Mutants in a macrophage-like cell line are defective in plasmalogen biosynthesis, but contain functional peroxisomes.

We have used a fluorescence-activated cytotoxicity protocol, 9-(1'-pyrene)nonanol (P9OH)/UV selection (Morand, O. H., Allen, L.-A. H., Zoeller, R. A., and Raetz, C. R. H. (1990) Biochim. Biophys. Acta 1034, 132-141), to isolate a series of plasmalogen-deficient mutants in a murine, macrophage-like cell line, RAW 264.7. Three of these mutants, RAW.7, RAW.12, and RAW.108, displayed varying degrees of plasmalogen deficiency (48, 17, and 14% of wild-type levels, respectively), and all three mutants were deficient in peroxisomal dihydroxyacetone phosphate (DHAP) acyltransferase activity (5% of wild-type). Unlike previously described Chinese hamster ovary (CHO) cell mutants, the RAW mutants contained intact, functional, peroxisomes and normal levels of alkyl-DHAP synthase activity, a peroxisomal, membrane-bound enzyme. In RAW.7 and RAW.108 cells, the loss of peroxisomal DHAP acyltransferase is the primary lesion. RAW.12 displayed not only a deficiency in the DHAP acyltransferase activity, but also displayed a second lesion in the biosynthetic pathway, a deficiency in delta 1'-desaturase activity (plasmanylethanolamine desaturase, EC 1.14.99.19), the final step in plasmenylethanolamine biosynthesis. The deficiencies expressed in the mutants represent unique lesions in plasmalogen biosynthesis. Since the RAW cell line is a macrophage-like responsive cell line, these mutants can be used to examine the role of plasmalogens in cellular functions such as arachidonic acid metabolism, prostaglandin synthesis, protein secretion, and signal transduction.

Raetz, C. R. H. (1990) Biochim. Biophys. Acta 1034,[132][133][134][135][136][137][138][139][140][141], to isolate a series of plasmalogen-deficient mutants in a murine, macrophage-like cell line, RAW 264.7. Three of these mutants, RAW.7, RAW.12, and RAW.108, displayed varying degrees of plasmalogen deficiency (48, 17, and 14% of wild-type levels, respectively), and all three mutants were deficient in peroxisomal dihydroxyacetone phosphate (DHAP) acyltransferase activity (5% of wild-type). Unlike previously described Chinese hamster ovary (CHO) cell mutants, the RAW mutants contained intact, functional, peroxisomes and normal levels of alkyl-DHAP synthase activity, a peroxisomal, membrane-bound enzyme. In RAW.7 and RAW.108 cells, the loss of peroxisomal DHAP acyltransferase is the primary lesion. RAW.12 displayed not only a deficiency in the DHAP acyltransferase activity, but also displayed a second lesion in the biosynthetic pathway, a deficiency in A1'desaturase activity (plasmanylethanolamine desaturase, EC 1.14.99.19), the final step in plasmenylethanolamine biosynthesis. The deficiencies expressed in the mutants represent unique lesions in plasmalogen biosynthesis. Since the RAW cell line is a macrophagelike responsive cell line, these mutants can be used to examine the role of plasmalogens in cellular functions such as arachidonic acid metabolism, prostaglandin synthesis, protein secretion, and signal transduction.
In animal cells, each class of amino phosphoglyceride can be divided into three subspecies, distinguished by the substituent associated with the sn-1 carbon of the glycerol backbone. Fig. 1 illustrates the three subspecies of ethanolamine-linked phospholipids. These are: 1) 1,2-diacyl-sn-glycero-3-phosphoethanolamine (phosphatidylethanolamine); 2) 1-alkyl-2-acylsn-glycero-3-phosphoethanolamine (plasmanylethanolamine); and 3) l-alk-l'-enyl-2-acyl-sn-glycero-3-phosphoethanolamine (plasmenylethanolamine). The latter two species are ether-linked lipids with the sn-1 position bearing a long * This work was supported by National Institutes of Health Grant DK40192-04. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
3 To whom correspondence and reprint requests should be addressed. Tel.: 617-638-4010. Fax: 617-638-4041. chain fatty alcohol attached through an ether bond. Plasmalogens (the general name given to the plasmenyl form of a lipid class; e.g. plasmenylethanolamine) have a cis-double bond between the first and second carbon of the fatty alcohol chain. In mammalian cells, plasmalogens are most commonly found as a subspecies of the ethanolamine phospholipids and can constitute a large portion of this class in certain cell types (1,2). In general, stimulatable cells (macrophages, neutrophils, muscle, brain, and neural tissue) contain high levels of plasmalogens (1)(2)(3). In neutrophils, for example, 65% of the ethanolamine phospholipids are plasmenylethanolamine (4). Macrophages show a similar profile, although not to as great an extent (2). Human heart muscle is unusual in that it contains very high levels of the plasmalogen form of choline head group species (plasmenylcholine) (1, 2).
Functional roles for plasmalogens are not known. Proposed functions include: 1) prostaglandin production and/or arachidonic acid metabolism (5); 2) membrane fusion-mediated events such as exocytosis and endocytosis (6); and 3) protection against active oxygen species such as singlet oxygen (7). The findings that the lipid portion of certain phosphatidylinositol-glycan-anchored proteins are ether-linked (8,9), and that a murine lymphoma cell line, unable to express such a protein (Thy-I), is deficient in ether lipid biosynthesis (IO), suggest that certain ether-linked lipids could be important for the proper localization of these proteins.
The murine macrophage-like cell line, RAW 264.7, responds to a variety of stimuli to release arachidonic acid and make prostaglandins (ll), phagocytize zymosan particles (12,13), and secrete proteins (13, 14). We have isolated a series of mutant derivatives from the RAW 264.7 cell line that are reduced in plasmalogen content in an effort to link plasmalogens with these, or other, cellular processes. These mutants are unique in that they contain intact, functional peroxisomes, unlike previously described mutants (15), and one of these mutants bears an as yet undescribed lesion in the biosynthesis of plasmenylethanolamine, a deficiency in A1'-desaturase activity. Using these mutants, we can evaluate the role of plasmalogens in cellular function without considering the loss of peroxisomes as a contributing factor. In this report, we describe the selection and biochemical characterization of these mutant cells. '"PIATP were obtained from Amersham and Du Pont-New England Nuclear, respectively. Dihydroxyacetone [32P]phosphate and glycerol-3-[:"P]phosphate were synthesized by enzymatic phosphorylation of dihydroxyacetone or glycerol using [Y-~'P]ATP and glycerol kinase (15,16 Phospholipid Standards-Plasmenylethanolamine from bovine brain was purchased from Sigma and was actually a mixture of plasmenylethanolamine (60%) and phosphatidylethanolamine (40%). T o produce lysoplasmenylethanolamine (l-alk-l'-enyl-2-lyso-snglycero-3-phosphoethanolamine), the brain lipids were subjected to mild base hydrolysis (0.5 N NaOH in 70% ethanol at 60 "C for 2 h) and extracted with chloroform after neutralization of this mixture. Lysoplasmenylethanolamine was purified by thin layer chromatography using Silica Gel H and chloroform: methanol:HPO (65:25:10) as the solvent system. Plasmanylethanolamine was produced by PtOz hydrogenation of the brain phosphatidylethanolamine mixture to reduce the vinyl ether double bond (19). Again, this was heavily contaminated with phosphatidylethanolamine. Lysoplasmanylethanolamine could be purified from this mixture by mild base hydrolysis, followed by TLC as described above for the purification of lysoplasmenylethanolamine. Egg yolk phosphatidylethanolamine (Sigma) was used as the diacyl species standard.
Cells and Culture Conditions-CHO.Kl (CCL61) and macrophagelike cell line, RAW 264.7 (ATCC TIB71), were obtained from the American Type Culture Collection. ZR-82 is a peroxisome/plasmalogen-deficient derivative of CHO.Kl (15). RAW cells were maintained in suspension culture at 37 "C in a 5% C02/95% air atmosphere, in Ham's F-12 medium containing 10% fetal bovine serum (Hazelton Laboratories) supplemented with 1 mM glutamine, penicillin G (100 units/ml), and streptomycin (75 units/ml). For all experiments, selections, and mutageneses, RAW cells were plated out in tissue culture dishes (to which they adhered). CHO cells were maintained under identical conditions as adherent monolayers on tissue culture plates for both growth and experiments. Mutageneses were performed using ethyl methanesulfonate as described (15).
Selection of Ether Lipid-deficient Mutants-Mutants were isolated from RAW 264.7 cells after three rounds of selection using the P90H/ UV selection technique (20). Mutagenized RAW 264.7 cells (5 X 10') were plated out in a 100-mm diameter tissue culture plate in 10 ml of medium and allowed to attach overnight. The following day, 5 ml of medium containing 30 p~ P90H was added (final concentration of 10 p M P90H), the cells were incubated at 37 "C for 20 h, and the P9OH-containing medium was then removed and replaced with 15 ml of P9OH-free medium. The cells were incubated at 37 "C for another 5 h and then irradiated for 5 min with long-wavelength ultraviolet light. Irradiation was accomplished by placing the dishes on a 1.5-mm thick glass plate which was suspended over a light source. The distance between the light source (Black-Ray UV lamp Model XX-15L; UVP, Inc., San Gabriel, CA) and the cells was adjusted to obtain an intensity of 2200 microwatts/cm2 at the surface of the tissue culture dish. The cells were allowed to grow out at 37 "C for 7-10 days after irradiation. The survivors from four plates were combined and placed through a second and third round of P9OH/UV selection. In these subsequent rounds of selection, fewer cells were treated (lo4 cells). On the third round of P9OH/UV selection, no cell killing was observed. Isolates from the P9OH/UV-resistant population were obtained by limiting dilution.
PSOH/UV Sensitivity Assays-Cells were plated out in 24-well tissue culture plates at a density of 500 cells/well in 0.5 ml of medium. After adjustment at 37 "C overnight, 0.25 ml of medium containing 3 times the final desired P90H concentration was added, and the cells were placed through the same PSOH/UV selection protocol as described above, with the exception that the final concentration of P90H was varied. The colonies which developed from the surviving cells were visualized after 7 days by staining with 0.1% Coomassie Brilliant Blue in methano1:HzO:acetic acid (45:45:10).
Enzymatic Assays-DHAP acyltransferase, glycerol-3-phosphate acyltransferase, alkyl-DHAP synthase, and alkyl/acyl-DHAP reductase activities were all measured in whole cell homogenates. Cells were grown as suspension cultures to near confluence, harvested by centrifugation for 7 min at 600 X g, and washed once with Trisbuffered saline (Sigma). The cells were resuspended in 0.05 M Tris-HCl, pH 7.4, and frozen at -80 "C. Cell suspensions were thawed on ice and resuspended using a Teflon-glass tissue homogenizer. Peroxisomal and microsomal DHAP acyltransferase and glycerol-3-phosphate acyltransferase activities were measured as described by Jones and Hajra (21). Alkyl-DHAP synthase was assayed as described by Davis and Hajra (22). 1-Acyl-DHAP/l-alkyl-DHAP reductase was assayed by the method of LaBelle and Hajra (23). Protein content was determined using the method of Lowry et al. (24).
Assay of A1 '-Desaturase Activity-The All-desaturase assays were performed on postnuclear supernatants. Cells were harvested from suspension culture, pelleted by centrifugation for 7 min at 600 X g, and resuspended in 0.1 M Tris-HC1, pH 7.4. The cell suspension was sonicated with a microprobe (Branson Sonic Power Co.) while on ice for 30 s (using 3-10-s bursts at a setting of 4). The lysate was centrifuged for 40 s, at 12,000 X g, in a Beckman Microfuge, and the supernatant was used for the assays. The A1'-desaturase assays were performed as described by Blank et al. (19). The assay mixture consisted of 0.1 M Tris-HC1, pH 7.4,2 mM NADH, 10 units of catalase, supernatant protein, and [ 1-3H]ethanolamine-labeled 1-alkyl-2-lysoglycero-3-phosphoethanolamine (10' cpm) in a total volume of 1 ml. Reactions were started by the addition of NADH. After 30 min at 37 "C, the reaction was stopped by the addition of 2.5 ml of methanol. 2.5 ml of chloroform and 1.0 ml of PBS were added to form a 2-phase Bligh and Dyer system (25). The lower organic phase was removed, the upper phase was washed once with an equal volume of preequilibrated lower phase, and the 2 lower phases were combined. This was dried under NP and developed on TLC using Silica Gel 60 and the double development, single dimension system described by Morand et al. (26). Briefly, samples were developed for 5 cm in chloroform:methanol:acetic acid:water (25:15:3:1.5). This separated the lyso-substrate from the acylated products. After drying for 5 min, the plate was sprayed with 10 mM HgC12 in acetic acid, cleaving any vinyl ether double bonds which may have been formed (27). After drying the plate for 20 min in a fume hood, the plate was redeveloped in the same dimension for 20 cm using the same solvent system to separate newly generated 1-lyso-2-acyl product. The radioactive products were localized by fluorography at -80 "C for 6 days after spraying the TLC plates with EN3HANCE (Du Pont-New England Nuclear). The bands of interest were scraped into liquid scintillation vials containing 1 ml of methanol, and the radioactivity was measured after the addition of 8 ml of scintillation mixture.
Preparation of Tritiated 1 -Alkyl-2-lyso-sn-glycero-3-phosphoethanolamine-RAW.12 cells were plated out as a suspension in a 60-mm diameter Petri dish at a density of 5 X lo6 cells/dish in 5 ml of medium containing 20 p~ 1-hexadecylglycerol and 100 pCi of [1-3HJ ethanolamine. Under these conditions, the RAW.12 cells accumulate plasmanylethanolamine (see Fig. 5). After 20 h at 37 "C, the cells by guest on March 22, 2020 http://www.jbc.org/ Downloaded from were harvested, pelleted, and resuspended in 0.8 ml of PBS. This was added to 3 ml of ch1oroform:methanol (1:2). After 10 min a t room temperature, 1 ml of chloroform and 1 ml of PBS were added to form 2 phases, and the lower phase was recovered after centrifugation. The upper phase was washed once with 2 ml of chloroform, and the 2 lower phases were combined and dried under NP. The lipids were applied to TLC, and the plate was sprayed with HgC1, to cleave any vinyl ether groups prior to development. The intact ethanolaminelinked phospholipids (plasmanylethanolamine and phosphatidylethanolamine) were isolated by development of the lipids on Silica Gel H in ch1oroform:methanol:acetic acidH20 (25:15:3:1.5) and recovered from the silica gel using ch1oroform:methanol (1:l) after visualization of the phospholipid bands (by spraying the plate with water). The sample was then hydrolyzed by incubation in 0.1 N NaOH in chloroform:methanol (1:4) for 1 h a t 40 "C, completely deacylating phosphatidylethanolamine and removing the acyl group from the 2-position of the plasmanylethanolamine. After neutralization of the hydrolysis mixture, the resulting 1-alkyl-2-lyso-sn-glycero-3phosphoethanolamine was extracted by the method of Bligh and Dyer (25) purified by development on Silica Gel H using chloroform:methanol:acetic acid:H20 (25:15:3:1.5) and recovered from the plate using ch1oroform:methanol (1:l). Approximately 8 pCi of this substrate could be recovered from a single preparation (using 5 X lo6 RAW.12 cells) at an approximate activity of 0.4 Ci/mmol.
Catalase Latency in Digitonin-permeabilized Cells-Cells were harvested from near confluent suspension cultures, pelleted by centrifugation, washed once with phosphate-buffered saline (PBS), and resuspended in PBS a t 1.5 x 10' cells/ml. To 0.3-ml aliquots of the cell suspension, 0.3 ml of digitonin in PBS was added. After 10 min a t room temperature, the digitonin-treated cells were centrifuged for 40 s a t 12,000 X g in a Microfuge (Beckman). The supernatant was used immediately for catalase and lactate dehydrogenase activity determinations. Catalase assays were performed by the modified method of Peters et al. (28). Lactate dehydrogenase was assayed as described previously (29). For a 100% release control, 1 aliquot of cells was treated with 0.1% Triton X-100.
Immunofluorescence Microscopy-Cells were grown on coverslips for 48 h, fixed with 4% paraformaldehyde, incubated with primary rabbit antiserum, and subsequently with a fluorescein-conjugated goat anti-rabbit IgG (30). Phorbol 12-myristate 13-acetate (1 p~) was added to the cell cultures for the last 12 h prior to fixation in order t o obtain a better spreading of the cells. The antiserum used was specific for bovine liver catalase (30). Antibodies against catalase were affinity-purified with bovine liver catalase bound to nitrocellulose.
Determinations of Plasmenylethnnolnmine Content-As a rapid, qualitative way to determine the relative plasmenylethanolamine content, RAW cells were grown for 48 h, a t 37 "C, in medium containing [1-'HJethanolamine (0.25 pCi/ml). The medium was removed, the cells were washed once with 2 ml of PBS, scraped in 0.8 ml of PBS, and added to a test tube containing 1 ml of CHCl.? and 2 ml of methanol. Samples were converted to a 2-phase Bligh and Dyer system (25) by the addition of 1 ml of chloroform and 1 ml of PBS. The lower (organic phase) was collected after centrifugation and evaporated to dryness under nitrogen, and phospholipids were separated by two-stage, single dimension, thin layer chromatography (26). Greater than 95% of the label was associated with either the plasmenylethanolamine or the combined phosphatidyl/plasmanylethanolamine bands. When cells were supplemented with 1-0-hexadecylglycerol, it was added to the medium from a 20 mM stock solution in ethanol to a final concentration of 20 @M. The labeled species were visualized by fluorography a t -80 "C after spraying the plates with EN:'HANCE. The bands of interest were scraped directly into scintillation vials containing 1 ml of methanol and counted after the addition of scintillation mixture.
When more quantitative determinations of all three subspecies (plasmenyl-, plasmanyl-, and phosphatidyl-) of either the ethanolamine or choline head group classes were required, the individual head group classes were isolated (from loR cells) using TLC, treated with phospholipase C, and benzoylated (31). The benzoylated subspecies were then separated using normal phase HPLC (31) and quantitated using the absorbance of the benzoyl moiety a t 230 nm. Alternatively, the benzoylated subspecies were separated on TLC using Silica Gel G and a development system consisting of benzene:n-hexane:diethyl ether (50:45:5). The bands corresponding to the three subspecies were visualized by spraying the TLC plate with H20, scraped, and eluted from the silica gel using ch1oroform:methanol (2:l). Samples were dried under N,, resuspended in 2 ml of 100% ethanol, centrifuged to remove any residual silica, and quantitated by measuring the absorbance at 230 nm in a spectrophotometer. This latter method of quantitation gave values comparable to those obtained by HPLC.
Quantitation of Very Long Chain Fatty Acids-Total cellular lipids were extracted according to Bligh and Dyer (25) and transesterified using 2% HzSOn in methanol (32), and the methyl esters were analyzed by capillary gas-liquid chromatography (32,33).
Oxidation of 'H-Labeled Phytanate by Intact Cells-Oxidation of phytanic acid was assayed by measuring the release of water-soluble radioactivity from [2,3-"H]phytanic acid-labeled cells. Cells were grown to confluency in 35-mm diameter culture dishes, and the medium was changed to 2 ml of Dulbecco's modified Eagle's medium containing 10% lipid-free fetal bovine serum 1 day prior to study. 3.0 pCi of [2,3-'H]phytanic acid was added to the medium in 5 p1 of ethanol (final phytanate concentration of 50 nM) in each culture dish, and cells were incubated for 24 h a t 37 "C. An aliquot (0.2 ml) of the radioactive medium was then mixed with 2.5 ml of chloroform:methanol (1:1), 1.25 ml of chloroform, and 0.75 ml of water. The solutions were vortexed for 1 min and centrifuged a t 400 X g for 5 min. The upper (aqueous) phase, containing water-soluble radioactivity, was removed. The lower phase was re-extracted with 2 ml of theoretical upper phase consisting of ch1oroform:methanol:water (15:240235). The aqueous phases were combined, an aliquot (1.0 ml) was transferred to a scintillation vial containing 10 ml of scintillation mixture, and the radioactivity was measured by scintillation spectroscopy. Dishes lacking cells served as controls.

Isolation of a PSOHIUV-resistant Population of RAW
Cells-A method to select for ether lipid-deficient Chinese hamster ovary cell mutants has been described (20). This involves treating cultured cells with a pyrene-labeled long chain fatty alcohol (9-(l'-pyrene)nonanol, P90H) followed by exposure to long-wavelength UV light. The fluorescent compound is taken up by the cells, incorporated into complex lipids as either the fatty alcohol or as the fatty acid (following oxidation by long chain fatty alcoho1:NAD' oxidoreductase). Pyrene-treated cells are killed during irradiation, presumably due to the generation of active oxygen species such as singlet oxygen. Plasmalogen-deficient cells take up less P90H than wild-type cells (20) and therefore are less susceptible to UV irradiation.
Mutagenized populations of wild-type RAW cells (RAW 264.7) were subjected to three rounds of P9OH/UV selection, resulting in the development of populations of P9OH/UVresistant cells (Fig. 2). Wild-type cells were killed using 5 pM P90H, whereas at least a portion of the resistant population survived using 30 PM P90H.
PSOHIUV Mutants Are Reduced in Ether Lipid Content-Thirteen clonal strains were isolated from resistant populations, and each was rapidly screened for its ability to synthesize plasmenylethanolamine by measuring the ratio of incor-   (Table I). In wild-type RAW cells, approximately 60% of the label was found in plasmenylethanolamine while the remainder was in the phosphatidylethanolamine/plasmanylethanolamine band. Relatively less ethanolamine was incorporated into the plasmenyl fraction in all of the PSOH/UV-resistant isolates, although there was a great deal of variation between strains. Three isolates, RAW.7, RAW.12, and RAW.108 were chosen for more detailed analyses. RAW.7 and RAW.12 were from the same PSOH/UV' population, while RAW.108 was isolated from a second mutagenesis stock.
We quantitated the relative mass of the three subspecies of choline and ethanolamine phospholipids in the wild-type and the mutant cell lines (Table 11). In wild-type cells, the ethanolamine phospholipids consisted of 36% plasmenylethanolamine, and there was little, or no, plasmanylethanolamine.  RAW.7 was 50% reduced in plasmenylethanolamine content, while the plasmenyl species in RAW.12 and RAW.108 was more severely reduced (3-6% of the ethanolamine phospholipid was plasmenylethanolamine). In RAW.12, approximately 6% of the ethanolamine phospholipid was found as plasmanylethanolamine. This accumulation of plasmanylethanolamine was interesting since even wild-type cells contained only 2.8% of this subspecies. There was no change in overall ethanolamine-containing phospholipid in the RAW mutants. Instead, the loss of plasmenylethanolamine was associated with an increase in phosphatidylethanolamine levels (not shown). The choline phospholipid fraction in wild-type cells consisted of lesser amounts of ether-linked lipid with plasmanylcholine making up 10% of the choline phospholipids and plasmenylcholine representing only 1% in wild-type cells. RAW.7 and RAW.12 were only partially reduced, while RAW.108 displayed very little of either of the ether lipid subspecies.

Quantitation of phosphatidyl-, plasmanyl-, and plasmenyl-species in ethanolamine and choline phospholipids The phospholipid head group species were isolated, and the individual subspecies were converted to the benzoylated derivatives as described under "Experimental Procedures." The derivatives were separated by high performance liquid chromatography and quantitated by the method of Blank et al. (31). Values represent duplicates and did not vary by more than 3%. Experiments using thin layer chromatography to separate the benzoylated species (see "Experimental Procedures") yielded very similar results. All values represent the percent mass of that head group species.
Peroxisomal DHAP Acyltransferase Activity Is Reduced in the RAW Mutants-We measured the activity of the first three reactions in plasmenylethanolamine biosynthesis in whole cell homogenates. Acylation of DHAP in peroxisomes is the first step in ether lipid biosynthesis (34). We were able to detect two DHAP acyltransferase activities in wild-type RAW cell homogenates; the peroxisomal form (measured at pH 5.5), as well as a DHAP acyltransferase which was active at pH 7.4. This latter activity was similar to that reported for ether lipid-deficient CHO cell mutants (15) and was probably due to the lack of specificity of the microsomal glycerol-3phosphate acyltransferase for acyl acceptor. All three mutants showed severely reduced peroxisomal DHAP acyltransferase activity (Table 111). RAW.7 and RAW.12 displayed a small amount of activity (5%), and no activity could be detected in RAW.108. There was only a moderate reduction (30%) in microsomal activity in all of the mutants (not shown). Glycerol-3-phosphate acyltransferase activity was unaltered in the mutant cells.
The second step in ether lipid biosynthesis is catalyzed by another peroxisomal, membrane-bound enzyme, alkyl DHAP synthase (22). This activity was normal in the whole cell homogenates from all of the RAW mutant strains. Another activity associated with ether lipid biosynthesis, acyl/alkyl DHAP reductase (the third step in the biosynthesis of ether lipids), was also found to be normal in the mutants.
Intact Peroxisomes Exist in the RAW Mutants-The previous use of the PSOH/UV selection technique on populations of CHO cells resulted in the production of mutants in which the primary lesion was peroxisomal deficiency (20). This resulted in the loss of the peroxisomal activities, DHAP acyltransferase and alkyl-DHAP synthase. The fact that the RAW mutants maintained normal levels of alkyl-DHAP synthase suggested that the mutants possessed peroxisomes.
Catalase is found primarily within the peroxisome, but in peroxisome-deficient cells this activity is found in the cytosol (15,35,36). The subcellular distribution of catalase was determined in peroxisome-containing and peroxisome-deficient CHO cell lines, as well as in the RAW strains, using digitonin permeabilization (Fig. 3) (36). Treatment of CHO cells with low levels of digitonin (10 pg/ml) causes disruption of the plasma membrane due to its high cholesterol content, resulting in the permeabilization of the cells and the release of soluble cytosolic proteins, such as lactate dehydrogenase (Panel A). Much higher levels of digitonin (150-300 pg/ml) were required to release soluble peroxisomal proteins, such as catalase, from peroxisome-containing, wild-type CHO cells  (Panel A ) . In a peroxisome-deficient CHO mutant ZR-82 (15), catalase was released from the cells along with lactate dehydrogenase (Panel A ) , indicating that catalase was cytosolic in those cells. In wild-type and all of the mutant RAW strains (Panel B), catalase was released only when high levels of digitonin were used, suggesting that catalase was still localized in the peroxisomes. Also, in studies using classical differential centrifugation techniques (15), the catalase distribution in the RAW mutants was similar to the wild-type strain (data not shown); 80-90% of the catalase activity was associated with subcellular organelles.
The catalase localization was confirmed using immunofluorescent microscopy. In wild-type cells, and all three RAW mutant strains, immunofluorescence micrographs using antibodies against catalase showed a concentration of catalase in discrete organelles, presumably peroxisomes (Fig. 4). Again, this is unlike peroxisome-deficient mutants which displayed no such subcellular concentration of catalase (15,20,37).
The Peroxisomal @-Oxidation System Is Functional in RAW Mutants-The peroxisome has a unique @-oxidation system which is responsible for the breakdown of very long chain fatty acids, such as 26:O. Every peroxisome-deficient tissue or cell line examined to date has displayed elevated levels of very long chain fatty acids, due to the loss of this system (38,39). This is true for peroxisome-deficient CHO mutants (Table  IV) which contained 5-20 times the wild-type levels of 26:O. Very long chain fatty acid levels in the RAW mutants, however, were not elevated in comparison to the wild-type strain (264.7), suggesting a functional peroxisomal @-oxidation system. Phytanic acid oxidation, another peroxisomal function which is defective in peroxisomal-deficient cells, is normal in all the RAW strains (Table IV) presence of functional peroxisomes.
Bypass of the Plasmalogen Deficiency Using I-Hexadecylglycerol-Supplementation of the growth medium with l-hexadecylglycerol(1-HG) bypasses the plasmalogen deficiency in human (40) and CHO (7) mutant cell lines. The 1-HG is phosphorylated by 1-alkylglycerol kinase (34) and enters the biosynthetic pathway downstream of the enzymatic lesions (as l-alkyl-2-lyso-sn-glycerol-3-phosphate). Although 1-HG supplementation restored the plasmenylethanolamine content of RAW.7 (not shown) and RAW.108 (Fig. 5 )    * Values for phytanate oxidation in peroxisome+ and peroxisomecultured cells were obtained from normal human fibroblasts (n = 9) and fibroblasts from patients suffering from Zellweger syndrome (n = 4), respectively. observed in RAW 264.7 cells, the plasmenylethanolamine deficiency could not be fully bypassed in RAW.12 cells. Instead, the immediate biosynthetic precursor of plasmenylethanolamine, plasmanylethanolamine, accumulated. During supplementation with 20 PM I-HG, this latter species represented 55% of the total ethanolamine-containing phospholipid mass (data not shown). Plasmanylethanolamine did not accumulate in RAW.108 or RAW.7, even when supplemented with 1-HG. 12 Is Deficient in A1 '-Desaturuse Activity-The inability to bypass the plasmalogen deficiency with 1-HG and the accumulation of plasmanylethanolamine suggested a loss of A1'-desaturase activity in RAW.12 cells. This enzyme (the final step in plasmenylethanolamine biosynthesis) converts plasmanylethanolamine to plasmenylethanolamine by introducing the cis double bond between the first and second carbon of the alcohol moiety to produce the HgCln-cleavable vinyl ether linkage. When tritium-labeled 1-alkyl-2-lyso-snglycero-3-phosphoethanolamine was incubated with a postnuclear supernatant from RAW 264.7, two products were formed (Fig. 6). These co-migrated with standards plasmanylethanolamine (acylated only; band 3 ) and plasmenylethanolamine (acylated and desaturated; band 2). This profile of product formation is similar to that reported for A1'-desaturase assays using microsomes from other cultured cell lines (19). Postnuclear supernatants from RAW.12 cells were able to acylate the substrate (to form plasmanylethanolamine), but they were not able to desaturate (form the vinyl ether double bond). This activity was reduced by at least 90% in RAW.12 (Fig. 7). When equivalent amounts of protein from wild-type and RAW.12 supernatants were mixed and used for these assays (not shown), intermediate activity was observed suggesting that no soluble, trans-acting inhibitor was present in the mutants.

DISCUSSION
Although several functions have been proposed for plasmalogens, a definitive proof for any of these has not been forthcoming. We've chosen to isolate animal cell lines that are deficient in plasmalogens and determine if any cellular processes are affected. Previous attempts, using the Chinese hamster ovary cell line, CHO.K1 (15,20,35), have resulted in the isolation of ether lipid-deficient mutants, but the primary lesion has been the loss of peroxisomes. In addition, all   7. Production of plasmalogen (Bands 1 and 2 in Fig.  6 ) by RAW 264.7 and RAW.12 lysates. Assays were performed and products were isolated as described in Fig. 6. Bands of interest were scraped from the TLC plate into scintillation vials containing 1 ml of methanol followed by 10 ml of scintillation mixture and counted. All values represent a combined value from the number of counts found in both Bands 1 and 2. Open symbols, RAW 264.7; closed symbols, RAW.12.
of the human genetic disorders which involve plasmalogen deficiency also involve a deficiency in peroxisomes and/or peroxisomal functions such as peroxisomal @-oxidation and phytanic acid oxidation (41). Interpretation of results with respect to plasmalogen or ether lipid function using fibroblasts from these patients or the CHO cell mutants is complicated by the possibility that the loss of the peroxisomes or peroxisomal function may be a contributing factor. Unlike any of the PSOH/UV-resistant CHO strains, the RAW mutants possessed intact, functional peroxisomes as indicated by: 1) the latency of catalase in digitonin permeabilization studies, 2) the localization of catalase using immunofluorescence microscopy, 3) the normal levels of very long chain fatty acids, and 4) normal phytanic acid oxidation rates.
The only biochemical lesion associated with plasmalogen biosynthesis in RAW.7 and RAW.108 was a loss of peroxisomal DHAP acyltransferase activity. The enzymes catalyzing the next two steps in plasmalogen biosynthesis (alkyl-DHAP synthase and alkyl-DHAP reductase) were active. Also, supplementation of the growth medium with 1-hexadecylglycerol (which enters the pathways just after the reduction of l-alkyl-DHAP) restored normal plasmenylethanolamine levels, indicating that subsequent steps are not affected. The lesion in RAW.108 appears to be quite stringent, in that there is little residual ether-linked ethanolamine or choline phospholipid, and the peroxisomal DHAP acyltransferase was virtually undetectable in our assays. RAW.7 cells were leaky with respect to plasmalogen biosynthesis. This may have been the result of residual peroxisomal DHAP acyltransferase activity.
RAW.12 displayed a deficiency in 2 steps of the pathway. Like the other two mutants, peroxisomal DHAP acyltransferase activity is greatly reduced however, 1-alkylglycerol supplementation only partially restored plasmenylethanolamine levels. Instead, the cells accumulated its immediate precursor, plasmanylethanolamine. Even without 1-hexadecylglycerol supplementation, there is a higher than normal level of this lipid species. These data, and the inability of RAW.12 lysates to convert this plasmanyl-intermediate to a HgC12-sensitive lipid in the presence of NADH, strongly suggest that RAW.12 is deficient in A1'-desaturase activity. Although the lesion appears to be quite severe (5% of wild-type activity), the presence of increased amounts of endogenous plasmanyle-thanolamine in the RAW.12 membranes makes an accurate determination of specific activities impossible using these assay conditions. However, since plasmanylethanolamine levels in RAW.12 cells are only 2-3-fold higher than wild-type cells (Table 11), endogenous substrate levels cannot account for the 20-fold decrease in A1'-desaturase activity. The fact that plasmanylethanolamine accumulates in cells which are restricted in an early step of plasmenylethanolamine biosynthesis (peroxisomal DHAP acyltransferase) suggests that the A1'-desaturase deficiency is quite severe. Further kinetic analysis using detergent-solubilized systems should aid in this determination. Also, the A1'-desaturase system is a membrane-bound, multicomponent, system consisting of an electron transport component (cytochrome bs and cytochrome bs reductase) and a terminal, cyanide-sensitive protein (42). Further analysis is required to determine if any of these components are defective, and whether other enzyme systems, which utilize the cytochrome bs-cytochrome b6 reductase electron transport system (e.g., the stearoyl-CoA desaturase system; Ref. 43) are affected.
Presently, we do not know whether the two lesions associated with RAW.12 are due to the alteration in one gene or two. A deficiency in A l ' desaturase activity should not prevent the incorporation of fatty alcohol into ether-linked ethanolamine phospholipids, and, therefore, the PSOH/UV protocol should not select for such mutants. It seems more likely that the loss of DHAP acyltransferase and A l ' desaturase activities are due to the alteration in one genetic locus which affects both activities. We are also unable to explain why the P90H/ UV selection technique generates peroxisome-deficient CHO mutants (20) while the first three PSOH/UV-resistant RAW mutants that we have examined contain functional peroxisomes.
In summary, we have described the isolation of somatic cell mutants which display at least two unique lesions with respect to plasmalogen biosynthesis, demonstrating that the P90H/ UV selection technique can be used to isolate mutants with defects specifically in the plasmalogen biosynthesis. Rigorous examination of other PSOH/UV-resistant cell lines, especially conditionally lethal mutants, may yield mutations in other steps of plasmalogen biosynthesis. Some of these activities are shared with diacylphospholipid biosynthetic pathway.