Genomic organization of the hamster phospholipase C-delta 1 gene: differential loss of separate alleles of the phospholipase C-delta 1 gene in two fibroblast mutants lacking phospholipase C-delta 1.

In earlier work from this laboratory, mutant Chinese hamster lung fibroblasts D1-6b and D1-9b were shown to lack phosphatidylinositol-specific phospholipase C-delta 1 (PLC delta 1). In the current study, Northern analysis of mRNA samples from these mutants fails to detect full-length transcripts for PLC delta 1. Southern analysis of Sac1 digests of wild type and mutant cell genomic DNA hybridized against probe made to full-length rat brain PLC delta 1 cDNA suggested that these mutants differ from one another and wild type controls. This observation is shown to be the result of the differential loss, in both mutants, of one of two alleles for the hamster PLC delta 1 gene normally present in the parental cell line. During the course of these studies, the hamster PLC delta 1 gene was cloned and sequenced. This genomic sequence spans 13 kb of DNA and encoded 98.5% of the predicted coding sequence for the hamster PLC delta 1 cDNA within 15+ exons.

A Chinese hamster fibroblast mutant defective in thrombin-induced signaling has a low level of phospholipase C-beta 1.

A mutant fibroblast, 2A4b, was isolated from the Chinese hamster lung cell line CCL39 by a previously described selection (Rath, H. M., Doyle, G. A. R., and Silbert, D. F. (1989) J. Biol. Chem. 264, 13387-13390) for cells deficient in thrombin-induced signaling. Although the antiporter activation by thrombin in 2A4b is only approximately 60% that in CCL39, the stimulation by serum is not significantly impaired, indicating that the defect in 2A4b lies upstream of the antiporter in the signaling pathway. The addition of thrombin to serum-starved 2A4b cells causes blunted responses both in production of inositol phosphates and in the cytosolic [Ca2+] transient, particularly when no Ca2+ is added to the external medium. The in vitro inositol phospholipid-specific phospholipase C (PLC) activity of 2A4b cytosol plus membrane extracts exceeds that in CCL39. However, immunoblots with antibodies to PLC isozymes show that although the levels of PLC-delta 1, PLC-gamma 1, and PLC-beta 3 are at least as great as those in CCL39, the amount of PLC-beta 1 in 2A4b is markedly deficient (< or = 10%). PLC-beta 1 is found primarily in the nucleus and in non-nuclear membranes of CCL39 and is proportionately low in these subcellular locations of 2A4b. Thrombin activation of phospholipases D and A2 is impaired in 2A4b. We postulate that the deficiency in PLC-beta 1 causes defective targeting of protein kinase C-alpha to specific membrane sites, which may be required for activation of these downstream phospholipases.

Characterization of a second hamster lung fibroblast mutant with defects in phosphatidylinositol-specific phospholipase C.

We report here further characterization of the Chinese hamster lung fibroblast mutant D1-9b displaying elevated agonist-induced phosphatidylinositol (PI) turnover responses relative to CCL39, the parental cell line (Rath, H. M., Doyle, G. A. R., and Silbert, D. F. (1989) J. Biol. Chem. 264, 13387-13390). These differences in PI turnover responses are further enhanced by long term (24 h) pretreatment of cells with the phorbol ester 4 beta-phorbol 12 beta-myristate 13 alpha-acetate (PMA). Short term pretreatment (15 min) with PMA attenuates the agonist-induced PI turnover response of both wild type and mutant D1-9b cell lines to that of unstimulated controls, suggesting that PMA-induced desensitization responses in D1-9b are intact. Both cytosolic and membrane/particulate fractions prepared from mutant D1-9b have decreased phospholipase C (PLC) activities relative to wild type when assayed with either of four different PI/phosphatidylinositol 4,5-bisphosphate-mixed micelle/vesicle substrate combinations. Thermolability studies, Mono Q anion-exchange chromatography, and Western blot studies identified the source of cytosolic PLC deficiency in D1-9b as being due to the absence of PLC delta 1, one of at least two PLC isozymes previously shown to be in wild type CCL39 cytosol (Rath, H. M., Fee, J. A., Rhee, S. G., and Silbert, D. F. (1990) J. Biol. Chem. 265, 3080-3087). We also report here the presence of two peaks of PLC activities (i.e. peaks mA and mB) following Mono Q chromatography of wild type CCL39 membrane/particulate extracts; membrane/particulate extracts prepared from mutant D1-9b are missing the first of these two peaks of activities (peak mA). Immunoblot analysis confirms the presence of PLC beta 1, but not PLC gamma 1 or PLC delta 1, in peak mB. Comparisons are made between D1-9b and the previously characterized mutant D1-6b, which displays diminished agonist-induced PI turnover responses yet in vitro biochemical deficiencies similar to those of mutant D1-9b. Somatic cell hybridization experiments suggest that the defects found in mutant D1-9b are codominant relative to wild type phenotype and that mutant D1-9b belongs to the same genetic complementation group as mutant D1-6b. Possible explanations to explain these different agonist-induced responses in light of the two mutants similar in vitro biochemical deficiencies are addressed.

Utilization of exogenously supplied sphingosine analogues for sphingolipid biosynthesis in Chinese hamster ovary and mouse LM cell fibroblasts.

Growth of Chinese hamster ovary and LM cells is inhibited by relatively low concentrations of sphingosine in the culture media. This effect is diminished by an order of magnitude by conversion of this positively charged long-chain amino alcohol to a number of N-acetylated analogues, such as N-acetylsphingosine, N-acetylsphingosine phosphate, and N-acetylsphingosine phosphorylcholine. Synthesis of sphinganine and its incorporation into ceramide, sphingomyelin, and glycosphingolipids (GSL) was monitored using a short pulse of [14C]serine together with a long pulse of [3H]-galactose. Compared to unsupplemented cultures, growth with 15-30 microM N-acetylsphingosine suppressed incorporation of 14C radioactivity into ceramide, sphingomyelin, and GSL by 75-95% without accumulation of labeled sphinganine and without any appreciable change in membrane phospholipid or total GSL content. Furthermore, when cells were cultivated with 15 microM [4,5-3H]N-acetylsphinganine to monitor its utilization for sphingolipid synthesis, considerable loss of radiolabel occurred due to desaturation of sphinganine to sphingosine. Nevertheless, most of the residual label was found in the long-chain base and not the acyl group of sphingomyelin, indicating that the exogenously supplied base was utilized intact for complex lipid synthesis. Radiolabel was also found in ceramide and glycosphingolipid fractions. Thus, established cell lines whose growth is very sensitive to long-chain amino alcohols can be cultivated with sphinganine (sphingosine) analogues at concentrations which suppress endogenous sphinganine production but support continued synthesis of complex sphingolipids.

PI-specific phospholipase C "alpha" from sheep seminal vesicles is a proteolytic fragment of PI-PLC delta.

Phosphatidylinositide-specific phospholipase C enzymes (PLCs) catalyze the conversion of the phosphoinositides to biologically important signal transducing molecules. These enzymes may be grouped into "families" which share similar structures and modes of regulation. The existence of a structurally distinct family of PLC termed "alpha" has been recently called into question. In the current paper we show by immunoblotting experiments that PLC "alpha" from sheep seminal vesicles is recognized by monoclonal antibodies raised against the delta 1 isoform of bovine brain PLC, and appears to be derived from a higher molecular weight band at 85 kDa. We also show that antibodies raised against PLC alpha efficiently immunoprecipitate the 85-kDa PLC delta 1 isoform from bovine brain and Chinese hamster lung fibroblasts. These data provide strong evidence that the PLC alpha from sheep seminal vesicles is a proteolytic fragment of PLC delta 1. Thus, there is still no conclusive evidence for a separate "alpha" class of PLC.

Deuterium NMR of Val1...(2-2H)Ala3...gramicidin A in oriented DMPC bilayers.

Deuterium NMR is used to study the selectively labeled Val1...(2-2H)Ala3...gramicidin A molecule to investigate the structure and dynamics of the C alpha-2H bond in the Ala3 residue of gramicidin. Val1...(2-2H)Ala3...gramicidin A is synthesized, purified, and characterized and then incorporated into oriented bilayers of dimyristoylphosphatidylcholine sandwiched between glass coverslips. Phosphorus NMR line shapes obtained from this sample are consistent with the presence of the bilayer phase and indicate that no nonbilayer phases are present in significant amounts. Deuterium NMR line shapes obtained from this sample indicate that the motional axis of the gramicidin Ala3 residue is parallel to the coverslip normal, that the distribution of motional axis orientations has a width of 2 degrees, and that only one major conformational and dynamical state of the Ala3 C alpha-2H bond is observed on the NMR time scale. Furthermore, the Ala3 C alpha-2H bond angle relative to the motional axis is 19-20 degrees if fast axial rotation is assumed to be the only motion present but is less than or equal to 19-20 degrees in the absence of such an assumption. This result indicates that various double-stranded, helical dimer models are very unlikely to represent the structure of gramicidin in the sample studied but that the single-stranded, beta 6.3 helical dimer models are consistent with the experimental data. However, a definitive distinction between the left-handed, single-stranded, beta 6.3 helical dimer model and the right-handed, single-stranded, beta 6.3 helical dimer model cannot be made on the basis of the experimental data obtained in this study.

Deuterium NMR of 2HCO-Val1...gramicidin A and 2HCO-Val1-D-Leu2...gramicidin A in oriented DMPC bilayers.

Deuterium NMR is used to study the structure and dynamics of the formyl C-2H bond in selectively deuterated gramicidin molecules. Specifically, the functionally different analogues 2HCO-Val1...gramicidin A and 2HCO-Val1-D-Leu2...gramicidin A are studied by 2H NMR so that any conformational or dynamical differences between the two analogues can be correlated with their difference in lifetime. These analogues are first synthesized, purified, and characterized and then incorporated into oriented bilayers of dimyristoylphosphatidylcholine sandwiched between glass coverslips. Phosphorous NMR line shapes obtained from these samples are consistent with the presence of the bilayer phase and indicate that the disorder exhibited by the lipid matrix is approximately of the same type and degree for both analogues. Deuterium NMR line shapes obtained from these samples indicate that the motional axis of the formyl group of gramicidin is parallel to the coverslip normal, that the distribution of motional axis orientations has a width of 7-9 degrees, and that a similar, major conformational and dynamical state exists for the formyl C-2H bond of both analogues. In this state, if the only motion present is fast axial rotation, then the experimentally derived angle between the formyl C-2H bond and the motional axis is consistent with the presence of a right-handed, single-stranded, beta 6.3 helical dimer but is not consistent with the presence of a left-handed, single-stranded, beta 6.3 helical dimer. However, if fast axial rotation is not the only motion present, then the left-handed, single-stranded, beta 6.3 helical dimer cannot be absolutely excluded as a possibility. Also, a second, minor conformational and dynamical state appears to be present in the spectrum of 2HCO-Val1-D-Leu2...gramicidin A but is not observed in the spectrum of 2HCO-Val1...gramicidin A. This minor conformational and dynamical state may reflect the presence of monomers, while the major conformational and dynamical state may reflect the presence of dimers.

Characterization of phosphatidylinositol-specific phospholipase C defects associated with thrombin-induced mitogenesis.

In a previous paper (Rath, H. M., Doyle, G. A. R., and Silbert, D. F. (1989) J. Biol. Chem. 264, 13387-13390), we reported a selection for the isolation of Chinese hamster lung fibroblasts (CCL39) defective in thrombin-induced mitogenesis. One mutant, D1-6b, had decreased production of inositol phosphates when challenged with activators of phosphatidylinositol turnover and extracts of this mutant showed a marked decrease in phospholipase C (PLC) activity toward phosphatidylinositol. In the current studies, the PLC activities of wild type CCL39 and D1-6b cytosolic extracts are further characterized. Wild type cytosol had at least two phosphatidylinositol-specific PLC isoenzymes, which could be separated by anion exchange chromatography and behaved differently in thermal inactivation studies. Since gel filtration of PLC activity in wild type extracts gave Mr values similar to that of previously characterized PLCs (140,000-200,000), immunoblots with antibodies to bovine brain isoenzymes were used to show that the PLC activities obtained by anion exchange chromatography were PLC-delta and PLC-gamma. Immunoblots with mutant D1-6b cytosol confirmed the presence of the PLC-gamma but showed no detectable PLC-delta. This activity in the mutant extracts eluted at the same conductivity on anion exchange columns and had the same kinetics of thermal inactivation as the PLC-gamma found in the wild type extracts. PLC-gamma from mutant extracts was active in assays containing phospholipid detergent mixed micelles but not in assays utilizing phospholipid vesicles, in sharp contrast to PLC-gamma from CCL39 extracts, which was active under either condition. Thus, the phosphatidylinositol-specific phospholipase C activity of mutant D1-6b is diminished both by the loss of PLC-delta and by the compromised behavior of PLC-gamma.

Hamster fibroblasts defective in thrombin-induced mitogenesis. A selection for mutants in phosphatidylinositol metabolism and other functions.

Growth of Chinese hamster lung fibroblasts (CCL39) on thrombin as sole mitogen is dependent on phosphatidylinositol (PI) metabolism and activation of the Na+/H+ antiporter. By modifying a H+ suicide selection developed for the isolation of antiporter mutants in these cells, we enriched for and isolated CCL39 variants deficient in the thrombin mitogenic response (thrombin nongrowers). These mutants retain alternate mitogenic mechanisms and, hence, grow well on media containing serum. When challenged with thrombin, the mutants show decreased, increased, or unchanged levels of inositol phosphates produced as compared with wild type cells. One of the mutants (D1-6b) has decreased inositol phosphates production not only with thrombin but also with serotonin (5-hydroxytryptamine) and AlF4-, suggesting a defect distal to the thrombin receptors. Extracts of this mutant reveal marked decreased phospholipase C activity toward PI. From the different phenotypes of the thrombin nongrowers, it is clear that the selection is general and that mutants with various biochemical defects should lead to a better understanding of the PI cycle as well as of functions essential to mitogenesis.

LM cell growth and membrane lipid adaptation to sterol structure.

Using a sterol auxotroph of the LM cell mouse fibroblast, we demonstrate that relatively few cholesterol analogues can substitute for cholesterol as a growth factor. The auxotroph grows normally on desmosterol and trans-22-dehydrocholesterol and at reduced rates on dihydrocholesterol, campesterol, and 22,23-dihydrobrassicasterol. It does not grow with beta-sitosterol, stigmasterol, ergosterol, or cis-22-dehydrocholesterol when the sterol is present as sole supplement but does grow at normal rates when the analogue is supplied with suboptimal amounts of cholesterol. Two contrasting types of membrane lipid changes are observed in cells grown on cholesterol analogues. In cells grown with dihydrocholesterol, a marked increase in desaturation and elongation of fatty acids is noted. Conversely, when cells are grown with cis-22-dehydrocholesterol, desaturation and elongation of fatty acids are severely curtailed. Cells grown on alkyl sterols respond like cells grown on cis-22-dehydrocholesterol but in a less pronounced fashion. The effects of sterol substitution in mammalian cells versus in lower eukaryotes are compared, and an explanation for the secondary changes in fatty acid composition in terms of phospholipid phase behavior is suggested.

Influence of sterol structure on phospholipid phase behavior as detected by parinaric acid fluorescence spectroscopy.

Phospholipid-sterol interactions were investigated using parinaric acid fluorescence spectroscopy. Cholesterol and cholesterol analogues which were modified in the sterol nucleus or side chain were added at 50 mol % to multilamellar vesicles of model phospholipids selected to be representative of major components in an LM cell plasma membrane. These included sphingomyelins and saturated and monounsaturated phosphatidylcholines and phosphatidylethanolamines. Based on the changes in cis-parinaric acid steady-state fluorescence polarization observed with addition of sterol, 50 mol % cholesterol abolished the phase transition of all the model phospholipids. Dihydrocholesterol and trans-22-dehydrocholesterol behaved like cholesterol in the two systems studied. 24-Methylcholesterols interacted well with all phospholipids except phosphatidylethanolamine which contained an unsaturated fatty acid. 24-Alkyl,trans-22-dehydrocholesterols abolished the phase transition in only two systems: sphingomyelins and phosphatidylcholines possessing relatively short saturated acyl chains. Since steady-state anisotropy is a function of fluorescence lifetime, rotational diffusion rates, and limiting anisotropy, we determined these parameters for two of the phospholipid systems. The results show that steady-state anisotropy values for phospholipid-sterol interactions correlate closely with limiting anisotropy and to a lesser extent with rotational relaxation time. The behavior of the sterols in the model phospholipids are used to interpret 1) fluorescence polarization measurements made with phospholipids extracted from LM cell plasma membranes, and 2) changes in membrane lipid composition which accompany growth of LM cells on various sterols.

Effect of sterol structure on the partition of sterol between phospholipid vesicles of different composition.

The partition of cholesterol analogues between dipalmitoylphosphatidylcholine and egg phosphatidylcholine vesicles was examined. Cholesterol, trans- and cis-22-dehydrocholesterols, and 24 alpha-ethyl,trans-22-dehydrocholesterol (stigmasterol) showed a preference from gel phase dipalmitoylphosphatidylcholine over fluid phase egg phosphatidylcholine at 37 degrees C. Within this group, the sterol concentration in DPPC relative to that in egg PC ranged from about 1.5 to 2.0. Cholesterol analogues with a 24 alpha-methyl or ethyl substituent (campesterol and beta-sitosterol, respectively) and cholestanol (dihydrocholesterol) distributed about equally between the two types of phospholipid. Thus, in this study involving two kinds of phospholipid and a small number of cholesterol analogues, there was no simple correlation between the sterol structure and its partition behavior. The combined results from studies on sterol partition behavior and on sterol interaction with individual phospholipids (Rujanavech, C., Henderson, P.A., and Silbert, D.F. (1986) J. Biol. Chem. 261, 7204-7214) provide an adequate basis to explain the different patterns of membrane lipid adaptation which accompany growth of LM cells on various cholesterol analogues (Rujanavech, C., and Silbert, D.F. (1986) J. Biol. Chem. 261, 7196-7203).

Conditions influencing formation of 16:0/16:0 molecular species in membrane phospholipids of Escherichia coli.

Growth of a beta-oxidation-negative (fadE) strain of Escherichia coli in liquid culture with exogenous palmitate leads to the accumulation of 16:0/16:0 molecular species of phospholipid resulting in a temporary decrease in growth rate and lysis of a variable fraction of the cell population. Under the same growth conditions, this behavior is not seen in the closely related fadE+ parent strain which accumulates more modest levels of 16:0/16:0 together with 16:0/14:0. Based on differential scanning calorimetric measurements, 75.8 and 17.5% of the lipids from 16:0-enriched fadE and fadE+ cells, respectively, were found to be in the gel state at the growth temperature. Kinetic studies reveal the translocation of 16:0/16:0 molecular species from inner to outer membrane delays briefly the accumulation of this species in the inner membrane. This extreme and deleterious change in membrane-lipid composition precludes cloning of the fadE strain on solid media containing 16:0 and, therefore, these conditions provide the basis for selection of mutants altered in the mechanisms which determine the synthesis or accumulation of membrane lipid. Three such mutants are described which display alterations in the normal distribution of molecular species.

Altered acyltransferase activity in Escherichia coli associated with mutations in acyl coenzyme A synthetase.

Growth of a temperature-sensitive general fatty acid synthesis mutant of Escherichia coli K12 at its restrictive temperature in the presence of exogenous palmitate results in lysis of the bacterium. Under these conditions, palmitate is incorporated into membrane phospholipid to a high level. Mutants of bacteria restricting this incorporation (having a palmitate-resistant phenotype) have been isolated and one such mutant, strain L8-2/3, has been further characterized. This mutant has lowered acyl-CoA synthetase (fadD) activity (25-33% of normal) and consequently is defective in fatty acid uptake. This lowered uptake could explain the palmitate-resistant phenotype of strain L8-2/3. However, both in vivo (fatty acid composition and positional distribution data) and in vitro (acyltransferase activity measurements) experiments suggest that this mutant is also altered in its acyltransferase activities. The mutation(s) of strain L8-2/3 appears to allow increased (approximately 2-fold) incorporation of myristate (and possible unsaturated fatty acids) into position 2 of 1-acyl-sn-glycerol 3-phosphate but normal palmitate incorporation into the same position. The incorporation of palmitate, myristate, and oleate into position 1 of sn-glycerol 3-phosphate by strain L8-2/3 is also higher than that observed with the parent, strain L8-2. Replacing the partially defective fadD gene of strain L8-2/3 with a wild type allele conferred on this strain the palmitate sensitivity and the acyltransferase activity of the parent strain L8-2. This finding, taken together with other data, suggests that acyl-CoA synthetase interacts with the acyltransferase(s) in some manner to influence the fatty acid specificity of the acyltransferase.

Effect of membrane sterol content on the susceptibility of phospholipids to phospholipase A2.

The effects of membrane sterol level on the susceptibility of LM cell plasma membranes to exogenous phospholipases A2 has been investigated. Isolated plasma membranes, containing normal or decreased sterol content, were prepared from mutant LM cell sterol auxotrophs. beta-Bungarotoxin-catalyzed hydrolysis of both endogenous phospholipids and phospholipids introduced into the membranes with beef liver phospholipid exchange proteins was monitored. In both cases, phosphatidylcholine (PC) and phosphatidylethanolamine (PE) were degraded at similar rates in normal membranes, while PC hydrolysis was specifically accelerated in sterol-depleted membranes. Additional data suggest that this preferential hydrolysis of PC is not a consequence of the phospholipid head group specificity of the phospholipase, nor of a difference in the accessibility of PC versus PE to the enzyme. Analysis of the reaction products formed during treatment of isolated membranes with phospholipase A2 showed almost no accumulation of lysophospholipids. This was found to be due to highly active lysophospholipase(s), present in LM cell plasma membranes, acting on the lysophospholipids formed by phospholipase A2 action. A soluble phospholipase A2 was partially purified from LM cells and found to behave as beta-bungarotoxin with regard to membrane sterol content. These results demonstrate that the nature of phospholipid hydrolysis, catalyzed by phospholipase A2, can be significantly affected by membrane lipid composition.

Partition of parinaroyl phospholipid probes between solid and fluid phosphatidylcholine phases.

The partitioning of parinaroyl phospholipid probes between solid- and fluid-phase phospholipid is examined. The immiscible model system dipalmitoylphosphatidylcholine and palmitoyldocosahexaenoylphosphatidylcholine is used. Fluorescence quantum yields and fluorescence polarization are used to calculate Kps/f, the solid to fluid phase partition coefficient of each probe [Sklar, L. A., Miljanich, G. P. & Dratz, E. A. (1979) Biochemistry 18, 1707-1716]. Like free trans-parinaric acid (9,11,13,15-all-trans-octadecatetraenoic acid) 1-palmitoyl-2-trans-parinaroylphosphatidylcholine and 1-palmitoyl-2-tra-parinaroylphosphatidylethanolamine partition strongly into solid phase with mean Kps/fs of 7 +/- 4 and 9 +/- 5, respectively. Like free cis-parinaric acid (9,11,13,15-cis-trans,trans,cis-octadecatetraenoic acid), another group of phospholipid probes prefer fluid phase just slightly or partition equally between fluid and solid phases. The Kps/f of 1-palmitoyl-2-cis-parinaroylphosphatidylcholine is 0.6 +/- 0.2, that of 1-palmitoyl-2-cis-parinaroylphosphatidylethanolamine is 1.0 +/- 0.3, that of 1-oleoyl-2-trans-parinaroylphosphatidylcholine is 0.6 +/- 0.3, and that of 1-oleoyl-2-trans-parinaroylphosphatidylethanolamine is 0.7 +/- 0.3. Two probes partition more strongly into fluid-phase phospholipid. These are 1-oleoyl-2-cis-parinaroylphosphatidylcholine with a Kps/f of 0.2 +/- 0.1 and 1-oleoyl-2-cis-parinaroylphosphatidylethanolamine with a Kps/f of 0.4 +/- 0.4.

Isolation and characterization of filipin-resistant LM cell variants not auxotrophic for sterol.

A series of LM cell variants resistant to filipin, but not auxotrophic for sterol, was isolated by plating mutagenized, filipin-treated cells on soft agar medium containing no sterol. Cloned variants were assayed for growth in the presence and absence of sterol or unsaturated fatty acid. Filipin-resistant clones whose growth rate was unaffected by the addition of sterol to the medium were further analyzed. Variants were cultured in minimal medium, and plasma membranes prepared from these cultures were analyzed for sterol content, phospholipid head group composition, and unsaturated fatty acyl content. All variants examined showed a decrease in membrane sterol in conjunction with an increase in unsaturated fatty acyl chains in the membrane phospholipids. In addition, several variants exhibited alterations in phospholipid head group composition, including changes in the phosphatidylcholine/phosphatidylethanolamine ratio, or a decrease in sphingomyelin content. These observations imply that several different metabolic lesions can give rise to decreased plasma membrane sterol content (and hence to filipin resistance). The range of phospholipid alterations observed in these sterol prototrophs emphasizes the complex interrelationship between membrane sterol and phospholipid structure.

Gel phase phospholipid in the plasma membrane of sterol-depleted mouse LM cells. Analysis by fluorescence polarization and x-ray diffraction.

Fluorescence polarization of cis- and trans-parinaric acid and high angle x-ray diffraction are used to examine the physical effects of in vitro sterol depletion of LM cell plasma membranes. Measurements on lipids isolated from the normal plasma membrane show that most of the lipid is fluid between 5 and 45 degrees C. When no sterol is present, approximately 15% of the phospholipids undergo gel phase formation, detectable by x-ray measurements, in this temperature range. Partial sterol repletion of the isolated phospholipids serves to lower the onset temperature of the phase transition, as well as to decrease the fraction of lipid undergoing a transition. Intact plasma membranes are sterol depleted by in vitro incubation with egg phosphatidylcholine liposomes. This induces a phase transition with an onset temperature increasingly above the physiological temperature as more sterol is removed from the membranes. The transition can be suppressed by returning sterol to the membranes. These experiments extend earlier work (Rintoul, D. A., Chou, S.-M., and Silbert, D. F. (1979) J. Biol. Chem. 254, 10070-10077) on isolated plasma membrane lipids by showing that sterol suppresses a phase transition in intact membranes as well as in isolated lipids. The correlation between gel phase formation upon sterol loss at the physiological temperature and loss of membrane glucose transport function is discussed.