of Psdp2 Enzyme Activity-Psd2p enzyme activity was measured by the 14 CO 2 trapping method with modifications (26). An aliquot of total cell-free extract from each of the different strains, containing 5 g of protein, was used for the enzyme assay. The extracts were mixed with NBD-Ptd[1Ј-14 C]serine (80,000 cpm) in the reaction buffer (25 mM KPO 4 , 10 mM 2-mercaptoethanol, 150 mM KCl, pH 7.0) and incubated at 30°C for 20 min. The 14 CO 2 produced by the decar-boxylation of NBD-Ptd[1Ј-14 C]serine was trapped on filter paper saturated with 100 l of 2 M KOH. The reaction was terminated by acidification with 0.25 M H 2 SO 4 , and the mixture was incubated further at 30°C for 1 h. The 14 C-containing filter paper was placed in 0.5 ml of H 2 O plus 4.5 ml of ScintiSafe TM 30% (Fisher) and incubated at 50°C for 30 min. The radioactivity bound to the filter paper was quantified by liquid scintillation spectrometry.
Determination of Ethanolamine Auxotrophy-The parental PTY44 strain and its derivatives harboring YEp352-PSD2, YEp352-PSD2/ GR⌬, and YEp352-PSD2/C2⌬ plasmids were cultured on plates supplemented with Etn (SCE -U) or without Etn (SC -U). Each of the strains (1 ϫ 10 6 cells) was streaked on one of four areas of the plate and incubated at 30°C for 2.5 days. These strains were also inoculated into liquid SCE -U or SC -U at an initial A 600 of 0.02 and grown at 30°C with shaking. The OD 600 of the culture media was continuously monitored for 72 h to assess the growth in liquid medium.
[ 3 H]Serine Incorporation into Aminoglycerophospholipids-PTY44, PTY41, and PTY44 strains transformed to express wild type and mutant PSD2 genes from high copy plasmids were grown overnight to log phase with SCE or SCE -U. After the cells were washed twice with sterile distilled water they were diluted to 2 ϫ 10 6 /ml in the aforementioned medium containing 20 Ci/ml of L-[3-3 H]serine (32 Ci/mmol) without Etn and non-radioactive serine. The cultures were incubated for 4 h at 30°C with shaking. At the end of the incubation, an aliquot of a 50% trichloroacetic acid solution was added to give a final concentration of 5%. The cells were harvested by centrifugation and washed three times with distilled water. The resultant pellet was subjected to lipid extraction (17). Total lipids were analyzed by thin layer chromatography on silica gel H plates (Uniplate TM , Analtech, Inc., DE) using a solvent system containing chloroform, methanol, 2-propyl alcohol, 0.25% aqueous KCl, and triethylamine (30:9:25:6:18, v/v/v/v/v). Individual phospholipids were identified under ultraviolet light by comigration with authentic standards (PtdSer, PtdEtn, PtdCho, and PtdIns) after the plates were sprayed with 0.2% 8-anilino-1-napthalene sulfonic acid. The lipids were scraped into vials containing scintillation fluid (Scinti-Safe TM 30%), and the radioactivity was quantified by liquid scintillation spectrometry.
Preparation of Permeabilized Yeast Cells-Permeabilized cells were prepared in a lysis buffer (20 mM HEPES, 0.15 M potassium acetate, 2 mM magnesium acetate, 0.5 mM EGTA, 0.4 M sorbitol, pH 6.8) following the method described by Achleitner et al. (31). In brief, yeast spheroplasts were prepared by treating cells with dithiothreitol under alkaline conditions followed by zymolyase treatment. The resulting spheroplasts were regenerated for 20 min at 30°C in the presence of 0.75% yeast extract, 1.5% peptone, 1% glucose, and 0.7 M sorbitol before they were washed and then resuspended in lysis buffer at a concentration of 0.5 g wet weight/ml. The cell suspensions were divided into 0.2-0.3-ml aliquots and frozen over liquid nitrogen vapor for 15 min. The frozen cells could be stored at Ϫ80°C for at least 3 months. The cells were thawed slowly on ice for 1 h prior to use in transport reactions.
PtdSer  30 Ci/ml [ 3 H]serine, pH 8.0, by the same method described previously (17). The final protein concentration of the permeabilized cells in the transport reaction was 0.9 mg/ml. The assay mixture was incubated at 30°C for 100 min. The incorporation of [ 3 H]serine was terminated by the addition of 1 ml of methanol, 0.5 ml CHCl 3 , and 0.3 ml of 0.2 M KCl to generate a monophase. The extraction was completed by the further addition of 0.45 ml of 0.2 M KCl and 0.5 ml CHCl 3 to create two phases that were separated by centrifugation. After aspiration of the upper phase, the lower chloroform phase was further washed with 1.9 ml of methanol/phosphate-buffered saline (1:0.9, v/v) twice. The resulting chloroform phase was dried under nitrogen gas, resuspended in 30 l of chloroform/methanol (9:1, v/v), and analyzed by thin layer chromatography as described above. The ratio of PtdEtn/PtdSer ϩ PtdEtn was used as a transport index for PtdSer.
Fractionation of the Permeabilized Cells-The permeabilized cells were prepared from strains harboring PSD2/GR⌬, PSD2/C2⌬, and PSD2 genes on YEp352 plasmids as described above. The fractionation procedure used methods previously described (14). The reaction mixture was immediately homogenized gently on ice with 15 strokes using the B pestle of a Dounce homogenizer. The cell debris was removed from the homogenate by centrifugation at 1,500 ϫ g for 5 min. The supernatant was further centrifuged at 30,000 ϫ g for 15 min to remove dense membranes. The supernatant was removed from the pellet, overlaid on a two-step gradient consisting of 1 ml of 80% sorbitol and 1 ml of 25% sorbitol, and centrifuged for 2 h at 280,000 ϫ g in a Beckman SW41 rotor. All sorbitol densities are given as the w/v in 10 mM triethanolamine, pH 7.2. The interface between the 25 and 80% sorbitol layers was collected and adjusted to 43% sorbitol using refractometry and layered on a gradient prepared in 40, 43, 60, 70, and 80% increments. The 40% step was 3.6 ml, and the 80% step was 1 ml. All other gradient steps were 2.3 ml each. Membranes were separated by centrifuging the gradient in an SW41 rotor at 280,000 ϫ g for 40 h. Fractions were collected by aspiration from the top of the gradient and stored at Ϫ20°C.

Wild Type Psd2p and the Structural Variants Psd2(C2⌬)p and Psd2(GR⌬)p Are Processed to Yield Mature Enzyme-The
yeast PtdSer decarboxylases contain several distinct structural domains as outlined in Fig. 1. The yeast Psd1p is organized in a similar manner to the mammalian enzyme that has been shown to be sequentially processed to yield mitochondrial outer-and inner-membrane transit intermediates and finally mature ␣and ␤-subunits (32). The generation of ␣and ␤-subunits in the eukaryotic enzymes is predicted to proceed via endoproteolytic cleavage of ␣and ␤-subunits by a serinolysis mechanism that generates the pyruvoyl prosthetic group essential for catalysis. Predictions for eukaryotic processing are derived from definitive work performed with the E. coli enzyme that demonstrates cleavage between Glu and Ser residues residing in an LGST motif (33). The yeast Psd2p contains a GGST motif that is proposed to be the cleavage site for generating ␣and ␤-subunits (12,13). In addition to predicted ␣and ␤-subunits, Psd2p contains within the ␤-subunit a putative Golgi localization/retention motif similar to that found in the Golgi protease Kex2p (18) and a C2 homology domain similar to that found in protein kinase C and synaptotagmin (20,21).
Analysis of C-terminal HA epitope-tagged versions of Psd2p shown in Fig. 2 reveals that the wild type protein is detectable as a 130-kDa precursor that is processed to yield a mature ␣-subunit of ϳ11 kDa. The size of the ␣-subunit is entirely consistent with the structural predictions from sequence data. The Psd2(GR⌬)p and the Psd2(C2⌬)p are identifiable as smaller molecular size precursors of ϳ127 and 117 kDa, respectively. Both the Psd2(GR⌬)p and the Psd2(C2⌬)p are processed to yield the mature ␣-subunit required for catalysis. From these data we conclude that the two structural mutants we have constructed are processed to the mature form of the protein in a process that is essentially indistinguishable from that occurring with the wild type protein. A minor immunoreactive band of slightly higher molecular size is also observed in the region of the ␣-subunit. The exact identity of this component is not known, but based upon the reaction sequence of pyruvoyl enzymes (13), it is probably the ␣-subunit containing PtdSer or PtdEtn in Schiff's base linkage to the active site.

Psd2(GR⌬)p and Psd2(C2⌬)p Are Catalytically Active in
Vitro-To determine whether the Psd2(GR⌬)p and Psd2(C2⌬)p expressed in yeast cells are catalytically active, we performed enzyme assays using cell-free extracts from the different strains. Total cell-free extracts derived from strains lacking PtdSer decarboxylases (psd1⌬ psd2⌬) or expressing only Psd2p from the chromosome (psd1⌬ PSD2) or expressing only high copy Psd2p, Psd2(GR⌬)p or Psd2(C2⌬)p were compared. As shown in Fig. 3, the psd1⌬ psd2⌬ strain had no decarboxylase activity, and the strain expressing Psd2p from the chromosome had the low but measurable activity previously described (12). High level expression of Psd2p gave 15 times the enzyme activity of the control strain (PTY41) expressing only the chromosomal copy. The Psd2(GR⌬)p and Psd2(C2⌬)p showed catalytic activity 12 and 10 times higher than that obtained for the chromosomal copy of the wild type gene. These data indicate that the Psd2(GR⌬)p and the Psd2(C2⌬)p are catalytically active and can be expressed at very high levels within the yeast. These data clearly demonstrate that neither the GR domain nor the C2 domain are required to produce a catalytically functional enzyme.
Psd2(GR⌬)p and Psd2(C2⌬)p Colocalize with Psd2p on Sorbitol Density Gradients-Subcellular fractionation of the permeabilized and homogenized cells was performed by density gradient centrifugation to define the localization of the Psd2(GR⌬)p, Psd2(C2⌬)p, and Psd2p expressed in yeast cells. The yeast strains were permeabilized and fractionated, and the localization of mature and immature Psd2p in fractionated samples was revealed by Western blot with anti-HA antibody. As shown in Fig. 4, the bands corresponding to the mature ␣-subunit of Psd2p, Psd2(GR⌬)p, and Psd2(C2⌬)p were observed from fractions 13 to 18, with the peak of the expression at fraction 16. Within the limits of resolution of this centrifugation method, we conclude that the subcellular localization of the Psd2p variants is identical to the wild type protein. The bands corresponding to immature forms of the decarboxylases were also centered around fraction 16, demonstrating that processing of Psd2p to the mature form is not required before localization. In addition, an 80-kDa band was detected in the peak fractions for both mature and immature Psd2p and LGST amino acid motif that identifies the proteolysis site necessary for formation of ␣and ␤-subunits and generation of the active site. Psd1p also has mitochondrial targeting and inner membrane (IM) sorting sequences. In contrast the Psd2p contains a C2 homology domain, a Golgi localization/retention sequence, and a GGST motif. The GGST motif is predicted to be the cleavage site for ␣and ␤-subunits and generation of the active site. FIG. 2. Psd2p, Psd2(C2⌬)p, and Psd2(GR⌬)p are expressed and proteolytically processed in yeast. The DNA constructs PSD2, PSD2/C2⌬, and PSD2/GR⌬ were inserted into a high copy YEp352 plasmid and engineered to contain a 3Ј sequence encoding a HA-epitope tag. The plasmids were transformed into PTY44 (psd1⌬psd2⌬). Cellfree extracts containing 75 g of protein were used for electrophoresis and Western-blot analysis with anti-HA monoclonal antibody and ECL detection. Ctrl, PTY44 plus YEp352. C2⌬, PTY44 plus YEp352-PSD2/ C2⌬. GR⌬, PTY44 plus YEp352-PSD2/GR⌬. WT, PTY44 plus YEp352-PSD2.
Psd2(C2⌬)p, but not in those of Psd2(GR⌬)p. It is likely that the ϳ80-kDa protein is a proteolytic product of the immature Psd2p and Psd2(C2⌬)p. This ϳ80-kDa band could be a spurious proteolytic product, but it may also be generated from a specific cleavage within the GR domain. The identical distribution of Psd2p, Psd2(GR⌬)p, and Psd2(C2⌬)p strongly suggests that the deletion of either the GR domain or the C2 domain does not affect either the maturation or the localization of Psd2p. Thus, the GR and C2 domains do not participate in targeting the decarboxylase to its subcellular location.
Psd2(C2⌬)p Cannot Fulfill the in Vivo Function of Psd2p-We next examined whether the Psd2(C2⌬)p and Psd2(GR⌬)p could replace the wild type enzyme in vivo. In these studies strains lacking chromosomal copies of PtdSer decarboxylases (psd1⌬psd2⌬) were transformed with high copy plasmids encoding either Psd2p, Psd2(GR⌬)p, or Psd2(C2⌬)p. Under the conditions used, cell growth is absolutely dependent upon the in vivo function of Psd2p when Etn is omitted from the medium. When Etn is included in the medium, cells can bypass the need for Psd2p function by synthesizing PtdEtn by the Kennedy pathway (5, 6, 30). High copy plasmids were chosen for this experiment to bias the results to reveal complementation of the PtdSer decarboxylase null mutant even with weakly active forms of the Psd2p variants lacking either the GR or C2 domains. In Fig. 5, A and B results are presented from experiments conducted with solid and liquid media, respectively. In the presence of Etn (permissive conditions) all strains harboring a plasmid bearing a Psd2p variant were capable of growing on either solid or liquid medium. In the absence of Etn (non-permissive conditions) the Psd2p-and Psd2(GR⌬)p-expressing strains grew equally well on both solid and liquid medium, and the growth was comparable to that found under permissive conditions. In contrast to the other constructs, the Psd2(C2⌬)p variant was unable to support growth on either solid or liquid medium in the absence of Etn. These data clearly demonstrate that the C2 domain of Psd2p plays an essential role in the in vivo function of the enzyme.
Yeast Cells Harboring Psd2(C2⌬)p Fail to Metabolize PtdSer to PtdEtn-The inability of the Psd2(C2⌬)p to support cell growth, despite high level catalytic activity and normal subcellular localization, suggested a defect in the transport-dependent decarboxylation of PtdSer. We next examined the effects of the Psd2p structural variants on aminophospholipid metabolism in vivo. In these experiments, shown in Fig. 6, cells were labeled for 1-2 generations with [ 3 H]serine in the absence of Etn. The majority of [ 3 H]serine is directly incorporated into the head groups of aminophospholipids via the de novo synthesis of PtdSer. In psd1⌬psd2⌬ strains lacking the decarboxylases, PtdSer is synthesized. However, there is little PtdSer turnover, and there is a marked defect in the formation of PtdEtn. The residual PtdEtn formed in psd1⌬psd2⌬ strains occurs via sphingolipid labeling and turnover (26). In strains containing a chromosomal copy of the PSD2 gene, PtdEtn labeling is four times the background labeling level, and there is significantly less PtdSer accumulation compared with the strains lacking PtdSer decarboxylases. In strains expressing the Psd2(GR⌬)p, PtdEtn formation is comparable to strains harboring the wild type chromosomal PSD2 gene. In contrast, strains expressing Psd2(C2⌬)p show the same profound defect in PtdEtn formation observed for the psd1⌬psd2⌬ mutant. For comparison, cells expressing high copy levels of Psd2p display elevated turnover of PtdSer and modestly enhanced production of Pt-dEtn. The PtdCho labeling is also presented in this figure to provide a complete data set for the aminoglycerophospholipids. Much of the labeling of PtdCho is derived from the [ 3 H]serine label entering the one-carbon pool and participating in the methylation of preexisting PtdEtn present within the cell as a consequence of the growth conditions (Etn supplementation) prior to the start of isotope addition. These data clearly dem-onstrate that cells expressing Psd2(C2⌬)p as the only PtdSer decarboxylase have a severe defect in transport-dependent aminoglycerophospholipid metabolism.
Permeabilized Cells Expressing Psd2(C2⌬)p Do Not Transport Nascent PtdSer to the Decarboxylase-Recently, we have developed a versatile system for examining nascent PtdSer transport to the locus of Psd2p in permeabilized cells. We used this system to examine whether cells expressing Psd2(C2⌬)p were capable of transporting PtdSer to the enzyme. Permeabilized cells were labeled with [ 3 H]serine, and the incorporation of the radiolabel into PtdSer and PtdEtn was measured. As shown in Fig. 7A, ϳ10% of the newly formed PtdSer was converted to PtdEtn during a 100-min incubation of permeabilized cells containing a wild type PSD2 gene. In contrast, cells expressing the Psd2(C2⌬)p as the only PtdSer decarboxylase converted less than 2% of the nascent PtdSer to PtdEtn. This difference between wild type and mutant Psd2p was not caused by intrinsic differences in catalytic activity in the permeabilized cells. Measurement of PtdSer decarboxylase activity in the permeabilized cells using the NBD-Ptd[1Ј-14 C]serine substrate that readily partitions into all membranes is shown in Fig. 7B. The cells expressing Psd2(C2⌬)p had demonstrably higher levels of decarboxylase activity than their wild type counterparts. These data clearly demonstrate that the nascent PtdSer generated within the permeabilized cells is not transported to the locus of Psd2(C2⌬)p. DISCUSSION Yeast genetics is an extremely powerful tool for dissecting complex biochemical and biological processes. We are using a combination of yeast genetics, molecular biology, and biochemistry to address fundamental mechanistic questions about interorganelle aminoglycerophospholipid traffic. We have devised and applied genetic screens that selectively use psd1⌬ and psd2⌬ mutants to examine aminoglycerophospholipid transport to and from specific organelles (30). In this report we focused upon the structure of the Psd2p enzyme and the potential role that specific subdomains of the protein play in the lipid transport process.
Examination of the deduced primary sequence of Psd2p reveals three potentially important sequences that could be involved in catalysis, protein traffic, and lipid recognition. In the initial work describing cloning and sequencing of the Psd2p gene (12), amino acid residues (1041-1044) comprising a GGST motif were proposed as candidates comparable to the LGST motif in E. coli (33). Cleavage of the G-S peptide linkage by serinolysis generates the large ␤Ϫsubunit and the smaller ␣-subunit containing a pyruvoyl residue that is part of the active site. In this report we now provide evidence, shown in Fig. 2, that Psd2p is initially made as a high molecular weight precursor of 130 kDa that is processed to yield an 11-kDa small subunit. This result is entirely consistent with previous structural predictions about the enzyme. The immature form of Psd2p appears relatively long lived, because the protein was detected by immunoblot analysis. The unprocessed form of Psd2p is also located on gradients where the mature ␣-subunit is found, suggesting that maturation is not a prerequisite for targeting Psd2p to Golgi/vacuolar membranes but can occur after arrival of the precursor.
A putative GR sequence (EFDIYNEDEREDSDFQSK) in Psd2p with homology to a Kex2p sequence (EFDIIDTDSEYD-STLDNK) implicated in its membrane targeting (18) was also identified in the original description of the PSD2 gene (12). Because subcellular fractionation indicated colocalization of a subpopulation of Psd2p molecules with the Kex2p, this GR motif was tested for function. Deletion of the GR motif in Psd2p did not alter the catalysis, processing, or subcellular localization of the enzyme. These findings clearly demonstrate that the putative GR motif of Psd2p is not required for any identifiable function.
The Psd2p also contains a C2 domain most closely related to the C2 domains of synaptotagmin III and PKC␣ (12). Although C2 domains are most often associated with Ca 2ϩ -mediated binding between proteins, or proteins and lipids (20,21), there is no requirement for Ca 2ϩ in Psd2p catalysis. We designed experiments to test the function of the C2 domain in Psd2p by deletion analysis. A gene encoding Psd2(C2⌬)p was expressed in psd1⌬ psd2⌬ cells, and the function of the mutant protein was examined. When the Psd2(C2⌬)p is expressed on high copy plasmids, a catalytically active protein results. By expressing Psd2(C2⌬)p on a high copy plasmid the resultant yeast strains produce ϳ10 times the amount of enzyme activity that is sufficient for the replication of cells under non-permissive growth conditions. These findings clearly demonstrate that the C2 domain is not essential for catalysis. Close inspection of the immunoblot data in Fig. 3 suggests that the processing of Psd2p to the mature form may be modestly retarded compared with the wild type protein, but the C2 domain is clearly not essential for the proteolytic cleavage of the enzyme to its mature form. Subcellular fractionation of cells expressing Psd2(C2⌬)p reveals that distribution of the mutant enzyme parallels that of its wild type counterpart. From this data we conclude that the C2 domain also does not participate in directing Psd2p to its correct subcellular location.
The in vivo function of Psd2(C2⌬)p was tested in a genetic background (psd1⌬psd2⌬) where cell growth requires either Psd2p activity or Etn supplementation (30). In the absence of Etn, the Psd2(C2⌬)p failed to support cell growth even though the catalytic activity present was 10-fold higher than what is sufficient for cell growth, and the subcellular location of the enzyme was normal. This striking effect on growth indicates that the C2 domain plays a crucial role in regulating a noncatalytic aspect of Psd2p function. Most significantly, experimentation reveals that permeabilized cells produce a pool of PtdSer that cannot be decarboxylated by Psd2(C2⌬)p. However, in this same reconstituted transport system, introduction of NBD-Ptd[1Ј-14 C]serine results in significant decarboxylation of the substrate by Psd2(C2⌬)p. Thus, the defect appears to be the access of the substrate to the active site. We are currently testing the idea that the C2 domain may function in docking the membranes containing Psd2p with membranes that are the proximal donors of PtdSer. The C2 domain or other elements of Psd2p may participate in the transfer of PtdSer to the membrane in which the decarboxylase resides. This proposed docking and transfer process appears to involve another recently described gene product, PstB2p, that is required on the acceptor membrane for the transport-dependent decarboxylation of PtdSer (17). We believe that the Psd2p decarboxylates PtdSer within its own bilayer (catalysis in "cis") rather than that present in an apposed bilayer (catalysis in "trans"), because the enzyme can use PtdSer added in detergent solutions or NBD-PtdSer that readily partitions into membranes, but not liposomal PtdSer. In addition, structural predictions place the active site at the N terminus of the ␣-subunit in a hydrophobic patch located in the cytoplasmic leaflet of the lipid bilayer (psort.nibb.ac.jp/formz.html). The simplest model based upon structural considerations favors catalysis in cis, but rigorous testing of this model is clearly needed. Ligands recognized by C2 domains include PtdSer, PtdIns, and polyphosphoinositides (27)(28)(29). All of the aforementioned ligands have been implicated by other criteria as important elements of aminophospholipid traffic along the phosphatidylserine B pathway (30). PtdSer is the substrate for Psd2p and as described above appears to require interaction with the C2 domain prior to catalysis. PtdIns is known to bind PstB2p, which must be present on the acceptor membrane for PtdSer to be decarboxylated. PtdIns-4-P has been implicated in the lipid transport to Psd2p because mutant strains with defects in the PtdIns 4-kinase (such as Stt4p) are defective in converting nascent PtdSer to PtdEtn. Collectively these data suggest that multiple protein-lipid interactions and probably protein-protein interactions occur between the donor and acceptor membranes to facilitate the interorganelle movement of PtdSer.
In summary, our data demonstrate that the C2 domain of Psd2p is dispensable for catalysis but essential for enzyme function in intact and permeabilized cells. These findings suggest that the C2 domain participates in interorganelle PtdSer transport, most likely in cooperation with other membraneassociated proteins.