Cloning of a Chinese Hamster Ovary (CHO) cDNA Encoding Phosphatidylserine Synthase (PSS) II, Overexpression of Which Suppresses the Phosphatidylserine Biosynthetic Defect of a PSS I-lacking Mutant of CHO-K1 Cells*

Phosphatidylserine (PtdSer) in mammalian cells is synthesized through the exchange of free l-serine for the polar head group (base) of preexisting phospholipid. We previously showed the presence of two different enzymes catalyzing the serine base exchange in Chinese hamster ovary (CHO) cells and isolated the cDNA of one of the enzymes, PtdSer synthase (PSS) I, which also catalyzes the exchange of the base moiety of phospholipid(s) for ethanolamine and choline. In this study, we cloned a CHO cDNA, designated aspssB, which encodes a protein exhibiting 32% amino acid sequence identity with CHO PSS I. Introduction of the pssBcDNA into CHO-K1 cells resulted in striking increases in both the serine and ethanolamine base exchange activities. In contrast to the PSS I cDNA, the pssB cDNA was incapable of increasing the choline base exchange activity. The expression of thepssB gene in Sf9 insect cells also results in striking increases in both serine and ethanolamine base exchange activities. ThepssB cDNA was found to transform a PtdSer-auxotrophic PSS I-lacking mutant of CHO-K1 cells to PtdSer prototrophy. The PtdSer content of the resultant transformant grown without exogenous PtdSer for 2 days was 4-fold that of the mutant and similar to that of CHO-K1 cells, indicating that the pssB cDNA complemented the PtdSer biosynthetic defect of the PSS I-lacking mutant. These results suggested that the pssB cDNA encoded the second PtdSer synthase PSS II, which catalyzed the serine and ethanolamine base exchange, but not the choline base exchange.


Phosphatidylserine (PtdSer) in mammalian cells is
Phosphatidylserine (PtdSer) 1 is one of the major membrane phospholipids in mammalian cells, comprising about 10% of the total phospholipids. To elucidate the biosynthetic pathway and biological function of PtdSer, we previously isolated a Chinese hamster ovary (CHO) cell mutant that requires exogenous Ptd-Ser for cell growth (1). This PtdSer-auxotrophic mutant, PSA-3, is strikingly defective in PtdSer biosynthesis, and its cell extract exhibits an about 50% decrease in the activity of the exchange between free L-serine and the polar head group (base) of preexisting phospholipids for PtdSer formation (1). The parental CHO-K1 cells, but not the mutant cells, have the ability to use phosphatidylcholine (PtdCho) as a phosphatidyl donor for serine base exchange (2). In contrast to PtdCho, phosphatidylethanolamine (PtdEtn) is used by both the CHO-K1 and mutant cells as a phosphatidyl donor for serine base exchange (2). Furthermore, the mutant has been shown to grow normally in a growth medium supplemented with PtdEtn, synthesizing a normal amount of PtdSer (2). These findings led us to conclude that PtdSer in CHO cells is synthesized through serine base exchange, which is catalyzed by two kinds of PtdSer synthase (PSS): one, PSS I, which is absent in mutant PSA-3, can use PtdCho as a substrate; and the other, PSS II, uses PtdEtn but not PtdCho as a substrate.
PSS I appears to contribute to the formation of both PtdSer and PtdEtn because mutant PSA-3 grown without exogenous phospholipids exhibited decreased levels of both PtdSer and PtdEtn (1). Since it has been suggested that much of the PtdEtn in CHO cells is synthesized through the decarboxylation of PtdSer (3), PtdSer synthesized by PSS I probably serves as an essential precursor of PtdEtn.
A CHO cDNA (designated as pssA) which complements the PSS I defect of mutant PSA-3 has been isolated (4). With antibodies raised against synthetic pssA peptides, we recently provided several lines of evidence that the pssA cDNA encodes PSS I (5). PSS I deduced from the cDNA sequence is composed of 471 amino acid residues and has several potential membrane-spanning domains (4). In addition to serine base exchange, PSS I catalyzes the exchange of the base moiety of phospholipid(s) for choline and ethanolamine in cell extracts (1,4,6). Subcellular fractionation followed by immunoblotting with an anti-(pssA peptide) antibody indicated that PSS I is enriched in mitochondria-associated membranes and microsomal membranes (5).
Although the function and structure of PSS I are being elucidated, as described above, less is known concerning PSS II, since neither the mutant nor the gene has been isolated. To address the role of PSS II in PtdSer metabolism and its biological significance, we have attempted to isolate the cDNA of PSS II. Here, we describe the cloning of the PSS II cDNA.

EXPERIMENTAL PROCEDURES
Strains and Culture Conditions-Strain CHO-K1 was obtained from the American Type Culture Collection. CHO-K1, mutant PSA-3 (1), and transformant PSA-3/pssB obtained in this study were maintained as described (1). For the ethanolamine supplementation experiment, 100 ml of newborn calf serum was dialyzed three times against 2 liters of phosphate-buffered saline for about 12 h and filter sterilized. Spodoptera frugiperda (Sf9) cells were provided by Dr. Yoshiharu Matsuura (National institute of Health, Tokyo, Japan) and maintained in TC-100 medium (Life Technologies, Inc.) supplemented with 10% (v/v) heatinactivated fetal bovine serum, 10 g/ml gentamycin, and 0.26% (w/v) tryptose phosphate broth (Life Technologies, Inc.) at 27°C.
Construction of a cDNA Library-Poly(A) ϩ RNA was prepared from CHO-K1 cells as described (7) and used for cDNA synthesis. The cDNA synthesis and construction of a cDNA library in a plasmid vector, pSPORT1, were performed with a SuperScript plasmid system (Life Technologies, Inc.) according to the manufacturer's instructions.
Cloning of a pssB cDNA-Oligonucleotides corresponding to parts of a human expressed sequence tag (EST) (GenBank number F11951) were used to amplify a pssB cDNA fragment from the CHO cDNA library by means of a two-stage polymerase chain reaction. The primers used for the first round of amplification were TCCAGACTGTCCAG-GACGGC (sense) and AGGAACTCGAACATCACGCT (antisense). The same antisense primer and TCCAGGACGGCCGGCAGTTT (sense) were used for the second round of amplification. The amplification reactions were performed for 35 cycles of denaturation at 94°C for 45 s, annealing at 55°C for 45 s, and elongation at 72°C for 90 s, with Taq polymerase (Perkin-Elmer) according to the manufacturer's instructions. For the second round of amplification, the first round reaction mixture was diluted 5,000-fold and used as the template. The 0.3kilobase pair product of the second round of amplification was subcloned into a plasmid, pBluescript II SKϩ (Stratagene), sequenced, and used as a hybridization probe for screening of the CHO cDNA library, after enrichment of hybridizing clones with a GeneTrapper cDNA-positive selection system (Life Technologies, Inc.). The enrichment was performed with a biotinylated oligonucleotide, GACTGGTG-GATGTGCATGATCATC, corresponding to a part of the 0.3-kilobase pair polymerase chain reaction product, according to the manufacturer's instructions except for omission of the repair reaction. Colony filter hybridization with the 32 P-labeled probe was performed as described (8); hybridization was performed for 22 h at 42°C in 5 ϫ SSPE (1 ϫ SSPE ϭ 0.15 M NaCl, 1 mM EDTA, 10 mM NaH 2 PO 4 , pH 7.4), 5 ϫ Denhardt's solution, 0.5% sodium dodecyl sulfate, 50% formamide, and 100 g/ml denatured salmon sperm DNA; the final wash was performed in 0.2 ϫ SSC (1 ϫ SSC ϭ 0.15 M NaCl, 15 mM sodium citrate, pH 7.0), and 0.1% sodium dodecyl sulfate at 50°C for 1 h.
DNA Sequencing-Both strands of the pssB cDNA were determined by the dideoxy chain termination method with Sequenase (U. S. Biochemical Corp.) according to the manufacturer's instructions, using a series of deletion mutants generated by exoIII/mung bean nuclease treatment (9), in combination with walking primers.
Transient Transfection of CHO-K1 Cells with the pssB and pssA cDNAs-A plasmid, pSPORT1/pssB, carrying the pssB cDNA, was cleaved at the SalI and NotI sites, and the resulting pssB cDNA fragment was inserted into these restriction enzyme sites of a mammalian expression plasmid vector, pSV-SPORT1 (Life Technologies, Inc.). The resulting construct, pSVpssB, and pcDPSSA encoding the pssA protein (CHO PSS I) (4) were introduced into CHO-K1 cells using Lipo-fectAMINE reagent (Life Technologies, Inc.) according to the manufacturer's instructions.
Heterologous Expression of the pssB gene-The plasmid, pSPORT1/ pssB, was cleaved at the SalI and HindIII sites, and the resulting pssB cDNA fragment was inserted into the XhoI and HindIII sites of a baculovirus transfer vector, pBlueBac4.5 (Invitrogen). A monolayer of Sf9 cells (35-mm-diameter dish) was cotransfected with the resulting construct, pBlueBac4.5/pssB and Bac-N-Blue Autographa californica DNA (Invitrogen), using Lipofectin reagent (Life Technologies, Inc.). For production of control virus, another monolayer of Sf9 cells was cotransfected with the transfer vector pBlueBac4.5 and Bac-N-Blue DNA. After 4 days of culture, the transfection supernatants containing pssB recombinant virus or control virus were collected. Fresh monolayers of Sf9 cells (75-cm 2 flask) were infected with the transfection supernatant containing pssB recombinant virus or control virus and cultivated in TC-100 insect medium (Life Technologies, Inc.) containing 10% heat-inactivated fetal bovine serum. After 4 days at 27°C, cells were harvested, washed two times with ice-cold phosphate-buffered saline, pH 7.4, resuspended in 0.25 M sucrose containing 1 mM EDTA, 2 g/ml leupeptin, 2 g/ml aprotinin, 1 mM (p-amidinophenyl)methanesulfonyl fluoride, and 10 mM HEPES, pH 7.5, and disrupted by two 15-s sonication bursts on ice. Phospholipid base exchange activity in the preparations was assayed as described (1).
Isolation of Transformant PSA-3/pssB-Mutant PSA-3 cells were transfected with pSVpssB by the calcium phosphate precipitation method (10), and the resultant transformant, designated as PSA-3/ pssB, which was able to grow in the growth medium without exogenous phospholipids, was purified by limited dilution of the transfected cells.
Other Methods-Radioactive labeling, extraction, separation, and quantitation of phospholipids were performed as described in Tables III  and IV and in the legends to Figs. 5, 6, and 7. [ 32 P]PtdCho was prepared from CHO-K1 cells metabolically labeled with 32 P i as described (2). Protein was measured according to Lowry et al. (11), using bovine serum albumin as a standard.

Isolation of a cDNA Clone Encoding a Protein Similar in
Sequence to PSS I-To identify cDNA clones encoding PSS II, the amino acid sequence of PSS I predicted from the cDNA sequence (4) was compared with the ESTs in DNA data bases using the TBLASTN search protocol at the National Center for Biotechnology Information. A human EST (GenBank number F11951) was found to encode a peptide that exhibited 30% sequence identity, in a stretch of 98 amino acids, with both CHO PSS I and a putative human PSS I (GenBank number D14694). A cDNA fragment of the CHO counterpart for human EST F11951 was generated from a CHO cDNA library, using a polymerase chain reaction and primers corresponding to parts of the EST sequence. The resulting cDNA fragment was used as a hybridization probe to screen the CHO cDNA library, after enrichment of hybridizing clones using a GeneTrapper cDNA-positive selection system with a biotinylated oligonucleotide specific to the cDNA fragment. Sequence analysis of a hybridizing cDNA clone revealed a large open reading frame encoding a protein of 474 amino acid residues with a calculated molecular mass of 55,003 Da (Fig. 1). The gene of this protein was designated as pssB. The predicted pssB protein exhibited 32% amino acid sequence identity with the pssA gene product, PSS I (Fig. 2). Hydrophobicity analysis of the predicted pssB protein by the method of Kyte and Doolittle (12) revealed a highly hydrophobic protein containing several potential membrane-spanning domains, the hydrophobicity profile of which was very similar to that of PSS I, as shown in Fig. 3. In addition, the pssB protein was found to have, at its NH 2 terminus, a Met-Arg-Arg-Ala-Glu sequence that corresponds to an NH 2 -terminal double arginine motif known as an ER targeting signal (13). The presence of the motif implied that the pssB protein is localized to ER. It is noteworthy that a strict positional requirement for the two arginines of the double arginine motif has been shown: namely, efficient targeting of recombinant reporter proteins to ER is only accomplished if two adjacent arginines or two arginines separated by one residue are located in the four residues following the initiator methionine (13). Thus, given that the double arginine motif of the pssB protein actually functions as an ER targeting signal (see "Discussion"), the putative initiator methionine codon appeared to be the true initiation codon, although the absence of a 5Јtermination codon raised the possibility that the isolated pssB cDNA was a partial clone.
Phospholipid Base Exchange Activities in pssB-transfected Cells-To determine whether the pssB gene product is relevant to serine base exchange for PtdSer formation, the pssB cDNA was placed downstream of the mammalian expression promoter of a plasmid, pSV-SPORT1, and the resulting construct, pSVpssB, was introduced into CHO-K1 cells. The transient transfectant with pSVpssB exhibited a 6-fold higher specific activity of serine base exchange for PtdSer formation than CHO-K1 cells transfected with the control vector (Table I). The ethanolamine base exchange activity also increased 10-fold upon transfection with the pssB cDNA (Table I). On the other hand, there was no significant difference in the choline base exchange activity between the transfectant with pSVpssB and the control CHO-K1 cells. In contrast to pSVpssB, a plasmid pcDPSSA (4) encoding CHO PSS I was capable of increasing the choline base exchange activity, as well as the serine and ethanolamine base exchange activities, upon transient transfection (Table I). These results, together with the sequence similarity of the pssB gene product to PSS I, suggested that pssB encoded an enzyme catalyzing both the serine and ethanolamine base exchange but not the choline base exchange in cell extracts.
Heterologous Expression of the pssB Gene-To obtain further evidence that pssB encodes an enzyme catalyzing both the serine and ethanolamine base exchange, we used heterologous expression of the pssB gene in insect cells. The pssB cDNA was placed within the genome of baculovirus under control of the polyhedrin promoter and expressed by viral infection of Sf9 cells. The serine and ethanolamine base exchange activities in a homogenate of Sf9 cells infected with the pssB-containing baculovirus were, respectively, 5.3-and 7.5-fold higher than those of Sf9 cells infected with a control virus (Fig. 4). On the other hand, the choline base exchange activity was not elevated by the pssB virus infection (Fig. 4). Thus, in this heterologous system the pssB cDNA was capable of increasing both the serine and ethanolamine base exchange activities, suggesting The amino acid sequence of the predicted pssB product is aligned with that of PSS I deduced from the pssA cDNA sequence. Gaps are inserted to allow alignment that gives the greatest identity. Identical amino acids are denoted by asterisks. FIG. 3. Hydrophobicity plots for the predicted pssB gene product (panel A) and the pssA gene product, PSS I (panel B). The average hydrophobicity (12) of a nanodecapeptide composed of amino acids n Ϫ9 to ϩ 9 is plotted against n, the amino acid number. that the cDNA encodes an enzyme catalyzing these two different base exchange reactions.
pssB cDNA Complements the Growth Defect of PSS I-lacking Mutant PSA-3-A PSS I-lacking mutant of CHO-K1 cells, PSA-3, requires the addition of either PtdSer or PtdEtn to the medium for cell growth (1, 2). To determine whether or not the pssB cDNA compensates for the lack of PSS I activity, the PSA-3 mutant cells were transfected with pSVpssB and then cultured in the absence of exogenous phospholipids. The transfection efficiently yielded transformants that were able to grow in the absence of exogenous phospholipids. The growth rate of the resulting transformant, PSA-3/pssB, was almost the same as that of CHO-K1 in the medium without exogenous phospholipids (Fig. 5). The extract of transformant PSA-3/pssB exhibited strikingly higher serine and ethanolamine base exchange activities than those of the CHO-K1 and mutant PSA-3 cells but remained defective in the choline base exchange activity (Table II). These results indicated that the pssB cDNA, which induced the stable overexpression of the serine and ethanolamine base exchange activities, was able to complement the growth defect of the PSS I-lacking mutant, PSA-3.
pssB cDNA Complements the PtdSer Biosynthetic Defect of PSS I-lacking Mutant PSA-3-The phospholipid composition and content of cells grown without exogenous phospholipids for 2 days were determined. Transformant PSA-3/pssB exhibited a phospholipid composition and content very similar to those of CHO-K1 cells, whereas the contents of PtdSer and PtdEtn of mutant PSA-3 were 21% and 51% of those of CHO-K1 cells, respectively (Table III). To determine the rate of biosynthesis of PtdSer, cells were pulse labeled with L-[U-14 C]serine. Transformant PSA-3/pssB, but not mutant PSA-3, was able to incorporate the label into PtdSer at a rate similar to that of CHO-K1 cells, as shown in Fig. 6. These results showed that the pssB cDNA was able to complement the PtdSer biosynthetic defect of the PSS I-lacking mutant, PSA-3.
The pssB-transformed PSA-3 Mutant Remains Defective in Conversion of PtdCho to PtdSer-Mutant PSA-3 is defective in the conversion of exogenous [ 32 P]PtdCho to [ 32 P]PtdSer because of a lack of PSS I activity (2). To determine whether or not the pssB cDNA complements this defect, cells were metabolically labeled with [ 32 P]PtdCho. CHO-K1 cells incorporated the radioactivity into PtdSer, in an amount comprising 3.6% of the radioactivity of cellular PtdCho (Fig. 7). In contrast to CHO-K1 cells, both transformant PSA-3/pssB and mutant PSA-3 were incapable of incorporating the radioactivity into PtdSer in significant amounts, although the level of cellular [ 32 P]PtdCho in the transformant and mutant was almost the same as that in CHO-K1 cells (Fig. 7). These results indicated that the transformant PSA-3/pssB remained defective in conversion of PtdCho to PtdSer.
Transformant PSA-3/pssB Cultivated in the Medium with Dialyzed Newborn Calf Serum Exhibits a Normal PtdSer Bio-  -SPORT1), pSVpssB, or pcDPSSA, as described under "Experimental Procedures," and were cultured for 3 days at 37°C. Then the cell extracts were prepared and assayed for phospholipid base exchange activities as described (1). Values indicate specific activities and are the averages of duplicate assays, with variation of Ͻ10% between duplicates. CHO (Fig. 8). When ethanolamine was added to the medium at a concentration of 10 M, mutant PSA-3 was able to grow normally for 5 days (Fig. 8). The addition of ethanolamine did not affect the cell growth of transformant PSA-3/pssB and CHO-K1 cells (Fig. 8). A labeling experiment with 32 P i for 48 h revealed that the PtdSer level of transformant PSA-3/pssB grown in the medium with the dialyzed serum was similar to that of CHO-K1 cells (Table IV). Although mutant PSA-3 cultivated in the medium with the dialyzed serum was defective in PtdSer biosynthesis, the addition of ethanolamine to the medium restored a normal level of PtdSer in the mutant (Table IV). Upon cultivation with ethanolamine, there was no significant difference in phospholipid composition among all three strains (Table IV). These results indicated that the restoration of PtdSer biosynthesis in transformant PSA-3/ pssB occurred without ethanolamine supplementation to the medium containing the dialyzed serum and that the addition of ethanolamine to the medium complemented the PtdSer biosynthetic defect of mutant PSA-3. DISCUSSION We have shown the presence of two different enzymes catalyzing the serine base exchange for PtdSer formation in CHO-K1 cells (1, 2). A CHO pssA cDNA encoding one of the and transformant PSA-3/pssB Cells were seeded at 1 ϫ 10 6 cells/150-mm-diameter dish in the growth medium supplemented with 30 M PtdSer at 37°C. After 2 days, the medium was replaced with fresh medium without PtdSer. After an additional 2 days, the cellular phospholipids were extracted and analyzed by two-dimensional thin layer chromatography as described (20). To quantitate the individual phospholipids, the phosphate in each spot on a chromatogram was determined chemically (21). The values in parentheses are expressed in nmol of phospholipid/mg of protein. PS, phosphatidylserine; PE, phosphatidylethanolamine; PC, phosphatidylcholine; SM, sphingomyelin; PI, phosphatidylinositol.  6. Incorporation of L-[U-14 C]serine into PtdSer. Cells were seeded into each of a series of 60-mm-diameter dishes and grown to a density of approximately 3 ϫ 10 5 cells/dish in the growth medium supplemented with 30 M PtdSer at 37°C. Then the medium was replaced with fresh growth medium without exogenous PtdSer. After 1 day, at time 0, the cells were metabolically labeled at 37°C by replacing the medium with fresh growth medium containing 0.2 Ci/ml L-[U-14 C]serine (Amersham). At the time indicated, one dish of each strain was removed, and then the cellular phospholipids were extracted and analyzed by one-dimensional thin layer chromatography as described (20). The number of cells at zero time was determined and used to standardize the results. Ç, CHO-K1; E, PSA-3; q, PSA-3/pssB.

FIG. 7. Conversion of exogenous [ 32 P]PtdCho to [ 32 P]PtdSer.
Approximately 2 ϫ 10 6 cells were seeded into 100-mm-diameter dishes containing the growth medium supplemented with 30 M PtdSer, followed by incubation for 1 day at 37°C. Then cells were washed twice with the growth medium without PtdSer and then incubated in the growth medium without PtdSer at 37°C. After 2 h, the cells were metabolically labeled at 37°C by replacing the medium with fresh growth medium containing 5.4 ϫ 10 5 cpm/ml [ 32 P]PtdCho. After labeling (24 h), the cells were washed three times with the growth medium, and then the cellular phospholipids were extracted and analyzed by two-dimensional thin layer chromatography as described (20).   (4). In this study, we have tried to isolate the cDNA of the second PSS, on the assumption that the second PSS is similar in sequence to PSS I. A CHO pssB cDNA isolated here encodes a protein showing a high (32%) amino acid sequence identity with the pssA-encoding PSS I. Transient transfection of CHO-K1 cells with the pssB cDNA results in a 6-fold increase in the serine base exchange activity. The pssB-transfected cells also exhibit a 10-fold elevated ethanolamine base exchange activity. However, the pssB cDNA is incapable of elevating the choline base exchange activity, in contrast to the pssA cDNA. The expression of the pssB gene in Sf9 insect cells also results in striking increases in both serine and ethanolamine base exchange activities. When the pssB cDNA is introduced into the mutant PSA-3, the mutant recovers a normal level of PtdSer biosynthesis. These results suggest that the pssB cDNA encodes the second PtdSer synthase PSS II, which catalyzes the serine and ethanolamine base exchange but not the choline base exchange.
Both the pssB and pssA proteins deduced from the cDNA sequences have several potential membrane-spanning domains, consistent with the fact that the serine and choline base exchange enzymes are firmly embedded in membranes (14). In addition, the pssB protein has a Met-Arg-Arg-Ala-Glu sequence, which corresponds to an NH 2 -terminal double arginine motif known as an ER targeting signal (13). Although the pssA protein does not have the double arginine motif, it has a unique Gly-Val-Gly-Lys-Lys sequence at its COOH terminus, which is similar to another ER targeting motif, a COOH-terminal double lysine motif (Lys-X-Lys-X-X or X-X-Lys-Lys-X-X) (15,16). Consistent with the presence of these sequences, the serine and choline base exchange activities are recovered in a microsomal fraction containing the bulk of ER and in a mitochondriaassociated membrane fraction containing a subfraction of ER (5,17,18).
PtdSer can be synthesized by the exchange of L-serine with the base moiety of both PtdCho and PtdEtn in CHO-K1 cells (2). Because mutant PSA-3 defective in the conversion of Ptd-Cho to PtdSer is incapable of synthesizing a normal amount of PtdSer (1, 2), the exchange of the choline moiety of PtdCho with L-serine appears to be a major route for PtdSer formation in CHO-K1 cells. Although the pssB-transformed PSA-3 mutant remains defective in the conversion of PtdCho to PtdSer, the transformant produces a normal amount of PtdSer. The transformant shows striking increases in both activities of serine base exchange and ethanolamine base exchange, which is supposed to be the reverse reaction of PtdSer formation from PtdEtn. These results suggest that the majority of PtdSer in the pssB-transformed PSA-3 mutant is produced through the exchange of the ethanolamine moiety of PtdEtn with L-serine.
PtdEtn can be synthesized by three pathways. First, in the CDP-ethanolamine pathway, ethanolamine is phosphorylated and converted to CDP-ethanolamine, and then the phosphoethanolamine moiety is transferred to diacylglycerol for PtdEtn formation (19). The second pathway is the decarboxylation of PtdSer. The third is the ethanolamine base exchange. PtdEtn formation through the decarboxylation and ethanolamine base exchange requires PtdSer. Thus, the exchange of L-serine for the ethanolamine moiety of PtdEtn made by the decarboxylation and ethanolamine base exchange does not yield a net increase in cellular PtdSer content. On the other hand, the ethanolamine-serine exchange using PtdEtn produced through the CDP-ethanolamine pathway results in a net increase in cellular PtdSer content. It is therefore likely that PtdEtn produced through the CDP-ethanolamine pathway contributes to the restoration of PtdSer biosynthesis in the pssB-transformed PSA-3 mutant. Because the restoration of PtdSer biosynthesis in the transformant occurs without ethanolamine supplementation to the medium containing dialyzed newborn calf serum, the level of endogenous ethanolamine appears to be sufficient for PtdSer biosynthesis in the transformant. In contrast, in the PSA-3 mutant cultivated without ethanolamine supplementation, the level of endogenous ethanolamine appears to be insufficient for PtdSer biosynthesis because the mutant is defective in PtdSer biosynthesis unless the growth medium is supplemented with either ethanolamine or PtdEtn (2). Why is the pssB-transformed PSA-3 mutant capable of synthesizing normal amounts of PtdEtn and PtdSer in the absence of ethanolamine supplementation? If the ethanolamine-serine exchange reaction is written together with the reutilization of ethanolamine for PtdEtn synthesis via CDP-ethanolamine pathway, Ethanolamine ϩ ATP 3 phosphoethanolamine ϩ ADP Phosphoethanolamine ϩ CTP 3 CDP-ethanolamine ϩ PPi CDP-ethanolamine ϩ diacylglycerol 3 PtdEtn ϩ CMP PtdEtn ϩ L-serine 3 PtdSer ϩ ethanolamine Diacylglycerol ϩ L-serine ϩ ATP ϩ CTP 3 PtdSer ϩ ADP ϩ CMP ϩ PPi REACTION 1 a catalytic amount of recycling ethanolamine is sufficient to achieve net synthesis of PtdSer from diacylglycerol and L-serine. Therefore, an increase in the overall rate of the ethanolamine cycle comprising the ethanolamine-serine exchange and the CDP-ethanolamine pathway could lead to increased cellular levels of PtdSer and its decarboxylation product, PtdEtn, without consumption of ethanolamine. Considering the above mentioned results, we speculate that the overexpression of PSS II which catalyzes the ethanolamine-serine exchange for Ptd-Ser formation induces an increase in the overall rate of the ethanolamine cycle, thereby complementing the defect of the PSA-3 mutant in PtdSer and PtdEtn biosyntheses. supplemented with dialyzed newborn calf serum Cells were seeded at 5 ϫ 10 6 cells/100-mm-diameter dish in Ham's F-12 medium supplemented with 10% (v/v) dialyzed newborn calf serum and 10 M ethanolamine or the dialyzed serum only at 37°C. After 1 day, 32 Pi was added into the medium at a final concentration of 2 Ci/ml. After an additional 48 h, the cellular phospholipids were extracted and analyzed by two-dimensional thin layer chromatography (19). Radioactivity of the individual phospholipids was analyzed with a bioimage analyzer (Fujix BAS 2000). The values are radioactivities of individual phospholipids, expressed as percentages of the total radioactivity of phospholipids. The abbreviations used are the same as in Table III