Lateral organization in Acholeplasma laidlawii lipid bilayer models containing endogenous pyrene probes Patrik Storm 1 ,LuLi 2 , Paavo Kinnunen 3,4 and A ˚ ke Wieslander 1 1 Department of Biochemistry and Biophysics, Stockholm University, Stockholm, Sweden; 2 Wallenberg Laboratory for Cardiovascular Research, Go ¨ teborg University, Sweden; 3 Department of Medical Chemistry, Institute of Biomedicine, Helsinki University, Finland; 4 Memphys – Center for Biomembrane Physics, University of Southern Denmark, Odense, Denmark In membranes of the small prokaryote Acholeplasma laid- lawii bilayer- and nonbilayer-prone glycolipids are major species, similar to chloroplast membranes. Enzymes of the glucolipid pathway keep certain important packing proper- ties of the bilayer in vivo, visualized especially as a monolayer curvature stress (Ôspontaneous curvatureÕ). Two key enzymes depend in a cooperative fashion on substantial amounts of the endogenous anionic lipid phosphatidylglycerol (PG) for activity. The lateral organization of five unsaturated A. laidlawii lipids was analyzed in liposome model bilayers with the use of endogenously produced pyrene-lipid probes, and extensive experimental designs. Of all lipids analyzed, PG especially promoted interactions with the precursor diacylglycerol (DAG), as revealed from pyrene excimer ratio (Ie/Im) responses. Significant interactions were also recorded within the major nonbilayer-prone monoglucosylDAG (MGlcDAG) lipids. The anionic precursor phosphatidic acid (PA) was without effects. Hence, a heterogeneous lateral lipid organization was present in these liquid-crystalline bilayers. The MGlcDAG synthase when binding at the PG bilayer interface, decreased acyl chain ordering (increase of membrane free volume) according to a bis-pyrene-lipid probe, but the enzyme did not influence the bulk lateral lipid organization as recorded from DAG or PG probes. It is concluded that the concentration of the substrate DAG by PG is beneficial for the MGlcDAG synthase, but that binding in a proper orientation/conformation seems most important for activity. Keywords: Acholeplasma; chemometrics; lipid heterogeneity; pyrene. Acholeplasma laidlawii A-EF22 is a simple cell-wall-less prokaryotic parasite. Its membrane lipid composition is metabolically adjusted in response to environmental and lipid-supply conditions. Due to this, A. laidlawii has been used as a model system to study plasma membrane properties and how these are maintained by the lipid synthesizing enzymes. Membrane lipids are synthesized in two competing pathways, both using phosphatidic acid (PA) as a precursor, with one branch resulting in glucolipids and the other in phosphatidylglycerol (PG) as shown in the diagram below. At least five enzymes constitute the glucolipid pathway. Phosphatidic acid phosphatase (PAP) makes diacylglycerol (DAG) from PA. 1,2-diacylglycerol-3-glucosyltransferase Correspondence to A ˚ ke Wieslander, Department of Biochemistry and Biophysics, Stockholm University, Stockholm, Sweden. Fax: + 46 8 15 36 79, Tel.: + 46 8 16 24 63, E-mail: ake@dbb.su.se Abbreviations: bis-PyrPC, 1,2-bis-[10-(pyren-1-yl)]decanoyl-sn-glycero-3-phosphatidylcholine; bis-PyrPG, 1,2-bis-[10-(pyren-1-yl)]decanoyl-sn- glycero-3-phospho-rac-glycerol; CL, cardiolipin; 1,2-DOG, 1,2-dioleoylglycerol; DGlcDAG, 1,2-diacyl-3-O-[a- D -glucopyranosyl-(1fi2)-O-a- D -glucopyranosyl]-sn-glycerol; MADGlcDAG, 1,2-diacyl-3-O-[a- D -glucopyranosyl-(1fi2)-O-(6-O-acyl-a- D -glucopyranosyl)]-sn-glycerol; MAMGlcDAG, 1,2-diacyl-3-O-[6-O-acyl(a- D -glucopyranosyl)]-sn-glycerol; MGlcDAG, 1,2-diacyl-3-O-(a- D -glucopyranosyl)-sn-glycerol; PA, phosphatidic acid; PC, phosphatidylcholine; PD, pyrenedecanoic acid; PG, phosphatidylglycerol; PyrDAG, 1-palmioyl-2-pyrenedecanoyl- glycerol; PyrPA, 1-palmioyl-2-pyrenedecanoyl phosphatidic acid; PyrPG, 1-palmioyl-2-pyrenedecanoyl-phosphatidylglycerol. Enzymes: 1,2-diacylglycerol-3-glucosyltransferase (MGlcDAG synthase; EC 2.4.1.157); 1,2-diacylglycerol-3-a-glucose (1fi2)-a-glucosyl transferase (DGlcDAG synthase; EC 2.4.1.208). (Received 30 November 2002, revised 22 January 2003, accepted 19 February 2003) Eur. J. Biochem. 270, 1699–1709 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03527.x (MGlcDAG synthase) (EC 2.4.1.157; I above), makes monoglucosyl diacylglycerol (MGlcDAG) from DAG plus UDP-Glc. 1,2-diacylglycerol-3-a-glucose (1fi2)-a-glucosyl transferase (DGlcDAG synthase) (EC 2.4.1.208), II above, makes diglucosyl diacylglycerol (DGlcDAG) from MGlc- DAG and UDP-Glc. Under certain circumstances, when MGlcDAG turns bilayer-prone by saturated chains, the more acylated and more nonbilayer-prone minor glu- colipids, i.e. MAMGlcDAG and MADGlcDAG, are synthesized [1]. Likewise, substantial amounts (20–30 mol/ 100 mol) of the normally minor precursor 1,2-DAG may accumulate in membranes with many saturated acyl chains [2]. It has been shown that the lipid composition is regulated to maintain certain properties: (a) a balance between bilayer and nonbilayer lipids (e.g. MGlcDAG/DGlcDAG) yielding phase equilibria close to a bilayer/nonbilayer transition; (b) a certain surface charge density through the ratio between the glucolipids (MGlcDAG, DGlcDAG and more acylated variants) and the charged lipids (PG and the phosphoryl- ated glucolipids). In vivo it has been shown that the ratio between DGlcDAG and PG is nearly constant [3]. The regulation of these properties is sensed and performed by the lipid synthesizing enzymes. Each enzyme acts on a lipid substrate with a specific headgroup, but are also sensitive to the type of acyl chains and lipid composition in the membrane [2]. MGlcDAG synthase (I) is activated by approximately 20 mol/100 mol PG or 10 mol/100 mol cardiolipin (CL), but is not critically dependent on the nature of the phosphate moiety and can be activated by other negatively charged lipids, however, not as efficiently [4–7]. The activation by CL indicates no specificity for the PG headgroup, but that the negatively charged phosphate is important for the enzyme. DGlcDAG synthase (II) is activated by PG or CL in the same way as MGlcDAG synthase, but also by other phosphate-containing species such as certain metabolites and dsDNA [8]. However, PG is the strongest activator among the naturally occurring lipids (strain A-EF22 does not make CL). As for the MGlcDAG synthase, this process is cooperative with respect to PG amounts and has a fairly high Hill coefficient (4–6 for MGlcDAG synthase and 3–7 for DGlcDAG synthase) [4]. Substrate fractions of MGlc- DAG up to five mol/100 mol raise the activity, which then levels out, most likely due to saturation of the active site [8]. More DGlcDAG is made from MGlcDAG in membranes with more unsaturated or longer acyl chains, increased temperature or increased amount of cholesterol. The shift in this lipid ratio stems from the more pronounced nonlamellar tendency of the membrane, compensated by making more DGlcDAG (from the nonbilayer prone MGlcDAG). This curvature sensitivity implies a sensing mechanism of mem- brane perturbation of nonbilayer-prone lipids. Analogous sensing features have been proposed for CTP:phosphocho- line cytidylyltransferase [9–11] or protein kinase C [12]. Could lipids adopting a heterogeneous lateral distribution have a bearing on the activity of the A. laidlawii glucosyl- transferases, in that substrates or surface charges become locally concentrated? Mammalian plasma membranes show transverse and lateral asymmetry. In the outer leaflet, ÔraftsÕ can form by the tight packing of saturated glycosphingo- lipids and cholesterol in a L o phase, possible to isolate [13–16]. The biological function seems to be enrichment of certain proteins, e.g. doubly acylated or GPI anchored in the ÔraftsÕ [17–21] involved in signaling and transport over the membrane. In the inner leaflet, lateral heterogeneity can form with phosphatidylserine and diacylglycerol, activating protein kinase C [22]. In this respect DAG is special in the interspacing, dehydration and altering conformation of lipid headgroups, as well as conferring a nonbilayer propensity for the membrane [23]. The reason for lateral heterogeneity is preferential interaction between headgroups (Coulombic forces, hydro- gen bonding, divalent cations, hydration level) or acyl chains (London forces) of certain lipids [24,25]. It is known that acyl hydrocarbon chain mismatch can cause lateral segregation, either by the length or the degree of saturation [5,26–30]. Indeed, stability of ÔraftsÕ demands a critical mismatch, as POPE (1-palmitoyl-2-oleoyl-sn-glycerophos- phatidylethanolamine) but not PDPE (1-palmitoyl- 2-docosahexanoyl-sn-glycerophosphatidylethanolamine) mix with raft lipids [31]. In vivo lipid mixtures from Micrococcus luteus and A. laidlawii, both containing gly- colipids and PG, reveal interactions between the individual lipids in monolayer experiments [32,33]. Analogous features are also recorded for plant galactolipids. The importance of the glycolipids for these properties are highlighted by the lower lateral diffusion for A. laidlawii in vivo glucolipids compared to the E. coli phospholipids [34]. To investigate whether lateral heterogeneity exists in the fluid glucolipid-rich membrane of A. laidlawii A-EF22 as a function of headgroup composition, liposomes were made where composition of five different lipids (major lipids in the membrane of A. laidlawii A-EF22), all with di-18:1c acyl chains, was varied according to a chemometrical experi- mental design. Pyrene-derivatives of the same lipids, inclu- ding endogenous major glucolipids synthesized by A. laidlawii, were used as fluorescent probes. A potential influence on the MGlcDAG synthase, the first regulating enzyme in the glucolipid pathway, was also investigated. Materials and methods Lipids and probes MGlcDAG and DGlcDAG were prepared from A. laidla- wii cells grown in a lipid-depleted medium supplemented with oleic acid [35]. 1,2-dioleoylglycerol (1,2-DOG) was pur- chased from Larodan (Malmo ¨ , Sweden). Phosphatidylgly- cerol (PG) was purchased from Avanti polar Lipids (USA). Pyrenedecanoic (PD) acid, 1-palmitoyl-2-pyrenedecanoyl- phosphatidylglycerol (PyrPG) and 1,2-bis-[10-(pyren-1-yl)] decanoyl-sn-glycero-3-phosphatidylcholine (bis-PyrPC) was purchased from Molecular Probes Inc. (Oregon, USA). 1-palmioyl-2-pyrenedecanoyl-glycerol (PyrDAG) and 1,2-bis-[10-(pyren-1-yl)]decanoyl-sn-glycero-3-phospho-rac- glycerol (bis-PyrPG) and 1-palmioyl-2-pyrenedecanoyl phosphatidic acid (PyrPA) were from KKV Bioware (Espos, Finland). Organism and growth conditions A. laidlawii strain A-EF22 was grown at 30 °C in a lipid- depleted tryptose/bovine serum albumin medium [36]. The 1700 P. Storm et al.(Eur. J. Biochem. 270) Ó FEBS 2003 fatty acids, oleic (18:1c) and palmitic (16:0) were supple- mented from sterile ethanol stock solutions and pyrene- decanoyl (PD) acid was supplemented from sterile dimethyl sulfoxide stock solution. Total concentration of fatty acids was 150 l M in the growth medium. Fatty acids were radiolabeled with 10 lCiÆL )1 [ 14 C]palmitic and 100 lCiÆL )1 [ 3 H]oleic acid (Amersham Pharmacia Biotech, Uppsala, Sweden), respectively, after four consecutive inoculations. 2-hydroxy-propyl-b-cyclodextrin (10 m M ) was used in the medium as a carrier for the PD acid [37]. Cell growth was monitored by absorbance and by phase contrast light microscopy. Contamination by any other bacteria was analyzed on standard bacteriological agar plates. Extraction and analysis of lipids Cells were harvested by centrifugation, washed twice in buffer, and frozen at )80 °C. Membrane lipids were extracted from the cell pellets using chloroform/methanol (2 : 1, v/v). One-dimensional thin layer chromatography (TLC) was used to separate and characterize the different lipids in the membrane. The TLC plates coated with silica gel 60 (Merck, Darmstadt, Germany) were developed in chloro- form/methanol/water (80 : 25 : 4, v/v/v). [ 14 C]-labeled lipids were visualized with electronic autoradiography (Packard Instant Imager). Excised gel lipid spots were digested in Soluene-350 (Packed) for 30 min at 37 °Cand quantified by double-channel liquid scintillation counting. To purify the pyrenyl lipids, the TLC plate was developed first in chloroform/methanol/water (80 : 25 : 4, v/v/v) and then in chloroform/methanol/ammonia (91 : 35 : 10, v/v/ v). Compared with a one-dimensionally developed TLC plate of extracted lipids from medium 18:1c/PD 120 l M :30l M , the spots of pyrenyl lipids could be separated better and become more concentrated in two dimensions. Excised gel spots of pure pyrenyl glucolipids (MGlcDAG and DGlcDAG) were extracted by chloro- form/methanol (2 : 1, v/v), and typical fluorescence spec- tra of mono-pyrenyl and bis-pyrenyl glucolipids visualized (Fig. 2B,C). Incorporation of PD and synthesis of pyrenyl glucolipids in vivo No growth of A. laidlawii could be observed with only 16:0 or PD, separately or together (Table 1). The presence of 18:1c fatty acid was very important for both the growth of A. laidlawii and the incorporation of PD into pyrenyl lipids. However, the yield of pyrenyl lipids was quite low (less than 10% on the basis of added fatty acids) compared to the nonpyrenyl lipids (30%)40%), but incorporation into MGlcDAG and DGlcDAG was fairly similar. The yield of nonpyrenyl MGlcDAG from A. laidlawii strain A-EF22 was much lower than that of nonpyrenyl DGlcDAG when PD in growth media; revealed by quantitation of nonpyrenyl lipids from excised gel spots (data not shown). In the same membrane 18:1c fatty acid preferred to incorporate PD acid to produce mono-pyrenyl glucolipid rather than nonpyrenyl glucolipid, whereas 16:0 dominated in the latter (data not shown). The Ie/Im ratio from the fluorescence spectra increased for the extracted lipid mixture from cells when increasing the PD ratio in the medium, showing that more bis-pyrenyl lipids were synthesized at higher PD acid to fatty acid ratios (data not shown). The yield of synthesized pyrenyl glucolipids was determined from standard fluorescence intensity curves, obtained from synthetic PyrDAG and bis-PyrPG. Enzymes and assays Mixed lipid micelles were made by swelling dry lipid to a final concentration of 10 m M (1 m M substrate) in a buffer of 110 m M Tris pH 8, 22 m M Chaps, 22 m M Mg 2+ .Purified MGlcDAG synthase (50 lL) or DGlcDAG synthase was incubated with 40 lL lipid micelles at 4 °Cfor30min.The enzyme reaction was started by adding 10 lLof10m M (0.5 CiÆmol )1 )UDP-[ 14 C]glucose. The reactions were ter- minated by the addition of 375 lL methanol/chloroform (2 : 1, v/v). Synthesized MGlcDAG or DGlcDAG was extracted according to a modified Bligh and Dyer method [38] and separated from other lipids by TLC. The 14 C- labeled glucolipid products were quantified using electronic autoradiography (Packard Instant Imager). Homogeneous MGlcDAG synthase for liposome binding was purified from detergent-solubilized A. laidlawii cells by three column chromatography methods, including ion exchange, gel filtration and hydroxyapatite chromatography [39]. Experimental design Chemometrics is how to design an experimental series in order to extract the maximum information from the minimum number of experiments [40]. Chosen factors are varied simultaneously in a randomized run order to reduce or eliminate unknown or uncontrolled influence on data. Table 1. Incorporation of PD and yield of A. laidlawii pyrenyl lipids. Fatty acids (l M ) PD/FA % Incorporation of PD into lipids16:0 18:1c PD Initial Harvest 1 120 30 0 – – – 2 90 30 30 0.25 0.024 3.83 3 30 30 90 1.5 0.21 8.13 4 0 60 90 1.5 0.31 6.17 5 0 30 120 4 0.68 7.03 6 0 10 140 14 1.62 4.22 Ó FEBS 2003 Lateral organization of A. laidlawii lipids (Eur. J. Biochem. 270) 1701 The response(s) Y is then fitted to the variables by a mathematical model, e.g. Y ¼ m + Xb + e;wereX is the model terms/variables, b is the coefficient of effect and e is the residuals. We have used the MODDE 3.0 package (Umetri AB, Umea ˚ , Sweden). Here, variables were changed from low to high and the response was plotted and analyzed in the computer to give a measure of effects. Variables in this case are the amounts of different lipid headgroups (as all acyl chains are 18:1c) and the amount of MGlcDAG synthase. Responses are excimer formation of pyrene-labeled phos- pholipid or anisotropy of diphenylhexatriene (DPH). DGlcDAG, considered the matrix lipid, was set as a filler and a full factorial design was chosen. In the simple case of three variables (dimensions, factors), a full factorial design is a cube in the experimental space, where data points are in the corners and center of the cube (Fig. 1), resulting in a linear interaction model. In a couple of cases the investigation was expanded to a response surface model, i.e. composite face-centered (CCF) design, where the design- cube also has data points on the face of the sides, making quadratic models possible to obtain. For the investigation of chain ordering for pure lipids a mixture ( D -optimal) design was chosen, where matrix lipid DGlcDAG is not set as a filler. Partial least squares (PLS) was used to fit the model. PLS finds the relationship between a matrix Y (response variables) and a matrix X (model terms). Measure of model fit is R 2 ¼ 1 ) (RSS/YSS), where RSS is residual sum of squares and YSS is the response sum of squares. Internal validation (crossvalidation or prediction ability) is measured by the Q 2 value, i.e. Q 2 ¼ 1 ) (PRESS/YSS), where PRESS is the predicted residual sum of squares. Rules of thumb are that R 2 should be at least 0.8 and Q 2 above 0.3 for linear models and even closer to one for quadratic models. R 2 and Q 2 are the overall parameters for model accuracy, which encompass analysis of variance ( ANOVA ), lack of fit, normal distribution of residuals. If these parameters are not satisfactory, then outliers, wrong metric, inhomogeneous data, range of the factors, etc., need to be investigated. The fitted model can then be presented as a response surface (Fig. 1B), a curve or a table. Furthermore, an important aspect of experimental design is that interaction effects can be detected; this would not be possible if only one variable at a time was changed. Inter- action means that the response of a variable is dependent on the level of another variable in a nonadditive fashion. Preparation of large unilamellar vesicles For each sample 0.25 lmol of total lipid was mixed to the desired composition according to the experimental design, where DOPG was varied 0–40 mol/100 mol, DOPA 0–10 mol/100 mol, MGlcDAG 0–30 mol/100 mol, DOG 0–10 mol/100 mol, and DGlcDAG was used as the (bal- ance) matrix. The content of fluorescence probe was constant at 1 mol/100 mol for mono-pyrenyl lipids, and 0.5 mol/100 mol for bis-pyrenyl lipids or DPH. The mixture was then dried under a nitrogen flow and then under reduced pressure (vacuum) overnight. The resulting lipid film was hydrated with intermittent vortexing during 45 min in filtered and deoxygenated 10 m M Hepes pH 8.0 with 5m M MgCl 2 , and then extruded with a LiposoFast Basic extruder (Avestin Inc., Canada) 19 times through two stacked polycarbonate filters (Millipore; pore diameter 100 nm). This yields large unilamellar vesicles (LUV) with an average diameter of nearly 100 nm [41]. The quality of vesicles for all data points was verified with dithionite quenching of an NBD-probe, showing that all vesicles were LUVs, as only the outer leaflet is quenched and roughly 50% of the signal remained after quenching (data not shown). Fluorescence and absorbance methods Absorbance measurements were performed with a Beckman DU 70 spectrophotometer. Fluorescence measurements with labeled vesicles were carried out so that 50 lL of prepared liposomes were added to 1950 lL buffer in an optical 1 · 1 cm fluorescence cuvette, and fluorescence measured with a Spex Fluoro- Max-2 fluorometer with magnetic stirrer and temperature control (28 °C). Samples with pyrene probes, were excited at 344 nm and emission spectra collected between 360 and 500 nm. Slits had a bandwidth of 1 nm for excitation and 4 nm for emission (step width 1 nm, integration time 0.5 s). Four scans were sampled, averaged, and subtracted by a blank consisting of the buffer, in order to obtain the fluorescence curve. Vesicles without a probe do not particularly affect the spectra, as verified in a control design showing only noise that was virtually the same as the blank; therefore no such reference was used in any of the runs. Ie/Im (excimer ratio) was calculated as the ratio between excimer emission at 480 nm (Ie), when two pyrenes are in close proximity ( 3.5 A ˚ ), and monomer emission at 398 nm (Im). Enzyme (MGlcDAG synthase) was incubated with 50 lL liposomes (protein : lipid 1 : 700–1 : 70) on ice for 30 min prior to measurement at room temperature in a 1 · 0.2 cm quartz fluorescence cuvette using a Perkin- Elmer LB50 spectrofluorimeter. DPH anisotropy [42,43] was analyzed using a Spex Fluorolog 12 fluorometer (Department of Biophysical Chemistry, Umea ˚ University), where bandwidths were 3.6 nm for excitation and 7.2 nm for emission. Sample Fig. 1. Experimental design. (A) Full factorial design cube with cen- terpoint. Variables, e.g. lipids, are changed from low to high amounts. (B) Example of a response surface plot showing response variation when varying two variables. The purpose of the design is to extract maximum information from a minimum number of experiments. 1702 P. Storm et al.(Eur. J. Biochem. 270) Ó FEBS 2003 solution was equilibrated for five minutes in the cuvette holder (no magnetic stirrer) to reach a temperature of 28 °C prior to measurement. Absorption at the excitation wave- length was less than 0.09, thus a minimal reabsorption. Anisotropy r ¼ (I I ) I ^ )/(I I +2I ^ ) for each datapoint was calculated and averaged in connection to the measurement by a computer program. Results Enzymes recognize pyrene derivatives For analysis of potential interactions between the various A. laidlawii membrane lipids we chose fluorescence spectro- scopy, with pyrene-labeled lipids containing one normal and one pyrene-labeled chain as proximity probes, as the studied phenomena may be transient and not possible to isolate, and too small (<300 nm) to be detected with microscopy. Pyrene-decanoyl chains locate in the membrane hydro- phobic core and are virtually nonperturbing at a fraction of 1 mol/100 mol or less, partitioning preferentially in the fluid phase [44–49]. To investigate chain ordering, i.e. membrane free volume (V f ), a bis-pyrenyl lipid, with a pyrene on both acyl chains, was used [47,50]. Native A. laidlawii pyrene-labeled glucolipids (not com- mercially available) were produced in vivo,andtestedas lipid enzyme substrates in vitro to monitor the impact of the pyrenyl chain moiety on headgroup organization. Pyrenyl- decanoic acid (PD) was used for the incorporation of the pyrene group into the lipids. 16:0 and 18:1c fatty acid were chosen for their approximately similar chain length to PD. Different compositions of growth medium fatty acids were used to optimize the incorporation of PD acid, with or without 16:0 and 18:1c, into the glucolipids (Table 1). Cell growth and size of cells were checked by routine light microscopy. PD or 16:0 could not support growth, alone or in combination. A. laidlawii cells became much bigger and less aggregated, and the density of the culture became lower,whentheratioofPDinfattyacidswasincreased. One-dimensional thin layer chromatography developed in chloroform/methanol/water (80 : 25 : 4, v/v/v) was used to characterize the lipids extracted from the cells (Fig. 2). The R f values of different lipids on a TLC plate were compared according to the standard samples characterized by NMR [1,51]. Fluorescent spots (under UV light) are marked by rings in Fig. 2A. Combined with the data from radiolabel analysis, it is obvious that without a pyrene group in the hydrocarbon chain of the lipid, there was no fluorescence. With one PD acyl chain and the other chain 16:0 or 18:1c, as in mono-pyrenyl lipids, both fluorescence and isotope signals were detected (data not shown). With two pyrenyl chains, only fluorescence but no isotope signal could be detected from the spot of the bis-pyrene lipid on the TLC plate (Fig. 2A). Note that mono-pyrenyl lipid migrated a little further than the nonpyrenyl lipid, and bis-pyrenyl lipid migrated even further, as expected from the larger hydrocarbon regions of the pyrenyl-containing lipids (Fig. 2A). Purified MGlcDAG synthase and partially purified DGlcDAG synthase were used to study the potential disturbance of the polar headgroup organization by the purified pyrenyl-labeled glucolipids in vitro (Fig. 3). The enzymatic products, pyrenyl-MGlcDAG or pyrenyl-DGlc- DAG, from the in vitro enzyme reactions, were extracted from TLC plates. Similar yields were obtained for pyrenyl- glucolipid and nonpyrenyl-glucolipid products, from both the MGlcDAG synthase and DGlcDAG synthase reactions (Fig. 3). Furthermore, the shape of the fluorescence spectra of the product depends on which type of pyrenyl lipid was used as substrate; mono- and bis-pyrenyl lipid substrate produced mono- or bis-pyrenyl glucolipid products, respectively (data not shown). Thus, these enzymes do not discriminate between substrates with a pyrene moiety in the acyl chain. Similar features have been observed for enzymes Fig. 2. In vivo synthesis of pyrenyl lipids. (A) A. laidlawii 14 C/ 3 H- labeled glucolipids and pyrenyl-glucolipids after TLC separation. Extracted lipids applied on TLC plates were developed in chloroform/ methanol/ammonia (91 : 35 : 10, v/v/v). The growth medium fatty acid composition (ratio 16:0/18:1c/PD) was from left to right: 120 : 30 : 0; 90 : 30 : 30; 30 : 30 : 90; 0 : 60 : 90; 0 : 30 : 120 and 0 : 10 : 140. Encircled spots represent the fluorescent mono- pyrenylMGlcDAG (lower) and bis-pyrenylMGlcDAG (upper), respectively. The fluorescence spectra of purified mono-pyrenyl (B) and bis-pyrenyl glucolipids (C) produced in vivo.Thesamples(0.1m M lipid) were excited at 344 nm in chloroform/methanol (2 : 1, v/v). Synthetic mono-pyrenylDAG (B), and bis-pyrenylPG (C), were used as references. Fig. 3. Pyrenyl lipids as enzyme substrates in vitro. Synthesis of DGlcDAG from MGlcDAG and UDP-[ 14 C]glucose by purified DGlcDAG synthase. The contents of pyrenyl glucolipids were less than 1% (mol/mol). (s)[ 14 C]DGlcDAG produced from di-18:1c- MGlcDAG; (m) mono-pyrenyl DGlcDAG produced from mono- pyrenyl MGlcDAG; and (d) bis-pyrenyl DGlcDAG produced from bis-pyrenyl MGlcDAG, respectively, by the DGlcDAG synthase. Ó FEBS 2003 Lateral organization of A. laidlawii lipids (Eur. J. Biochem. 270) 1703 acting on phosphatidylinositol lipids [52]. This seems logical in that the MGlcDAG synthase is attached to the membrane interface [7] and does not recognize the acyl chain region close to the bilayer center, where the pyrene moiety is. Lipid organization as seen with pyrene derivatives Potential interactions between the A. laidlawii membrane lipids were analyzed in liposome bilayer models containing various pyrene-labeled probes of synthetic and in vivo origin. Excimer formation (Materials and Methods) for lipids with one pyrene-acyl chain is an intermolecular event, depending on the collision rate, and Ie/Im hence monitor lateral mobility and concentration of these molecules. A series of full factorial designs were made, where the compositions (87 conditions in total) were varied to cover the limits occurring in vivo for the five important lipids of the glucolipid pathway in the membrane of A. laidlawii (PG, PA, DAG, MGlcDAG and DGlcDAG). In in vitro bilayer (liposome) models, the lipids adopted a heterogeneous organization, as was seen with excimer formation of the pyrene-labeled probes for each lipid type. Table 2 lists lipid composition and the statistically significant changes in Ie/Im (the excimer ratio) for different pyrene derivatives as a function of different headgroups, going from a low mole content to a high mole content for each lipid, and where all acyl chains are 18:1c (dioleoyl). In two cases this investiga- tion was expanded to a composite face-centered (CCF) design. The distribution of DAG, as monitored by Ie/Imofthe PyrDAG probe (Table 2) was affected significantly by PG and DAG (increased Ie/Im), but none of the other lipids, according to a CCF design. Interestingly, the model also revealed a dependence (interaction) between the variables PG and DAG (data not shown), i.e. increasing both PG and DAG does not increase the Ie/Im additively. In accordance with this model a one-variable titration, where DOPG was varied 0–40 mol/100 mol (at constant 5 mol/100 mol DOPA, 5 mol/100 mol DOG and DGlcDAG as balance), showed a 1.5-fold increase in Ie/Im (0.080–0.124) for the PyrDAG probe between 0 and 25 mol/100 mol PG, and then a decrease to 0.105 between 25 and 40 mol/100 mol PG (Fig. 4). We therefore conclude that DAG has a heterogeneous distribution, strongly promoted by increas- ing amounts of PG. It is interesting to note that increasing the amount of CL (0–20 mol/100 mol) in a DGlcDAG matrix with 5 mol/100 mol DOG and 1 mol/100 mol PyrDAG, did not produce any change in the Ie/Im(data not shown). For the PyrPG probe Ie/Im increased 1.35-fold (0.051– 0.069) between 0 and 40% PG (Table 2). It has been noted before that decreasing amounts of DGlcDAG (the balance here) upon increasing PG, increased the collision rate between pyrenes [5]. All other lipids had an insignificant effect on excimer formation. As a comparison, increasing PG amounts in a matrix with DGlcDAG did not change the Fig. 4. Glucosyltransferases and lipid organization. (A) Response of pyrene-derivatives of activator PG and substrate DOG lipids, and order-sensing bis-PyrPC, upon increase in the DOPG content. (B) Normalized enzyme activities (adapted from Dahlqvist et al.[3])ofthe MGlcDAG and DGlcDAG synthases. Table 2. Lateral interactions between A. laidlawii lipids in liposome bilayers. Changes in Ie/Im of pyrene probes upon variation in lipid amounts according to a factorial design or, in the case of PyrDAG and PyrPA probes, to a composite face-centered design (Materials and methods). The balance (matrix) in the various lipid mixtures was always DGlcDAG. Only statistically significant changes are shown. NT, not tested. R 2 and Q 2 are measures of model fit and vary between 0 and 1 (Materials and methods). Probe MGlcDAG 0–30 mol/100 mol DAG 0–10 mol/100 mol PG 0–40 mol/100 mol PA 0–10 mol/100 mol Replicate error (Ie/Im) R 2 /Q 2 PyrDAG – 0.078–0.087 (11.5%) 0.078–0.110 (41%) – ± 0.0017 0.97/0.72 PyrPG – – 0.051–0.069 (35%) – ± 0.0031 0.81/0.59 PyrPA 0.058–0.062 (7%) 0.058–0.064 (10%) – – ± 0.0027 0.90/0.51 PyrMGDG 0.081–0.107 (32%) – – – ± 0.0027 0.78/0.64 PyrPC 0.064–0.068 (6%) 0.064–0.065 (1.5%) 0.064–0.073 (14%) NT ± 0.0029 0.94/0.42 1704 P. Storm et al.(Eur. J. Biochem. 270) Ó FEBS 2003 excimer ratio for pyrene-labeled DGlcDAG, Ie/Imwas always approximately 0.06. As the PyrPG signal increased during these conditions (Table 2, Fig. 4), this may indicate affinity between like molecules for these two lipids. The Ie/Im for pyrene probes of MGlcDAG and PA were not changed when increasing the amount of the enzyme activator lipid PG (Table 2), indicating no interactions. However, increasing the amounts of normal MGlcDAG gave an increase in Ie/Im for PyrMGlcDAG probe (Table 2), supporting an interaction between the MGlc- DAG molecules (note that decreasing DGlcDAG amounts was the balance). Also, headgroup interaction for MGlc- DAG lipids is indicated in a monolayer study, showing a much more compact fluid state than the equivalent phosphatidylcholine (PC) lipid [53]. For PA, precursor to both the glucolipid and phospholipid pathways, an increase in Ie/Im for PyrPA probe was also observed when increasing the amount of DAG or MGlcDAG (Table 2) in a CCF design, supporting a heterogeneous distribution, however, the effect was small. Note that there are only a few mol percent of this lipid naturally occurring in the membrane. A small effect was also seen with PyrPC as a probe (Table 2), probably reflecting exclusion from the domains formed. Bilayer chain ordering A starting point is the response at a homogenous distribu- tion of probe in the membrane, as is the case for 1 mol/100 mol PyrPC in a matrix of DOPC [46]. This gave an Ie/Im of 0.07 in PC-matrix and 0.06 in DGlcDAG- matrix (data not shown). The difference may reflect ease of diffusion. Diffusion is also indirectly related to chain ordering in the membrane. This property decreased with increasing content of DGlcDAG and increased with increasing content of DOPG or DOG when measured with a bis-PyrPC probe. For bis-pyrenyl lipid probes the Ie/Im reflects an intramolecular event, where an increase corresponds to increased chain-chain contacts (collision rates). The Ie/Im was 1.4 at 40 mol/100 mol DOPG, and 1.05 at 0 mol/100 mol DOPG (Table 3), indicating a decreased V f by increased PG (or increased V f by DGlcDAG). A complete inverse of this property was indicatedwhenmeasuredwithDPH(Table3).Steadystate anisotropy r for DPH was between 0.10 at high DOPG (low DGlcDAG) content and 0.14 at low DOPG content (high DGlcDAG). DOG had the same effect as DOPG, although the smaller fraction in the membrane made its effect less pronounced. This has been observed before [26] and was addressed to chain splaying motions of the bis-PyrPC. The location of pyrene has nevertheless been determined to be located close to the bilayer center. MGlcDAG, shown to increase order [54], is organized laterally in this five-lipid model system in a way that gives no significant contribution to the ordering of the membrane bilayer, according to the experimental design model. Enzyme binding, lateral organization and chain order The MGlcDAG synthase (purified without detergent [55]), when binding to the lipid bilayers at initial protein : lipid- ratios of 1 : 700, 1 : 350 or 1 : 70 (mol/mol), did not affect the Ie/Im of PyrDAG or PyrPG (data not shown). Thus, the substrate lipid DAG and the activator lipid PG were not concentrated on a large scale by the enzyme. Yet, binding and activity with liposomes, as here, was strongly correlated with increasing amounts of anionic lipid ([55] and Li et al. submitted), with a Hill coefficient of 4–6, which is the potential number of PG molecules associated to the enzyme. However, the number of lipid molecules bound under an enzyme makes it very unlikely that these be two probe molecules (1 mol/100 mol concentration), even at the highest enzyme-lipid ratio. Binding was practically irrever- sible and more enzymes bound to liposomes as a function of mol/100 mol anionic lipid (promoted by nonbilayer lipid), revealed by surface plasmon resonance experiments using Biacore (Li et al. submitted). Furthermore, the Ie/Imfroma bis-PyrPC probe was reduced by the addition of enzyme, as seen in a design (R 2 ¼ 0.93, Q 2 ¼ 0.41) with LUVs composed of DOPG (0–40%), CL (0–20%), DAG (0– 10%), bis-PyrPC (0.1%) and DGlcDOG as balance, and 1 : 700–1 : 70 (mol/mol) enzyme : lipid (Fig. 5). PG and DAG increase the order, as seen in Table 3, as do CL. There was also synergism between PG and the enzyme, and antagonism between DAG and enzyme, with respect to chain-ordering effects (Fig. 5). Hence, interfacial binding of fairly large amounts of the MGlcDAG enzyme reduced chain order (increased V f ) but did not detectably change the lateral distribution of the A. laidlawii A-EF22 polar lipid species. Table 3. Lipid composition and chain ordering. Excimer formation (Ie/Im) and anisotropy (r) were monitored by bis-PyrPC and DPH, respectively. A mixture design was made (Materials and methods) where DOPG 0–40 mol/100 mol, DODAG 0–10 mol/100 mol, DODGlcDAG 40–90 mol/ 100 mol (40–99 mol/100 mol with DPH as a probe), DOMGlcDAG 0–30 mol/100 mol, DOPA 0–10 mol/100 mol. Only lipids with a significant effect are shown, with variables from the modeled data. The model is linear with R 2 ¼ 0.95 and Q 2 ¼ 0.91 for bis-PyrPC as the probe, and R 2 ¼ 0.89 and Q 2 ¼ 0.79 for DPH as the probe. Lipid Content level Ie/Im r DOPG low (0 mol%) 1.09 0.122 high (40 mol%) 1.37 0.101 DODAG low (0 mol%) 1.2 0.115 high (10 mol%) 1.262 0.108 DODGlcDAG low (40 mol%) 1.38 0.103 high (90 or 99 mol%) 1.02 0.14 Replicate error ± 0.024 ± 0.0014 Ó FEBS 2003 Lateral organization of A. laidlawii lipids (Eur. J. Biochem. 270) 1705 Discussion Synthesis of pyrenyl glucolipid The existence of 18:1c fatty acid in the medium is very important for both cell growth and the incorporation of PD into glucolipids (Materials and Methods). A higher ratio of PD in the growth medium produces more bis-pyrenyl glucolipids, as A. laidlawii must choose PD as the side chain, as there are not enough of the other fatty acids in the medium. However, the cell cannot grow well with a high PD ratio. In vivo it might be difficult to incorporate PD acid into the precursor PA as the occupied volume and hydropho- bicity are larger than for 18:1 or 16:0 fatty acids, resulting in low yields of pyrenyl lipids. In our experiment, less than 10% of PD could be incorporated into lipids. Media with a 30 : 120 ratio of 18:1c to PD is appropriate to obtain a reasonable yield of mono-pyrenyl and bis-pyrenyl gluco- lipids. In vitro, using PyrDAG, purified mono-pyrenyl MGlcDAG or bis-pyrenyl together with extra large amounts of nonpyrenyl lipids as substrates, the product yields were approximately the same for these two glucolipid synthases (Fig. 2A). This suggests that the enzymes do not discriminate between pyrenyl lipid and nonpyrenyl lipid in a micelle system, which makes pyrenyl-glucolipid probes possible to use in the study of metabolic stages of glucolipids and biophysical properties of membranes. This is also the first time mono- and bis-pyrenyl glucolipids from A. laidlawii have been synthesized and purified. Lateral organization of A. laidlawii lipids The metabolism of glucolipids in A. laidlawii depends on several factors such as growth temperature, presence of foreign molecules, and unsaturation and length of the fatty acids [2,3,56]. Not all lipids adopt a homogeneous distribution in liquid-crystalline liposome model membranes with A. laidlawii lipids having 18:1c acyl chains, as indicated from the present investigation. The most prominent effect was for DAG when increasing the molar fraction of PG. There was a good agreement between results from the factorial design models (Table 2) and a one-variable titra- tion showing (at most) a 1.5-fold increase in the excimer ratio (Ie/Im) (Fig. 4). Similar increase in Ie/Im was observed when ceramide (DAG analogue) was enzymatically split from sphingomyelin and forming microdomains [57,58], or the patching of DOPG in liquid-crystalline PC due to a hydrophobic (chain length) mismatch [5]. This is most probably not due to an increased diffusion in the mem- brane. Given an excited state lifetime of 100 ns for pyrene and diffusion coefficient of lipids approximately 5 · 10 )8 cm 2 Æs )1 , the pyrene derivative move at most two lipid diameters and should not form excited dimers unless already being close [59]. This is supported by PyrPC showing a small change in Ie/Im (Table 2) and by the ordering of lipids seen with bis-PyrPC (Table 3). Indeed, total lipid extracts from A. laidlawii grownattwodifferent growth temperatures, thereby containing different lipid compositions, show a fairly similar lateral diffusion coeffi- cient [34]. An interesting observation is that anisotropy of DPH is decreasing when Ie/Im of bis-PyrPC is increasing (Table 3). However, Ie/Im increases when cholesterol is included in the alloys with sugar lipids (data not shown), giving reason to believe that bis-PyrPC gives a good indication of the fluidity (free volume) in the membrane. No further investigation was made into this phenomenon, but hypothesizing that it may have to do with the matrix of sugar lipids with 18:1c acyl chains, as a report by Kaiser and London [60] states that DPH is located close to bilayer center in DOPC bilayers. Thus, DAG is segregated in what seems like micro- domains by increasing the amount of PG in the membrane, possibly due to favorable hydrogen bonding between DAG and PG. This is not due to charge entirely, as PA had no appreciable effect on the PyrDAG-probe response. The decrease in Ie/Im after 25 mol/100 mol PG (Fig. 4) can be that a microdomain formed preferably by the DAG lipid is diluted. DAG is a special lipid, as pointed out in a review by Goni and Alonso [23]. Constituting only a small fraction of membranes, it can act as an intracellular second messenger or metabolic intermediate and is involved in enzyme modulation, membrane fusion and membrane physical properties. For membrane physical properties, unsaturated DAG imposes no phase separation at low molar fractions but does increase the chain order in an unsaturated phosphatidylcholine membrane. This ordering was also Fig. 5. Enzyme binding and chain ordering. Plot showing the effects of varying amount of MGlcDAG synthase (MGS) and three different lipids in LUVs with DGlcDAG as balance, going from low to high amount, and 0.1 mol/100 mol bis-PyrPC as probe (computed (R 2 ¼ 0.93, Q 2 ¼ 0.41) in MODDE 3.0). Range: CL 0–20 mol/100 mol, DOG 0–15 mol/100 mol, DOPG 0–40 mol/100 mol and MGS/Lipid 1 : 700–1 : 70 (mol/mol). Interaction effects between pairs of variables are also plotted, where positive value means synergism and negative value antagonism. The effect in Ie/Im is the sum of the response difference between high and low variable levels divided by two. Effects for which error bars encompass zero are insignificant. 1706 P. Storm et al.(Eur. J. Biochem. 270) Ó FEBS 2003 observed in the system investigated here (Table 3 and Fig. 5). DAG also increases the spacing between phospho- lipid (or glucolipid) headgroups, with its hydroxyl proton participating in hydrogen bonding. This is part of the reason why ceramide assemble laterally when enzymatically released from sphingomyelin [58,61]. Hydrogen bonding in the lipid interface plays a role in the interaction between sugar headgroups, where a subtle difference as between galactose and glucose may be important [62]. For other lipid species, modulation by divalent cations (in our case Mg 2+ ) coordinating negatively charged lipids and decreasing the headgroup repulsion, is also important. PG, which has a flexible glycerol moiety in the headgroup, is shown to take part in hydrogen bonding [63,64] and it has been noticed that PG and dimannosyl-DAG is heterogeneously distri- buted, due to interactions in the headgroup region [65]. Furthermore, lipids in total lipid mixtures from A. laidlawii have a smaller interfacial area than any of the individual lipids [33,65], due to interactions in the headgroup region causing a lateral condensation. An interaction between MGlcDAG or DGlcDAG species was revealed here from the responses of the corresponding pyrene probes (Table 2 and Results above). The patching of substrate lipid DAG with activator PG (Fig. 4), in combination with an increased charge density creating an electrostatic interaction between the membrane and the enzyme (L. Li, unpublished observation), contribute to the explanation of the anionic lipid preference for the MGlcDAG synthase. An increase in acyl chain order also precedes enzyme activity (bis-PyrPC signal in Fig. 4). For the DGlcDAG synthase the patching of its substrate MGlcDAG when the amount increases (Table 2) seems not biologically relevant as PG is the only naturally occurring activator, and that 3 mol/100 mol MGlcDAG saturates the DGlcDAG synthase [8]. However, the order increase probably contributes to the activity (Fig. 4). With respect to the precursor PA, the PyrPA show an addi- tively increased Ie/Im by DAG and MGlcDAG (Table 2), but the biological function is less clear as the PA phospha- tase is not regulated by changes in lipid composition [66]. In conclusion, this is the first time that up to five lipids have been varied simultaneously in vitro and that a fluorescent glucolipid probe has been synthesized in vivo. We find that DAG does not mix ideally but forms microdomains, possibly in weak interaction with PG. As PG is the strongest activator in vivo for the two glucolipid- synthesizing enzymes, this phenomenon has a biological relevance in concentrating DAG, the substrate for MGlc- DAG synthase, which is rate-limiting in glucolipid synthesis. Purified MGlcDAG synthase, free of lipids and detergent [55], did not affect the organization (Ie/Im) for the tested PyrDAG and PyrPG probes, but affects order in the membrane. Activity therefore seems to depend more on the ability to bind to the membrane in a proper orientation/ conformation as Fig. 5 suggests, and due to the fact that CL, a strong activator, did not affect the Ie/Imfor PyrDAG. Acknowledgements This work was supported by the Swedish Natural Science Research Council, and the K & A Wallenberg foundation. References 1. Andersson, A.S., Rilfors, L., Lewis, R.N., McElhaney, R.N. & Lindblom, G. (1998) Occurrence of monoacyl-diglucosyl-diacyl- glycerol and monoacyl-bis-glycerophosphoryl-diglucosyl-diacyl- glycerol in membranes of Acholeplasma laidlawii strain B-PG9. Biochim. Biophys. Acta 1389, 43–49. 2. Wieslander, A ˚ ., Nordstro ¨ m,S.,Dahlqvist,A.,Rilfors,L.& Lindblom, G. (1995) Membrane lipid composition and cell size of Acholeplasma laidlawii strain A are strongly influenced by lipid acyl chain length. Eur. J. Biochem. 227, 734–744. 3. Dahlqvist, A., Nordstro ¨ m, S., Karlsson, O.P., Mannock, D.A., McElhaney, R.N. & Wieslander, A ˚ . (1995) Efficient modulation of glucolipid enzyme activities in membranes of Acholeplasma laidlawii by the type of lipids in the bilayer matrix. Biochemistry 34, 13381–13389. 4. Karlsson, O.P., Dahlqvist, A. & Wieslander, A ˚ . (1994) Activation of the membrane glucolipid synthesis in Acholeplasma laidlawii by phosphatidylglycerol and other anionic lipids. J. Biol. Chem. 269, 23484–23490. 5. Karlsson, O.P., Ryto ¨ maa, M., Dahlqvist, A., Kinnunen, P.K. & Wieslander, A ˚ . (1996) Correlation between bilayer lipid dynamics and activity of the diglucosyldiacylglycerol synthase from Acho- leplasma laidlawii membranes. Biochemistry 35, 10094–10102. 6. Karlsson, O.P. (1997) Maintenance of Lipid Bilayer Properties. University of Umea ˚ , Sweden. 7.Berg,S.,Edman,M.,Li,L.,Wikstro ¨ m, M. & Wieslander, A ˚ . (2001) Sequence properties of the 1,2-diacylglycerol 3-glucosyl- transferase from Acholeplasma laidlawii membranes. Recognition of a large group of lipid glycosyltransferases in eubacteria and archaea. J. Biol. Chem. 276, 22056–22063. 8. Vikstro ¨ m, S., Li, L., Karlsson, O.P. & Wieslander, A ˚ . (1999) Key role of the diglucosyldiacylglycerol synthase for the nonbilayer- bilayer lipid balance of Acholeplasma laidlawii membranes. Bio- chemistry 38, 5511–5520. 9. Arnold, R.S., DePaoli-Roach, A.A. & Cornell, R.B. (1997) Binding of CTP:phosphocholine cytidylyltransferase to lipid vesicles: Diacylglycerol and enzyme dephosphorylation increase the affinity for negatively charged membranes. Biochemistry 36, 6149–6156. 10. Cornell, R.B. & Northwood, I.C. (2000) Regulation of CTP:phosphocholine cytidylyltransferase by amphitropism and relocalization. Trends Biochem. Sci. 25, 441–447. 11. Davies, S.M., Epand,R.M., Kraayenhof, R. & Cornell, R.B. (2001) Regulation of CTP:phosphocholine cytidylyltransferase activity by the physical properties of lipid membranes: an important role for stored curvature strain energy. Biochemistry 40, 10522–10531. 12. Newton, A.C. (2001) Protein kinase C: structural and spatial regulation by phosphorylation, cofactors, and macromolecular interactions. Chem. Rev. 101, 2353–2364. 13. Simons, K. & Ikonen, E. (1997) Functional rafts in cell mem- branes. Nature 387, 569–572. 14. Harder, T. & Simons, K. (1997) Caveolae, DIGs, and the dynamics of sphingolipid-cholesterol microdomains. Curr. Opin. Cell Biol. 9, 534–542. 15. London, E. (2002) Insights into lipid raft structure and formation from experiments in model membranes. Curr. Opin. Struct. Biol. 12,480. 16. Anderson, R.G. & Jacobson, K. (2002) A role for lipid shells in targeting proteins to caveolae, rafts, and other lipid domains. Science 296, 1821–1825. 17. Varma, R. & Mayor, S. (1998) GPI-anchored proteins are orga- nized in submicron domains at the cell surface. Nature 394, 798–801. 18. Friedrichson, T. & Kurzchalia, T.V. (1998) Microdomains of GPI-anchored proteins in living cells revealed by crosslinking. Nature. 394, 802–805. Ó FEBS 2003 Lateral organization of A. laidlawii lipids (Eur. J. Biochem. 270) 1707 19. Brown, D.A. & London, E. (1998) Functions of lipid rafts in biological membranes. Annu.Rev.CellDev.Biol.14, 111–136. 20. Brown, D.A. & London, E. (2000) Structure and function of sphingolipid- and cholesterol-rich membrane rafts. J. Biol. Chem. 275, 17221–17224. 21. Lang, T., Bruns, D., Wenzel, D., Riedel, D., Holroyd, P., Thiele, C. & Jahn, R. (2001) SNAREs are concentrated in cholesterol- dependent clusters that define docking and fusion sites for exo- cytosis. EMBO J. 20, 2202–2213. 22. Yang, L. & Glaser, M. (1996) Formation of membrane domains during the activation of protein kinase C. Biochemistry 35, 13966– 13974. 23. Goni, F.M. & Alonso, A. (1999) Structure and functional properties of diacylglycerols in membranes. Progr. Lipid Res. 38, 1–48. 24. Kinnunen, P.K. (1991) On the principles of functional ordering in biological membranes. Chem. Phys, Lipids 57, 375–399. 25. Edidin, M. (1997) Lipid microdomains in cell surface membranes. Curr. Opin. Struct. Biol. 7, 528–532. 26. Lehtonen, J.Y., Holopainen, J.M. & Kinnunen, P.K. (1996) Evi- dence for the formation of microdomains in liquid crystalline large unilamellar vesicles caused by hydrophobic mismatch of the constituent phospholipids. Biophys. J. 70, 1753–1760. 27. Killian, J.A. (1998) Hydrophobic mismatch between proteins and lipids in membranes. Biochim. Biophys. Acta 1376, 401–415. 28. Dumas, F., Lebrun, M.C. & Tocanne, J.F. (1999) Is the protein/ lipid hydrophobic matching principle relevant to membrane organization and functions? FEBS Lett. 458, 271–277. 29. Piknova ´ ,B.,Pe ´ rochon, E. & Tocanne, J F. (1993) Hydrophobic mismatch and long-range protein/lipid interactions in bacterio- rhodopsin/phosphatidylcholine vesicles. Eur. J. Biochem. 218, 385–396. 30. Tocanne, J.F., Cezanne, L., Lopez, A., Piknova, B., Schram, V., Tournier, J.F. & Welby, M. (1994) Lipid domains and lipid/pro- tein interactions in biological membranes. Chem. Phys. Lipids 73, 139–158. 31. Shaikh, S.R., Brzustowicz, M.R., Gustafson, N., Stillwell, W. & Wassall, S.R. (2002) Monounsaturated PE does not phase- separate from the lipid raft molecules sphingomyelin and choles- terol: role for polyunsaturation? Biochemistry 41, 10593–10602. 32. Welby, M., Poquet, Y. & Tocanne, J.F. (1996) The spatial dis- tribution of phospholipids and glycolipids in the membrane of the bacterium Micrococcus luteus varies during the cell cycle. FEBS Lett. 384, 107–111. 33. Andersson, A.S., Demel, R.A., Rilfors, L. & Lindblom, G. (1998) Lipids in total extracts from Acholeplasma laidlawii Apackmore closely than the individual lipids. Monolayers studied at the air– water interface. Biochim. Biophys. Acta 1369, 94–102. 34. Lindblom, G., Ora ¨ dd, G., Rilfors, L. & Morein, S. (2002) Reg- ulation of lipid composition in Acholeplasma laidlawii and Escherichia coli membranes: NMR studies of lipid lateral diffusion at different growth temperatures. Biochemistry 41, 11512–11515. 35. Christiansson, A. & Wieslander, A ˚ . (1980) Control of membrane polar lipid composition in Acholeplasma laidlawii A by the extent of saturated fatty acid synthesis. Biochim. Biophys. Acta 595, 189– 199. 36. Rilfors, L. (1985) Difference in packing properties between iso and anteiso methyl-branched fatty acids as revealed by incorporation into the membrane lipids of Acholeplasma laidlawii strain A. Biochim. Biophys. Acta 813, 151–160. 37. Greenberg-Ofrath, N., Terespolosky, Y., Kahane, I. & Bar, R. (1993) Cyclodextrins as carriers of cholesterol and fatty acids in cultivation of Mycoplasmas. Appl. Environ. Microbiol. 59, 547– 551. 38. Kates, M. (1972) Techniques of Lipidology: Isolation, Analysis and Identification of Lipids. North-Holland Publishing Co., Amsterdam. 39. Karlsson, O.P., Dahlqvist, A., Vikstro ¨ m, S. & Wieslander, A ˚ . (1997) Lipid dependence and basic kinetics of the purified 1,2- diacylglycerol 3-glucosyltransferase from membranes of Achole- plasma laidlawii. J. Biol. Chem. 272, 929–936. 40. Eriksson, L., Johansson, E., Kettaneh-Wold, N., Wikstro ¨ m, C. & Wold, S. (2000) Design of Experiments: Principles and Applica- tions. Learnways, Stockholm. 41. MacDonald, R.C., MacDonald, Ruby, I., Menco, Bert, PhM., Takeshita, Keizo, Subbarao, Nanda, K. & Hu, Lan-rong. (1991) Small-volume extrusion apparatus for preparation of large uni- lamellar vesicles. Biochim. Biophys. Acta 1061, 297–303. 42. Boldyrev, A.A., Lopina, O.D. & Prokopjeva, V.D. (1984) Diphenylhexatriene and pyrene as tools for characterization of biological membranes. Biochem. Int. 8, 851–859. 43. Gidwani, A., Holowka, D. & Baird, B. (2001) Fluorescence ani- sotropy measurements of lipid order in plasma membranes and lipid rafts from rbl-2h3 mast cells. Biochemistry 40, 12422–12429. 44. Galla, H.J. & Sackmann, E. (1974) Lateral diffusion in the hydrophobic region of membranes: use of pyrene excimers as optical probes. Biochim. Biophys. Acta 339, 103–115. 45. Jones, M.E. & Lentz, B.R. (1986) Phospholipid lateral organiza- tion in synthetic membranes as monitored by pyrene-labeled phospholipids: effects of temperature and prothrombin fragment 1 binding. Biochemistry 25, 567–574. 46. Somerharju, P.J., Virtanen, J.A., Eklund, K.K., Vainio, P. & Kinnunen, P.K. (1985) 1-Palmitoyl-2-pyrenedecanoyl glycer- ophospholipids as membrane probes: evidence for regular dis- tribution in liquid-crystalline phosphatidylcholine bilayers., Biochemistry 24, 2773–2781. 47. Vauhkonen, M., Sassaroli, M., Somerharju, P. & Eisinger, J. (1990) Dipyrenylphosphatidylcholines as membrane fluidity probes. Relationship between intramolecular and intermolecular excimer formation rates. Biophys. J. 57, 291–300. 48.Barenholz,Y.,Cohen,T.,Haas,E.&Ottolenghi,M.(1996) Lateral organization of pyrene-labeled lipids in bilayers as determined from the deviation from equilibrium between pyrene monomers and excimers. J. Biol. Chem. 271, 3085–3090. 49. Engelke, M., Bojarski, P., Diehl, H.A. & Kubicki, A. (1996) Protein-dependent reduction of the pyrene excimer formation in membranes. J. Membr. Biol. 153, 117–123. 50. Templer, R.H., Castle, S.J., Curran, A.R., Rumbles, G. & Klug, D.R. (1998) Sensing isothermal changes in the lateral pressure in model membranes using di-pyrenyl phosphatidylcholine. Faraday Discussions 41–53, discussion 69–78. 51. Hauksson, J.B., Rilfors, L., Lindblom, G. & Arvidson, G. (1995) Structures of glucolipids from the membrane of Achole- plasma laidlawii strain A-EF22. III. Monoglucosyldiacylglycerol, diglucosyldiacylglycerol, and monoacyldiglucosyldiacylglycerol. Biochim. Biophys. Acta 1258, 1–9. 52. Tuominen, E.K., Holopainen, J.M., Chen, J., Prestwich, G.D., Bachiller, P.R., Kinnunen, P.K. & Janmey, P.A. (1999) Fluor- escent phosphoinositide derivatives reveal specific binding of gel- solin and other actin regulatory proteins to mixed lipid bilayers. Eur. J. Biochem. 263, 85–92. 53. Asgharian, B., Cadenhead, D.A., Mannock, D. & McElhaney, R.N. (2000) A comparative monolayer film behaviour study of monoglucosyl diacylglycerols containing linear, methyl iso-, and x-cyclohexyl fatty acids. Langmuir 16, 7315–7317. 54. Eriksson, P.O., Rilfors, L., Wieslander, A ˚ ., Lundberg, A. & Lindblom, G. (1991) Order and dynamics in mixtures of mem- brane glucolipids from Acholeplasma laidlawii studied by 2H NMR. Biochemistry 30, 4916–4924. 55. Li, L. (2001) Functional Regulation of Glucolipid Synthases in Membranes.Umea ˚ University, Sweden. 56. Vikstro ¨ m, S., Li, L. & Wieslander, A ˚ . (2000) The nonbilayer/ bilayer lipid balance in membranes. Regulatory enzyme in 1708 P. Storm et al.(Eur. J. Biochem. 270) Ó FEBS 2003 [...]... Holopainen, J.M., Subramanian, M & Kinnunen, P.K (1998) Sphingomyelinase induces lipid microdomain formation in a fluid phosphatidylcholine/sphingomyelin membrane Biochemistry 37, 17562–17570 Hinderliter, A., Almeida, P.F., Creutz, C.E & Biltonen, R.L (2001) Domain formation in a fluid mixed lipid bilayer modulated through binding of the C2 protein motif Biochemistry 40, 4181–4191 Kaiser, R.D & London,... 61 Lateral organization of A laidlawii lipids (Eur J Biochem 270) 1709 Acholeplasma laidlawii is stimulated by metabolic phosphates, activator phospholipids, and double-stranded DNA J Biol Chem 275, 9296–9302 Holopainen, J.M., Lehtonen, J.Y.A & Kinnunen, P.K.J (1997) Lipid microdomains in dimyristoylphosphatidylcholine-ceramide liposomes Chem Phys Lipids 88, 1–13 Holopainen, J.M., Subramanian, M & Kinnunen,... B., Beaucourt, J.P & Tocanne, J.F (1989) Transverse and lateral distribution of phospholipids and glycolipids in the membrane of the bacterium Micrococcus luteus Biochemistry 28, 3728–3737 ˚ 66 Berg, S & Wieslander, A (1997) Purification of a phosphatase which hydrolyzes phosphatidic acid, a key intermediate in glucolipid synthesis in Acholeplasma laidlawii A membranes Biochim Biophys Acta Bio Membr 4,... diphenylhexatriene (DPH) and its derivatives within membranes: comparison of different fluorescence quenching analyses of membrane depth Biochemistry 37, 8180–8190 Holopainen, J.M., Brockman, H.L., Brown, R.E & Kinnunen, P.K (2001) Interfacial interactions of ceramide with dimyristoylphosphatidylcholine: impact of the N-acyl chain Biophys J 80, 765–775 62 Hinz, H.-J., Kuttenreich, H., Meyer, R., Renner,... groups determine structure and phase behaviour of glycolipid membranes: Densitometric, calorimetric, and X-ray studies Biochemistry 30, 5125–5138 63 Pascher, I., Lundmark, M., Nyholm, P.G & Sundell, S (1992) Crystal structures of membrane lipids Biochim Biophys Acta 1113, 339–373 64 Garidel, P & Blume, A (1998) Miscibility of phospholipids with identical headgroups and acyl chain lengths differing by two . Lateral organization in Acholeplasma laidlawii lipid bilayer models containing endogenous pyrene probes Patrik Storm 1 ,LuLi 2 , Paavo Kinnunen 3,4 and. synthases. Table 2. Lateral interactions between A. laidlawii lipids in liposome bilayers. Changes in Ie/Im of pyrene probes upon variation in lipid amounts according