Identification of a Novel Ca 2 (cid:49) -dependent, Phosphatidylethanolamine-hydrolyzing Phospholipase D in Yeast Bearing a Disruption in PLD1 *

We have previously reported the identification and partial characterization of a gene encoding a phospholipase D activity ( PLD1 ) in the yeast, Saccharomyces cerevisiae. Here we report the existence of a second phospholipase D activity, designated PLD2, in yeast cells bearing disruption at the PLD1 locus. PLD2 is a Ca 2 (cid:49) -dependent enzyme which preferentially utilizes phosphatidylethanolamine over phosphatidylcholine as a substrate. In contrast to PLD1, the activity of PLD2 is insensitive to phosphatidylinositol 4,5-bisphosphate, and the enzyme is incapable of catalyzing the transphosphatidylation reaction with short chain alcohols as ac- ceptors. Subcellular fractionation shows that PLD2 lo-calizes mainly to the cytosol, but could also be detected in the particulate fraction. Thus, the biochemical properties of PLD2 appear to be substantially different from those of PLD1. PLD2 activity is significantly and transiently elevated upon exit of wild type yeast cells from stationary phase, suggesting that it may play a role in the initiation of mitotic cell division in yeast. In view of the significantly different properties of PLD1 and PLD2, and because the yeast genome contains PLD1 as the sole member of the recently defined PLD gene family, it may be concluded that PLD2 is structurally unrelated to PLD1. Thus, the novel PLD2 activity described herein is likely to represent the first identified member of a Phospholipase D Assays— Total cell lysates de- C 6 -NBD-PC was in In solubi-lization of C 6 -NBD-PE, 1.5 m M Triton X-100 was added. The final concentration of Triton X-100 in assay reactions containing C 6 -NBD-PE was 0.25 m M . For determination of substrate specificity, all substrates were solubilized in 6 m M Triton X-100. The final concentration of the detergent in the reaction was 1.2 m M . The hydrolysis of C 6 -NBD-phospholipids was monitored by the production of C 6 -NBD-PA, as by For assaying PLD1 activity, we prepared cell lysates from wild-type JC1 cells. The PLD1 reaction contained 0.3 mg/ml protein, 35 m M HEPES, m M NaCl, (cid:109) M C 6 -NBD-PC, m M EDTA, 5 m M EGTA, and 4 mol % PIP 2 . The surface concentration of PIP 2 is ex-*

Phospholipase D catalyzes the hydrolysis of phospholipids at their distal phosphodiester bond to yield phosphatidic acid (PA) 1 and a free polar head group (1). Most PLDs can utilize primary alcohols as acceptors of the phosphatidyl moiety, to yield the corresponding phosphatidylalcohols (1). In mammals, different PLD activities with various substrate specificities, activation requirements, and subcellular localization have been described (2). PLD activity can be rapidly activated upon receptor activation or other types of cell stimulation. Receptor activation of PLD is probably mediated by multiple factors including small GTP-binding proteins, heterotrimeric GTPbinding proteins, protein kinase C, phosphatidylinositol 4,5bisphosphate (PIP 2 ), tyrosine phosphorylation, and changes in the intracellular concentration of Ca 2ϩ (3,4). Phosphatidic acid, the product of the reaction catalyzed by PLD, is likely to serve as a lipid second messenger (5) and may mediate a variety of biological responses, including mitogenesis (6) and the respiratory burst (7). In addition, an ADP-ribosylation factor (ARF)-activated phosphatidylcholine (PC)-specific PLD which was discovered recently (8,9) has been implicated in vesicular trafficking (10). The cloning of the first plant PCspecific PLD (11) enabled the identification of additional eukaryotic PC-specific PLDs, including a yeast PLD (12)(13)(14) and a human ARF-dependent PLD (15). The eukaryotic PC-PLD gene family shares a number of common homology domains (15)(16)(17). A short common motif is shared with certain phospholipid biosynthetic enzymes that catalyze a phosphatidyl transfer reaction, but are not PLDs (15,16). In yeast, PLD1 has no direct homologs and is therefore the sole member of this family. This paper reports the identification of a second PLD activity in yeast and describes its biochemical properties, distribution in subcellular fractions, and activation during early growth phase.
Media-Wild-type yeast were maintained on synthetic complete minimal medium (SC). PLD1⌬FS-1 cells were maintained on SC drop-out medium, lacking histidine. Media were prepared essentially according to Rose et al. (20).
Phospholipase D Assays-Total cell lysates were prepared as described previously (13). C 6 -NBD-PC was dissolved in water. In solubilization of C 6 -NBD-PE, 1.5 mM Triton X-100 was added. The final concentration of Triton X-100 in assay reactions containing C 6 -NBD-PE was 0.25 mM. For determination of substrate specificity, all substrates were solubilized in 6 mM Triton X-100. The final concentration of the detergent in the reaction was 1.2 mM.
The hydrolysis of C 6 -NBD-phospholipids was monitored by the production of C 6 -NBD-PA, essentially as described by Danin et al. (21). For assaying PLD1 activity, we prepared cell lysates from wild-type JC1 cells. The PLD1 reaction mixture contained 0. For the transphosphatidylation assays, 300 mM concentrations of the indicated primary alcohols were added as acceptors. The reaction mixtures were incubated at 30°C for 30 min at final volume of 120 l. Termination, TLC separation, and quantification of the fluorescent lipid products were conducted as described (13,21). Activity is expressed as the mean of two duplicate samples measured in arbitrary fluorescence units. Specific activity is expressed as the PA-derived fluorescence units per g of protein.

RESULTS AND DISCUSSION
We have previously reported that disruption of the PLD1 gene results in complete loss of PLD activity in ⌬pld1 cell lysates (13), when measured under assay conditions which included EGTA, a Ca 2ϩ chelator (Fig. 1, left). In subsequent experiments designed to further characterize PLD1 activity and in which EGTA was omitted from the assay, we found elevated PLD activities in both wild type and ⌬pld1 cell lysates (data not shown). These results suggested that yeast cells express an additional, Ca 2ϩ -dependent PLD activity, which is independent of PLD1. Indeed, addition of CaCl 2 at a concentration that allowed 1 mM free Ca 2ϩ in the presence of EGTA and EDTA resulted in a substantial increase in PLD activity which was evident in both wild type and ⌬pld1 cell lysates (Fig.  1, right). Total PLD activity was consistently lower in ⌬pld1 cell lysates, reflecting the loss of PLD activity contributed by the disrupted PLD1 gene product. Therefore, the residual Ca 2ϩdependent PLD activity observed in ⌬pld1 cell lysates represents a second PLD enzyme which we have designated PLD2. Thus, use of the yeast system has allowed us to assay the activities of PLD1 and PLD2 separately: PLD1 activity could be assayed in wild type cell lysates in the presence of EGTA (which abolishes PLD2 activity), whereas PLD2 activity could be assayed in the presence of Ca 2ϩ in ⌬pld1 cell lysates (in which PLD1 activity is abolished as a consequence of gene disruption).
The substrate specificity of PLD2 was next compared with that of PLD1. This was accomplished by assaying both PLD activities with 400 M C 6 -NBD-PC, C 6 -NBD-PG, and C 6 -NBD-PE in the presence of 1.2 mM Triton X-100 (Fig. 2). At a concentration of 0.5 mM, Triton X-100 inhibited PLD1 activity by 25% and activated PLD2 activity using C 6 -NBD-PE by 30% (data not shown). We found that the production of C 6 -NBD-PA was up to 75-fold higher with C 6 -NBD-PE as a substrate compared with C 6 -NBD-PC. In subsequent experiments, when Triton X-100 was used in a concentration that is below its critical micelle concentration value, PLD2 activity with C 6 -NBD-PE was up to 10-fold higher than with C 6 -NBD-PC as a substrate (data not shown). In contrast, PLD1 could not utilize C 6 -NBD-PE as a substrate. Similar results were reported previously for PLD1 with long chain PE (12). Furthermore, compared with PLD1, PLD2 produced 4 and 11 times as much C 6 -NBD-PA from hydrolyzing C 6 -NBD-PC and C 6 -NBD-PG, respectively. These data suggest that a considerable difference exists in the catalytic properties of PLD1 and PLD2. In addition, the results provide a tool by which PLD1 and PLD2 activities could be discriminated in wild type cell lysates, where C 6 -NBD-PE can be used as a substrate for determination of PLD2 activity exclusively, without the interference of PLD1. PLD2 activity increased when increasing concentrations of either C 6 -NBD-PC or C 6 -NBD-PE were added to the reaction mixture. Using C 6 -NBD-PC as a substrate, PLD2 activity reached saturation at 200 M. In contrast, saturation could not be reached with C 6 -NBD-PE concentrations up to 400 M (data not shown). C 6 -NBD-PE was utilized routinely as a substrate in PLD2 reactions, as it is not used by PLD1 and because of the greater activity that can be achieved with this substrate. Throughout the remainder of PLD2 characterization assays, we used both C 6 -NBD-PC and C 6 -NBD-PE. The concentration used for C 6 -NBD-PC (400 M) allowed enzyme saturation, and the concentration used for C 6 -NBD-PE was submaximal (40 M) in the presence of 0.25 mM Triton X-100.
There is evidence for PLD-mediated hydrolysis of PE in intact mammalian cells (22, 23). Furthermore, a cytosolic PLD activity identified in various bovine tissues exhibited some preference for PE over PC as substrate (24). The mammalian cytosolic PLD was stimulated by Ca 2ϩ (24,25). However, in contrast to yeast PLD2, the mammalian cytosolic PLD could efficiently catalyze a transphosphatidylation reaction (24,25).
To examine the quantitative relationship between Ca 2ϩ con-  Yeast 37 centration and PLD2 activity, enzyme activity was determined in the presence of increasing free Ca 2ϩ concentrations in the presence of EGTA (Fig. 3). PLD2 activity was elevated substantially and dose-dependently as a function of free Ca 2ϩ . Significant activation was observed at free Ca 2ϩ concentrations above 0.1 M while a maximal effect was obtained at a Ca 2ϩ concentration of 1 mM. The effect of Ca 2ϩ was largely independent of the substrate used. The mode of PLD2 activation by Ca 2ϩ is not clear. It is known that intracellular Ca 2ϩ can be elevated in vivo to levels sufficient to activate PLD2 (26). One possibility is that Ca 2ϩ may be involved in mediating a translocation of PLD2 from the cytosol to the membrane, as is the case for conventional protein kinase C isoforms. PIP 2 was shown to act as a cofactor of two membranal PCspecific PLDs from brain (27,28) and an activator of yeast PLD1 (13). The effect of increasing surface concentrations of PIP 2 on yeast PLD1 and PLD2 was determined using either C 6 -NBD-PC or C 6 -NBD-PE as substrates. PLD2 activity was not affected by up to 10 mol % PIP 2 (data not shown). In contrast, PLD1 activity was doubled in the presence of as little as 0.01 mol % PIP 2 . Thus, unlike the PC-specific PLD enzymes, PLD2 appears to be insensitive to PIP 2 . Consequently, PIP 2 was omitted from the reaction mixture in PLD2 activity assays.

Identification and Characterization of a Novel PLD in
The production of phosphatidylalcohols by a transphosphatidylation reaction is a unique property of PLDs which had been utilized extensively as a specific marker of PLD activity in vitro and of PLD activation in situ (see Refs. 1 and 3). Therefore, we have studied the efficiency by which PLD2 could use primary alcohols as substrates for transphosphatidylation. Ethanol, 1-propanol, and 1-butanol were added to the PLD2 reaction mixture at a concentration of 300 mM. Interestingly, PLD2 activity of ⌬pld1 cell lysate produced no phosphatidylalcohols whether C 6 -NBD-PC or C 6 -NBD-PE was utilized as substrate (data not shown). Thus, PLD2 appears to be incapable of transphosphatidylation when simple short chain primary alcohols are added as acceptors. This contrasts sharply with the ability of yeast PLD1 to catalyze transphosphatidylation (12)(13)(14) and with the efficient transphosphatidylation reaction carried out by the mammalian PC-hydrolyzing enzymes. Our data imply that certain forms of PLD lack this property. This conclusion is supported by previous reports showing that some PLD activities isolated from variant strains of the bacterium Streptomyces chromofuscus exhibit a greatly reduced or lack of ability to produce phosphatidylalcohols (29 -31). Thus, whereas the production of phosphatidylalcohols (with simple short chain alcohols) remains a unique property of PLDs, not all PLDs can catalyze a transphosphatidylation.
Mammalian cells contain both cytosolic (24,25,32,33) and membrane-bound (34,35) forms of PLD. To examine the subcellular localization of PLD2, its activity was determined in soluble and particulate fractions derived from yeast cell lysates (Table I). Subcellular fractionation shows that most of the PLD2 activity was soluble and found in the 100,000 ϫ g supernatant. However, a significant amount of PLD2 activity was membrane-associated.
It has already been demonstrated that PLD1 activity is required for meiotic cell division in yeast (12)(13)(14). It was suggested that PLD1 may inhibit mitotic cell division (36). We decided, therefore, to check whether PLD2 activity is altered during mitotic cell division. PLD2 activity was determined at different time points after the dilution of stationary-phase (G 1arrested) wild-type cells into fresh medium (Fig. 4). It was found that PLD2 was transiently activated upon culture dilution in fresh media, both before and during the interval in which the first cell division took place. Two peaks of activation could be seen. The first peak represents rapid, transient 2-fold activation of PLD2 measured within 20 min after the transfer to fresh medium and before initiation of the first cell division. This peak declines after 40 min. The second peak of activity is  4. Time course of PLD2 activation during exit from stationary phase. PLD2 activity was determined at different stages of growth in culture. A 12-h-old stationary phase culture of wild-type W303-1B cells was diluted in fresh YPD medium to 0.6 ϫ 10 6 cells/ml and grown at 30°C. Samples were taken at the indicated times for PLD2 activity assays (solid circles) and for cell density determination (open squares). PLD2 activity was assayed as described under "Experimental Procedures" except 20 M C 6 -NBD-PE was used. The arrow indicates the approximate time of entry into the first cell division cycle.

TABLE I Distribution of PLD2 activity in subcellular fractions
Stationary phase W303-1B wild-type cells were diluted and grown to midlog phase. Whole-cell lysates were centrifuged at 8,000 ϫ g for 10 min. The supernatant was collected and pellet was washed by resuspension in lysis buffer for assay. The 8,000 ϫ g supernatant was ultracentrifuged for 1 h at 100,000 ϫ g. The supernatant was collected and the pellet was washed as above and resuspended in lysis buffer for assay. Fractions were assayed for PLD2 activity as described under "Experimental Procedures." Total activity is expressed in phosphatidic acid-derived fluorescence units at 520 nm. Specific activity is expressed as fluorescence units per g of protein. Identification and Characterization of a Novel PLD in Yeast 38 observed 2-4 h after the transfer and then declines. These results suggest a biphasic response of PLD2 to induction of mitotic cell division in G 1 -arrested cells and indicate that PLD2 is a regulated enzyme. Interestingly, this activation of PLD2 was detected under in vitro assay conditions. Hence, the stimulation of PLD2 activity during early growth phase may involve a persistent modification of the enzyme. The rapid activation upon growth stimulation suggests that PLD2 may mediate early growth signals and play a role in the initiation of the mitotic cell cycle. In mammalian cells, PLD is activated by growth factors and other mitogens, suggesting a role in mediating G 0 -G 1 transition (6). In addition, there is evidence for a negative modulatory effect of PLD also on G 2 -M phase transition (37). Presumably, the PA produced by PLD2 may regulate one or more of the proteins involved in cell cycle control.
The radically different catalytic properties of PLD1 and PLD2 indicate that these two enzymes are probably structurally unrelated. Indeed, an extensive analysis of the complete S. cerevisiae genome by the BLAST algorithm, using several highly conserved domains present in known PLD family members as well as the putative Ca 2ϩ -binding domain of plant PLD, has demonstrated that PLD1 has no direct homologs. Thus, in yeast, PLD1 is the sole member of the recently defined gene family (15,16) that, in addition to PLD1, includes the plant PLDs, the human ARF-dependent PLD, and a putative Caenorhabditis elegans PLD gene. Consequently, we postulate that the gene encoding PLD2 does not belong to the PLD/phosphatidyltransferase gene family and, thus, may represent the first identified member of a novel PLD family. Ongoing work in our laboratory is aimed at identifying and cloning the PLD2 gene.