Báo cáo khoa học: Transphosphatidylation activity of Streptomyces chromofuscus phospholipase D in biomimetic membranes docx

Transphosphatidylation activity of Streptomyces chromofuscus phospholipase D in biomimetic membranes Karim El Kirat 1 , Annie-France Prigent 2 , Jean-Paul Chauvet 3 , Bernard Roux 1 and Franc¸oise Besson 1 1 Laboratoire de Physico-Chimie Biologique, UMR CNRS 5013, Villeurbanne, Lyon, France; 2 Laboratoire de Biochimie et Pharmacologie, UMR INSERM 585, Villeurbanne, Lyon, France; 3 Laboratoire d’Inge ´ nierie et de Fonctionnalization des Surfaces, UMR CNRS 5621, Ecully, Lyon, France The phospholipase D (PLD) from Streptomyces chromo- fuscus belongs to the superfamily of PLDs. All the enzymes included in this superfamily are able to catalyze both hydrolysis and transphosphatidylation activities. However, S. chromofuscus PLD is calcium dependent and is often described as an enzyme with weak transphosphatidylation activity. S. chromofuscus PLD-catalyzed hydrolysis of phospholipids in aqueous medium leads to the formation of phosphatidic acid. Previous studies have shown that phos- phatidic acid–calcium complexes are activators for the hydrolysis activity of this bacterial PLD. In this work, we investigated the influence of diacylglycerols (naturally occurring alcohols) as candidates for the transphosphati- dylation reaction. Our results indicate that the transphos- phatidylation reaction may occur using diacylglycerols as a substrate and that the phosphatidylalcohol produced can be directly hydrolyzed by PLD. We also focused on the surface pressure dependency of PLD-catalyzed hydrolysis of phospholipids. These experiments provided new informa- tion about PLD activity at a water–lipid interface. Our findings showed that classical phospholipid hydrolysis is influenced by surface pressure. In contrast, phosphatidyl- alcohol hydrolysis was found to be independent of surface pressure. This latter result was thought to be related to headgroup hydrophobicity. This work also highlights the physiological significance of phosphatidylalcohol produc- tion for bacterial infection of eukaryotic cells. Keywords: phospholipase D; Langmuir films; transphos- phatidylation; Streptomyces chromofuscus; diacylglycerol. Phospholipase D (PLD) catalyzes the hydrolysis of the phosphoester bond between the phosphatidyl moiety and the choline headgroup of phosphatidylcholine (PtdCho), liberating choline and phosphatidic acid (PA). More precisely, PLD catalyzes the cleavage of the P-O bond of PtdCho, as demonstrated previously [1]. The mechanism of this reaction involves a molecule of water for the nucleophile substitution on the phosphatidyl–enzyme intermediate. When the nucleophile is an alcohol, a phosphatidylalcohol is produced. This latter activity is called transphosphatidy- lation and is specific for the PLD [2]. The PLD from Streptomyces chromofuscus belongs to the PLD superfamily, together with some endonucleases, some helicases, and some lipid synthases. Most of these enzymes are capable of catalyzing the hydrolysis and/or formation of phosphodiester bonds [3]. However, S. chromofuscus PLD also presents interesting characteristics that allow it to be distinguished from other enzymes of the PLD superfamily. Yang & Roberts [4] have recently shown that S. chromo- fuscus PLD does not bear the classical HKD motif and this may explain its strong dependency on calcium. Phosphatase activity has also been attributed to S. chromofuscus PLD, which was able to catalyze para-nitrophenyl-phosphate hydrolysis but was inactive on lipidic phosphomonoesters, such as PA [5]. Most work concerning S. chromofuscus PLD has focused on its hydrolytic activity and especially on its activation by anionic lipids [6]. An involvement of this bacterial PLD on the aggregation, leakiness and fusion of vesicles has also been demonstrated [7]. PLD-catalyzed hydrolysis has been measured using several techniques and with biomimetic substrates [8]. In the case of phospholipid hydrolysis, the production of PA has been determined by radioactive assay using radiolabelled PtdCho, by pH-stat and by [ 1 H] NMR [9], by a choline oxidase electrode [10], by the Langmuir film technique on phospholipidic monolayers [11], and by polarization modulation infrared reflection-absorption Correspondence to F. Besson, Laboratoire de Physico-Chimie Bio- logique, UMR CNRS 5013, Bat. Chevreul, 43 Bd du 11/11/1918, F-69622 Villeurbanne, UCB-Lyon 1, France. Fax: + 33 4 72431543, Tel.: + 33 4 72431542, E-mail: f-besson@univ-lyon1fr Abbreviations: MLV, multilamellar vesicles; PA, phosphatidic acid; PLD, phospholipase D; PtdCho, phosphatidylcholine; Pam 2 (Pam[ 3 H]N)GroChoP, L -a-dipalmitoyl-[2-palmitoyl- 9,10- 3 H(N)]-sn-glycero-3-phosphocholine; PamOleGroEtP, 1-O-palmitoyl-2-O-oleoyl-sn-glycero-3-phosphoethanol; PamOle- GroBuP,1-O-palmitoyl-2-O-oleoyl-sn-glycero-3-phosphobutanol; Myr 2 GroChoP, L -a-dimyristoyl-sn-glycero-3-phosphocholine; Myr 2 Gro, 1,2-dimyristoyl-rac-glycerol; Myr 2 GroPA, L -a-dimyristoyl- sn-glycero-3-phosphatidic acid; PamLinGroEtnP-HNE, 1-O-palmi- toyl-2-O-linoleoyl-sn-glycero-3-phosphoethanolamine-4-hydroxy- nonenal; SM, sphingomyelin from bovine brain. (Received 11 July 2003, revised 17 September 2003, accepted 22 September 2003) Eur. J. Biochem. 270, 4523–4530 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03841.x spectroscopy at the air–water interface [12]. All of these studies concluded that PLD is activated by PA and calcium. However, no factor has as yet been identified which enhances the transphosphatidylation reaction. PLD from S. chromofuscus is often described as an enzyme that has only weak transphosphatidylation activity. However, Fried- man et al. [13] have shown that this PLD is able to catalyze an intramolecular transphosphatidylation reaction within lyso-PtdCho. This activity is responsible for the formation of cyclic lyso-PA, which can be further hydrolyzed by PLD to produce lyso-PA. In other words, PLD from S. chromo- fuscus is able to form transphosphatidylation products and can also catalyze their hydrolysis. In the present work, the S. chromofuscus PLD trans- phosphatidylation activity was investigated in model mem- branes (i.e. vesicles and monolayers). Our results highlight the involvement of diacylglycerols in a naturally occuring transphosphatidylation reaction catalyzed by PLD. More- over, monolayer experiments allowed us to estimate phos- phatidylalcohol hydrolysis under PLD action at the air– water interface. Comparison with the results obtained on lipidic vesicles led to the elaboration of a model mechanism of transphosphatidylation reaction catalyzed by S. chromo- fuscus PLD. Materials and methods Materials L -a-Dipalmitoyl-[2-palmitoyl-9,10- 3 H(N)]-sn-glycero-3-pho sphocholine [Pam 2 (Pam[ 3 H]N)GroChoP] was purchased from NEN (Life Science Products, Inc., Boston, MA, USA). Coomassie Brilliant Blue R and TLC aluminium sheets (Silica gel 60F 254 ) were from Merck (Darmstadt, Germany). 1-O-palmitoyl-2-O-oleoyl-sn-glycero-3-phos- phoethanol (PamOleGroEtP)and1-O-palmitoyl-2-O- oleoyl-sn-glycero-3-phosphobutanol (PamOleGroBuP) were purchased from Biomol Research Laboratories Inc. (Plymouth Meeting, PA, USA). L -a-Dimyristoyl-sn-glyc- ero-3-phosphocholine (Myr 2 GroChoP), 1,2-dimyristoyl- rac-glycerol (Myr 2 Gro), L -a-dimyristoyl-sn-glycero-3-phos- phatidic acid (Myr 2 GroPA), sphingomyelin from bovine brain (SM) and PLD from S. chromofuscus were purchased from Sigma Chemical Co. and used without further purification. SDS/PAGE analysis of PLD gave the same three bands as those obtained by Geng et al.[14].Triswas purchased from Roche Diagnostics. 1-O-palmitoyl-2-O- linoleoyl-sn-glycero-3-phosphoethanolamine-4-hydroxynonenal (PamLinGroEtnP-HNE) was a gift from M. Guichardant (Laboratoire de Biochimie et Pharmacologie, UMR INSERM 585, INSA de Lyon, France). Monolayer technique All experiments were performed at a constant temperature of 21 ± 0.1 °C. The film balance was built by R&K (Wiesbaden, Germany) and equipped with a Wilhemy-type surface-pressure measuring system. The subphase was aqueous buffer containing 120 l M CaCl 2 ,150m M NaCl, and 10 m M Tris/HCl, pH 8.0. The calcium concentration was sufficient to allow maximum PLD activity [11]. Phospholipids were spread at the air–water interface in hexane/ethanol (9 : 1, v/v), at a concentration of 0.175 m M , to reach a final quantity of 8.75 nmol of lipids. After 15 min of solvent evaporation, the monolayer was compressed to a lateral pressure of 35 mNÆm )1 to obtain a control p-A isotherm. Then, the pressure was fixed at 30 mNÆm )1 and theenzyme(15 lg of protein) was injected into the subphase after monolayer stabilization. The subphase was stirred using a magnetic stirrer spinning at 100 r.p.m. Surface compressional moduli were calculated from the pressure- area data obtained from the monolayer compressions, using the following equation [15]: Ks ¼ÀAxdp=dA where A is the molecular area at the indicated surface pressure p. High Ks values correspond to low interfacial elasticity among packed lipids forming a monolayer [16]. This suggests that the higher the Ks value of a monolayer, the greater the monolayer rigidity. Vesicle preparation Lipids (i.e. Myr 2 GroChoP and Myr 2 Gro) were dissolved in chloroform at 5 m M . The lipid mixture, containing varying molar ratios of Myr 2 Gro, were dried under nitrogen for 2 h. Then, the lipidic film was resuspended by vigorous agitation for 1 min in 120 l M CaCl 2 , 150 m M NaCl and 100 m M Tris/HCl, pH 8.0, to achieve a 5 m M final concentration of total lipids. The lipid suspension was frozen in liquid nitrogen for 5 min and then heated for 10 min at 40 °Cina thermostated bath. This vortex freeze-warming was repea- ted three times to obtain the multilamellar vesicles (MLV). Radioactive assay PLD activity on lipidic bilayers was determined using mixed radiolabelled vesicles of Myr 2 GroChoP/Myr 2 Gro, at various ratios, with 20 lCi Pam 2 (Pam[ 3 H]N)GroChoP, in a final volume of 200 lL. The MLV were prepared in exactly the same way as described above (Vesicle prepar- ation) to achieve the same final concentration (5 m M ). To measure PLD activity, vesicles (corresponding to 2 lCi) were incubated with 1 lg of PLD at 37 °C in a final volume of 200 lL. The reaction was stopped with 2 mL of chloroform/methanol/0.1 M HCl (1 : 1 : 0.002, v/v/v) and 1 mL of 1 M HCl containing 5 m M EDTA. The tubes were agitated vigorously and then centrifuged at 400 g for 5min at 4°C. The organic phases containing the lipids were then transferred to a TLC plate (Silica gel 60F 254 ) with Myr 2 GroChoP,Myr 2 GroPA, PamOleGroEtP, PamOleGroBuP and Myr 2 Gro standards. After develop- ment in ethyl acetate/isooctane/acetic acid (90 : 50 : 20, v/v/v), the plate was developed using Coomassie Brilliant Blue R250 [17]. The spots were then scraped off and counted for radioactivity determination (Wallac Winspec- tral TM 1414 Liquid Scintillation Counter; Wallac, Turku, Finland). Calcium content determination Plasma emission spectroscopy (Service Central d’Analyse, CNRS, Vernaison, France) was used to determine the calcium content of buffers. 4524 K. El Kirat et al. (Eur. J. Biochem. 270) Ó FEBS 2003 Results Influence of surface pressure on PLD-catalyzed hydrolysis of phospholipids In these experiments, PLD activity was recorded as described previously [11]. First, the compression isotherm of the lipid was measured and then the monolayer was stabilized at a constant surface pressure. Under these conditions, and after PLD injection into the subphase, the apparent molecular area of the monolayer decreases with time, indicating the formation of a lipidic product with a smaller polar head and thus presenting a lower molecular area than the phospholipidic substrate. For example, if Myr 2 GroChoP (60 A ˚ 2 per molecule at 30 mNÆm )1 surface pressure) is the substrate, then it will be converted into Myr 2 GroPA (42 A ˚ 2 per molecule at 30 mNÆm )1 )atthe air–water interface under PLD-catalyzed hydrolysis [11]. Therefore, this reaction, at a constant surface pressure, will lead to a decrease of molecular area with time. Three lipids (PtdCho, SM and PamOleGroBuP)were tested for PLD interfacial activity (Fig. 1). All of these lipids are substrates of S. chromofuscus PLD. For the two naturally occuring lipids – PtdCho and SM (Fig. 1A,B) – PLD activity is strongly dependent on surface pressure. The representation of PLD activity as a function of surface pressure (Fig. 2) shows sigmo curves for these two lipids, with a cut-off at % 25 mNÆm )1 for PtdCho and SM. As shown in Fig. 1C, PLD-catalyzed hydrolysis of PamOleGroBuP is not dependent on surface pressure, as the slope of apparent molecular area decrease was similar at different pressures. According to Fig. 2, PLD-catalyzed hydrolysis of PamOleGroBuP always occurs with approxi- mately the same slope, even at high surface pressures. Lee et al. [18] have previously described the headgroup insertion of phosphatidylalcohols within the hydrophobic core of membrane bilayers. This phenomenon is obviously a result of the unusual hydrophobicity of the phosphatidylalcohols’ headgroup. Such hydrophobic interactions, whilst driving the alkyl headgroup of PamOleGroBuP into the membrane, lead to an unusual exposure of the phosphate group at the membrane–water interface. Hence, it may be postulated that, regardless of the surface pressure, the phosphate group of PamOleGroBuP would always be exposed to the water and that this would enhance PLD activity. Fig. 1. Influence of surface pressure on phospholipase D (PLD)-cata- lyzed hydrolysis of phospholipids. Three lipids were tested as substrates for PLD activity at the air–water interface: phosphatidylcholine (A), sphingomyelin from bovine brain (B) and 1-O-palmitoyl-2-O-oleoyl- sn-glycero-3-phosphobutanol (C). After isotherm measurement, the monolayer was stabilized at constant surface pressure: 10 (d), 15 (n), 20 (m), 25 (j)and30(h)mNÆm )1 . Then, the enzyme was injected into the subphase and PLD activity was recorded continuously as the apparent molecular area decreases with time. The apparent molecular area was normalized to compare the different reactions. The subphase was 120 l M CaCl 2 ,150m M NaCl and 10 m M ,TrispH8.0,andthe temperature was fixed at 21 °C. Fig. 2. Influence of surface pressure on phospholipase D (PLD) activity at the air–water interface. PLD activity was expressed in min )1 as the velocity of normalized apparent molecular area decrease. These values were calculated from the curves given in Fig. 1 for phosphatidylcholine (j), sphingomyelin from bovine brain (h)and1-O-palmitoyl-2-O- oleoyl-sn-glycero-3-phosphobutanol (m). Subphase and temperature were as described in the legend to Fig. 1. Ó FEBS 2003 S. chromofuscus PLD activation by diacylglycerol (Eur. J. Biochem. 270) 4525 Influence of phospholipid headgroup hydrophobicity on PLD activity In these experiments, we measured PLD activity at the air– water interface using substrates with different degrees of headgroup hydrophobicity (Fig. 3). These results allow comparison to be made between a naturally occuring phospholipid, such as PtdCho, and phospholipids present- ing an aliphatic chain as a headgroup, such as PamOle- GroEtP,PamOleGroBuP and PamLinGroEtnP-HNE. The isotherms of these three atypical phospholipids were measured. PamOleGroEtP and PamOleGroBuP present approximately the same molecular area (60 and 70 A ˚ 2 per molecule, respectively) for a surface pressure of 30 mNÆm )1 . The molecular area of PamLinGroEtnP-HNE was % 90 A ˚ 2 per molecule (data not shown). Their Ks values were also calculated from the p-A isotherms for a surface pressure of 30 mNÆm )1 . According to previously published results [19,20], a surface pressure of 30 mNÆm )1 is thought to approximate to the internal pressure of biological mem- branes. Ks values were found within surface pressures of 70–80 mNÆm )1 , which reveal a compressibility similar to that of Myr 2 GroChoP. As shown in Fig. 3, PtdCho is slowly hydrolyzed at a surface pressure of 30 mNÆm )1 , while the nonclassical phospholipids are rapidly converted into PA. Morever, EtP, which presents only a two-carbon chain headgroup, seems to be a better substrate of PLD than PtdCho. When the hydrophobic headgroup is composed of at least four carbons, which is the case for PamOleGroBuP and PamLinGroEtnP-HNE, the rate of their PLD-cata- lyzed hydrolysis is maximal (60 times higher than that of PtdCho). One may postulate that the acyl-chain headgroup has a tendency to penetrate into the hydrophobic core of the membrane and that this could expose the phosphate group to PLD activity. Under these conditions, phosphatidyl- alcohols would always be hydrolyzed by PLD activity as soon as they were synthesized within membranes. Properties of mixed Myr 2 GroCho P /Myr 2 Gro monolayers Diacylglycerols are naturally occuring alcohols that are transiently produced by phospholipase C in biological membranes [21]. Therefore, we can assume that these neutral lipids could serve as natural nucleophiles within membranes in a PLD-catalyzed transphosphatidylation reaction. In order to test this hypothesis, mixed monolayers of Myr 2 GroChoP/Myr 2 Gro were prepared and compressed at the air–water interface (Fig. 4A). These results indicate that Myr 2 Gro presents a molecular area of % 40 A ˚ 2 per molecule. This value is consistent with the small polar head of Myr 2 Gro. The isotherms show that increasing the amount of Myr 2 Gro in Myr 2 GroChoP monolayers leads to a decrease of the mixed film molecular area. The isotherms present a modification of slope during compression (Fig. 4A). This could be the result of an Fig. 3. Influence of the headgroup hydrophobicity on phospholipase D (PLD) activity. Phosphatidylcholine (a), 1-O-palmitoyl-2-O-oleoyl- sn-glycero-3-phosphoethanol (b), 1-O-palmitoyl-2-O-linoleoyl-sn- glycero-3-phosphoethanolamine-4-hydroxynonenal (c) and 1-O-pal- mitoyl-2-O-oleoyl-sn-glycero-3-phosphobutanol (d) were tested as substrates for PLD activity at a constant surface pressure of 30 mNÆm )1 . The apparent molecular area was normalized to compare the different reactions. The chemical structure of PamLinGroEtnP- HNE is given within the figure. Subphase and temperature were as described in the legend to Fig. 1. Fig. 4. Pressure-area isotherms (A) and surface compressional modulus (B) of L -a-dimyristoyl-sn-glycero-3-phosphocholine (Myr 2 GroChoP) monolayers containing an increasing molar percentage of 1,2-dimyris- toyl-rac-glycerol (Myr 2 Gro). Myr 2 Gro molar ratios in Myr 2 GroChoP monolayer were: a (0%); b (5%); c (10%); d (20%); e (40%); f (50%); and g (100%). The surface compressional modulus (Ks) value was calculated from pressure-area isotherms and was plotted as a function of surface pressure. Subphase and temperature were as described in the legend to Fig. 1. 4526 K. El Kirat et al. (Eur. J. Biochem. 270) Ó FEBS 2003 expanded-liquid to condensed-liquid phase transition that seems to occur for low pressure values at high Myr 2 Gro percentages. This transition could be explained by the highly ordered chain organization induced by Myr 2 Gro at the air– water interface, which would lead to a global rigidization of the monolayer. It should also be borne in mind that diacylglycerol induces lateral segregation of lipids, which leads to domains enriched in diacylglycerol with a low headgroup steric encumbrance [22]. Surface compressional moduli were calculated from the p-A isotherms for each monolayer in order to estimate membrane rigidity caused by the presence of Myr 2 Gro in the mixtures (Fig. 4B). These results confirm the phase transition caused by Myr 2 Gro mixed with Myr 2 GroChoP in monolayers. This transition corresponds to a sudden change in the slope of the curve; for example, % 17 mNÆm )1 in surface pressure for the Myr 2 GroChoP/Myr 2 Gro 90 : 10 (mol/mol) mixture. The surface pressure value of this transition decreases with increasing amounts of Myr 2 Gro in the monolayer. Morever, the curves show that the presence of Myr 2 Gro does not change the physical properties of the monolayer until a molar percentage of 5–10 is reached. Above this limit, Myr 2 Gro induces an increase in Ks values, indicating an increased rigidity of the monomolecular film. Influence of Myr 2 Gro on PLD activity at the air–water interface Mixed Myr 2 GroChoP/Myr 2 Gro monolayers were com- pressed and stabilized at a surface pressure of 30 mNÆm )1 . PLD was injected into the subphase and its activity against the lipidic monolayers was measured by monitoring the decrease in apparent molecular area, along with time, at 30 mNÆm )1 (Fig. 5A). These results indicate an increase of PLD-catalyzed Myr 2 GroChoP hydrolysis occuring in the presence of Myr 2 Gro. By increasing the initial Myr 2 Gro content in the monolayer, the apparent molecular area decrease seems to occur more rapidly and with a greater slope (Fig. 5). PLD activation starts at low percentages (1.25 molar percentage) of Myr 2 Gro and seems to be maximal between 3 and 5 molar percentage. In previous work [11], we reported that PLD activity is dependent on membrane compressibility, so we represented the slope of the reaction and the variation of the monolayer compressibility (Ks) as a function of Myr 2 Gro molar percentage (Fig. 5B). This result showed that 3 molar percentage is almost sufficient for the maximal activation of PLD, while the Ks value is only slightly modified. As shown previously, only low values of Ks (ranging from 30 to 60 mNÆm )1 ) are able to induce maximal PLD activity [11]. Therefore, it seems that the Myr 2 Gro- induced activation of PLD is not dependent on membrane compressibility, but may occur through a transphosphati- dylation activity. The effect of another alcohol on PLD activity at the air–water interface was also tested to determine a transphosphatidylation mechanism. Monolay- ers of octanol (OctOH) mixed with Myr 2 GroChoP were prepared and compressed to 30 mNÆm )1 . After pressure stabilization, PLD was injected into the subphase. In this case, we obtained PLD activities of % 0.006 min )1 and 0.0128 min )1 for 4 and 10 molar percentage OctOH within the monolayer, respectively. These values are in the range of those obtained for Myr 2 Gro (Fig. 5B) and no change of the Myr 2 GroChoP/OctOH isotherms could be observed as compared to Myr 2 GroChoP alone. Influence of Myr 2 Gro on PLD activity in vesicles van Blitterswijk & Hilkmann [21] have proved the existence of a transient lipid, called bis(PA), in mammalian cells, which is produced via the PLD-catalyzed transphosphati- dylation reaction in the presence of diacylglycerol. In order to detect bis(PA) production in biomimetic membranes incubated with S. chromofuscus PLD, we prepared vesi- cles, containing different molar percentages of Myr 2 Gro mixedwithMyr 2 GroChoP, in the presence of Pam 2 (Pam[ 3 H]N)GroChoP radiolabelled on its sn-2 fatty acid. The reaction was stopped at different time-points using a concentrated EDTA solution, lipids were separated by TLC, and the levels of diacylglycerol, bis(PA) and PA were measured by radioactivity. Low initial percentages of Myr 2 Gro mixed in Myr 2 GroChoP vesicles were tested, Fig. 5. Influence of 1,2-dimyristoyl-rac-glycerol (Myr 2 Gro) molar per- centage in the monomolecular film of L -a-dimyristoyl-sn-glycero-3- phosphocholine (Myr 2 GroChoP) on phospholipase D (PLD) activity. The apparent molecular area was normalized to compare the different reactions at a constant surface pressure of 30 mNÆm )1 .TheMyr 2 Gro molar ratio in the Myr 2 GroChoP monolayer was: a (0%); b (1.25%); c (2.5%); d (3%); e (5%); f (10%). Figure 5B shows PLD activity (d) and Ks (h) dependency on the Myr 2 Gro molar ratio. Ks values were reported from Fig. 4 and PLD activity was expressed as the decrease in slope of normalized molecular area observed in Fig. 5A. Subphase and temperature were as described in the legend to Fig. 1. Ó FEBS 2003 S. chromofuscus PLD activation by diacylglycerol (Eur. J. Biochem. 270) 4527 but no significant production of radiolabelled lipids was observed (data not shown). However, at 10–30% Myr 2 Gro initially present in Myr 2 GroChoP vesicles, we observed the simultaneous production of PA, diacylglycerol and bis(PA) (Fig. 6). Increasing the initial content of Myr 2 Gro led to higher amounts of each of these three lipids. Therefore, diacylglycerol was considered as an activator for PLD activity. Furthermore, it was impossible to produce PA in amounts greater than 30% (Fig. 6A). This is because of the shape of this lipid, which is not compatible with the existence of a bilayer in the presence of calcium. Under these conditions, lipidic structures enriched in PA would be stabilized in inverted micelles that have a tendency to form aggregates. In such aggregates, PLD would not be able to reachthepolarheadofPtdChotocatalyzeitshydrolysis. Diacylglycerol is the second major product of this reaction (Fig. 6B). However, only small amounts of this neutral lipid were detected after the reaction. Under optimal conditions, i.e. 30% of Myr 2 Gro initially present in the vesicles, 5% radiolabelled diacylglycerol is detected after 10 min vs. 30% radiolabelled PA. Radioactive diacylglyc- erol can be produced in two different ways: first, through the PLD-catalyzed hydrolysis of bis(PA), which can lead, in theory, to equal production of radiolabelled PA or diacyl- glycerol, as bis(PA) is symmetric; or, second, via the phosphatase activity of S. chromofuscus PLD that would have converted PA into diacylglycerol. However, as PA is not a substrate for S. chromofuscus PLD phosphatase activity, according to the observations of Zambonelli & Roberts [5], this latter explanation is not feasible. Concern- ing bis(PA) (Fig. 6C), it seems that it is only produced in small amounts, from 0.6 to 0.7% of the radioactivity recovered, indicating that it is a reaction intermediate, which cannot accumulate in membranes. Discussion The monolayer technique is a powerful method for assaying the PLD-catalyzed hydrolysis of various lipids. This method requires only small amounts of lipids and provides infor- mation on the compressibility of lipidic membranes. Another important advantage of this technique is the possibility of forming monolayers with lipids that cannot form vesicles. This is because the monolayer interface is planar, unlike the liposome interface which is curved. Under these conditions, PLD activity can lead to total hydrolysis of phospholipids, which cannot be obtained when vesicles are used. Furthermore, all the lipids spread at the air–water interface are in contact with the subphase, whereas in vesicles, lipids can divide between the two leaflets of the bilayer. In the latter case, only a fraction of the lipids are accessible to proteins. Phosphatidylalcohols are substrates of S. chromofuscus PLD Surface pressure dependency of PtdCho hydrolysis suggests that insertion of the enzyme into the membrane is a prerequisite for PLD activity [7]. This phenomenon has previously been described on PLC activity towards phos- phatidylinositol 4,5-bisphosphate monolayers [23]. PLD from S. chromofuscus does not present the classical HKD motif and is calcium dependent [4,5]. Furthermore, transphosphatidylation activity of this PLD has been widely discussed and several authors concluded on a low ability of this enzyme to catalyze this type of reaction. Friedman et al. [13] have reported a transferase activity observed by NMR with lyso-PtdCho as substrate. They have shown that Fig. 6. Role of 1,2-dimyristoyl-rac-glycerol (Myr 2 Gro) on L -a-dimyris- toyl-sn-glycero-3-phosphocholine (Myr 2 GroChoP) hydrolysis catalyzed by phospholipase D (PLD) in liposomes. We prepared vesicles of Myr 2 GroChoP [radiolabelled with L -a-dipalmitoyl-[2-palmitoyl- 9,10- 3 H(N)]-sn-glycero-3-phosphocholine (Pam 2 (Pam[ 3 H]N)Gro- ChoP)]mixedwith10%(s), 20% (h)and30%(j)molarratioof Myr 2 Gro. The final lipid concentration for all the assays was 5 m M in 120 l M CaCl 2 ,150m M NaCl, 10 m M Tris, pH 8.0, and the tempera- ture was fixed at 37 °C. Aliquots were taken at 2.5-min intervals and the reaction was stopped with concentrated EDTA solution. After lipid extraction and TLC separation, radioactivity was counted for phosphatidic acid (PA) (A), diacylglycerol (B) and bis(PA) (C). According to van Blitterswijk & Hilkmann [21], bis(PA) comigrates with 1-O-palmitoyl-2-O-oleoyl-sn-glycero-3-phosphobutanol in our TLC system. Radioactivity for each lipid was expressed as the per- centage of total radiactivity recovered after TLC (100% corresponds to 250 000 d.p.m.). 4528 K. El Kirat et al. (Eur. J. Biochem. 270) Ó FEBS 2003 conversion of this substrate into lyso-PA occurs via the formation of cyclic lyso-PA obtained by intramolecular transphosphatidylation of the lyso-PtdCho. Then, lyso-PA is produced from cyclic lyso-PA by hydrolase activity of PLD. Here, we report the ability of PLD to catalyze the hydrolysis of phosphatidylalcohol without any dependency on the surface pressure. This could be a result of the hydrophobicity of the phosphatidylalcohol headgroup. Indeed, several physical studies with liposomes showed the tendency of such types of phospholipid headgroup to insert into the membranes, leading to the exposure of the lipid phosphate group to the buffer [18]. Such behavior of phosphatidylalcohols in a monolayer would result in an increase of the PLD activity, independently of the surface pressure, as compared with the phosphate group of other phospholipids. The comparison between PtdCho, PamOleGroEtP, PamOleGroBuP and PamLinGroEtnP-HNE provided information on the influence of carbon-chain headgroup length on PLD activity. It seems that maximum PLD activity is obtained with a hydrophobic headgroup contain- ing at least four carbons. Therefore, if PLD from S. chro- mofuscus can catalyze a transphosphatidylation reaction, the phosphatidylalcohol produced will be a better substrate than PtdCho. As a consequence, a phosphatidylalcohol cannot accumulate in membranes in the presence of S. chromofuscus PLD. Furthermore, our results show a rapid hydrolysis of PamLinGroEtnP-HNE catalyzed by PLD at the air–water interface at a surface pressure of 30 mNÆm )1 . This adduct is generated by condensation of 4-HNE and PE, and this reaction can also lead to Schiff base adduct formation [24]. Those N-acylated lipids are generated after cell oxidative stress induced by reactive oxygen species. Hence, the presence of such adducts within cell membranes could favor PLD activity. S. chromofuscus PLD activation by diacylglycerol Previous work reported the activation of S. chromofuscus PLD by diacylglycerol [10] but the mechanism of this still remains unknown. We tested PLD activity on mixed Myr 2 GroChoP/Myr 2 Gro monolayers. An activating effect of the diglyceride was detected at low molar ratios (1.25%), with a maximal effect occurring at 3 molar percentage Myr 2 Gro. These low molar fractions of diacylglycerol are close to physiological concentrations resulting from PLC activity on membranes. On the basis of these low propor- tions of diacylglycerol, we conclude that steric encumbrance and physical properties of the membranes are not respon- sible for the increased PLD activity. Experiments with mixed PtdCho/diacylglycerol vesicles also showed an increased PLD-catalyzed hydrolysis of PtdCho, at higher percentages than those observed with monolayers. The liposomes used here are multilamellar vesicles, so all the Myr 2 Gro present in the membranes is not accessible to PLD as the vesicles can be encapsulated one inside the other. Furthermore, the situation in bilayers is different from that obtained with monolayers: Myr 2 Gro can be partitioned between the two leaflets of the bilayer, thus only a fract- ion of total Myr 2 Gro is exposed to PLD. Therefore, the maximum enzyme activation will be obtained in bilayers with higher percentages of diacylglycerol than in monolayers. However, radioactivity detection permitted the quantification of radiolabelled PA, diacylglycerol and bis(PA). As PLD from S. chromofuscus does not possess a phosphatase activity with the ablility to generate diacyl- glycerol from PA [5], all the radiolabelled diacylglycerol must be produced through another reaction. A second possible explanation for the production of diacylglycerol could have been the presence of a contaminant PLC activity, but no radiolabelled diacylglycerol (and no radio- labelled bis(PA) or PA could be detected using pure PtdCho liposomes. The production of radiolabelled bis(PA) allowed us to elaborate a mechanism for diacylglycerol activation of S. chromofuscus PLD. In a first-step reaction, PLD cata- lyzes a transphosphatidylation reaction involving PtdCho as a substrate and diacylglycerol as the nucleophile. This reaction will produce bis(PA), which could be radiolabelled if the substrate is radiolabelled on its acyl-chains. The bis(PA) generated in the membranes should fully expose its phosphate group to PLD activity, as observed with other phosphatidylalcohols in monolayer experiments. Therefore, in a second step of the reaction, bis(PA) will be converted into PA plus diacylglycerol via PLD-phosphodiesterase activity. One should bear in mind that PA or diacylglycerol, owing to their small headgroup, cannot be produced together up to more than 40–50 molar percent in vesicles. As bis(PA) is a symmetric substrate of PLD, we can presume that its hydrolysis will lead to equal proportions of PA and diacylglycerol (Fig. 7). However, this was not observed in our results. A possible explanation for this could be that PA is also an activator of S. chromofuscus PLD; Fig. 7. Schematic mechanism of the transphosphatidylation reaction involving diacylglycerol and radiolabelled phosphatidylcholine. The radiolabelled acyl chain of L -a-dipalmitoyl-[2-palmitoyl-9,10- 3 H(N)]- sn-glycero-3-phosphocholine is represented in shaded boxes. In a first step, radiolabelled dipalmitoylphosphatidyl-phospholipase D (PLD) is produced. Then, a nucleophilic substitution occurs with 1,2-dimyris- toyl-rac-glycerol (Myr 2 Gro) as an alcohol within the membrane. This yields a phosphodiester intermediate, bis(PA). This product is a symetric substrate for Streptomyces chromofuscus PLD. Thus, its cleavage can occur, even in positions 1 and 2, with equal probability around a phosphate group. Therefore, this reaction will theoretically produce equal proportions of two lipid couples: radiolabelled diacyl- glycerol plus cold PA and cold diacylglycerol plus radiolabelled PA (see text for more details). Ó FEBS 2003 S. chromofuscus PLD activation by diacylglycerol (Eur. J. Biochem. 270) 4529 when a critical percentage of PA is reached, hydrolysis will be the major reaction catalyzed by PLD instead of the transphosphatidylation involving diacylglycerol. In conclusion, bacterial PLDs are often excreted as well as PLC and are described as virulence determinants [25]; therefore these two enzymes could act in synergy to permit internalization of bacteria into host cells. As a result of its ability to form a complex with calcium, PA can favor divalent cation-dependent fusion of membranes. Thus, PLD-catalyzed formation of PA activated by diacylglycerol could enhance fusion between bacteria and the host cell membrane. Acknowledgements We thank Dr Caroline Elston for reviewing the English version of this manuscript. References 1. Holbrook, P.G., Pannell, L.K. & Daly, J.W. (1991) Phospholipase D-catalyzed hydrolysis of phosphatidylcholine occurs with P-O bond cleavage. Biochim. Biophys. Acta 1084, 155–158. 2. Yu, C.H., Liu, S.Y. & Panagia, V. (1996) The transphos- phatidylation activity of phospholipase D. Mol. Cell. Biochem. 157, 101–105. 3. Ponting, C.P. & Kerr, I.D. (1996) Thermal stability of the three domains of streptokinase studied by circular dichroism and nuclear magnetic resonance. Protein Sci. 5, 914–922. 4. Yang, H. & Roberts, M.F. (2002) Cloning, overexpression, and characterization of a bacterial Ca 2+ -dependent phospholipase D. Protein Sci. 11, 2958–2968. 5. Zambonelli, C. & Roberts, M.F. (2003) An iron-dependent bac- terial phospholipase D reminiscent of purple acid phosphatases. J. Biol. Chem. 278, 13706–13711. 6. Stieglitz, K.A., Seaton, B.A. & Roberts, M.F. (1999) The role of interfacial binding in the activation of Streptomyces chromofuscus phospholipase D by phosphatidic acid. J. Biol. Chem. 274, 35367– 35374. 7. Stieglitz, K.A., Seaton, B.A. & Roberts, M.F. (2001) Binding of proteolytically processed phospholipase D from Streptomyces chromofuscus to phosphatidylcholine membranes facilitates vesicle aggregation and fusion. Biochemistry 40, 13954–13963. 8. Morris, A.J., Frohman, M.A. & Engebrecht, J. (1997) Measure- ment of phospholipase D activity. Anal. Biochem. 252,1–9. 9. Geng, D., Chura, J. & Roberts, M.F. (1998) Activation of phos- pholipase D by phosphatidic acid. Enhanced vesicle binding, phosphatidic acid–Ca 2+ interaction, or an allosteric effect? J. Biol. Chem. 273, 12195–12202. 10. Yamamoto, I., Konto, A., Handa, T. & Miyajima, K. (1995) Regulation of phospholipase D activity by neutral lipids in egg- yolk phosphatidylcholine small unilamellar vesicles and by cal- cium ion in aqueous medium. Biochim. Biophys. Acta 1233, 21–26. 11. ElKirat,K.,Besson,F.,Prigent,A.F.,Chauvet,J.P.&Roux,B. (2002) Role of calcium and membrane organization on phospholipase D localization and activity. Competition between a soluble substrate and an insoluble substrate. J. Biol. Chem. 277, 21231–21236. 12. Estrela-Lopis, I., Brezesinski, G. & Mohwald, H. (2001) Dipalmitoyl–phosphatidylcholine/phospholipase D interactions investigated with polarization-modulated infrared reflection absorption spectroscopy. Biophys. J. 80, 749–754. 13. Friedman, P., Haimovitz, R., Markman, O., Roberts, M.F. & Shinitzky, M. (1996) Conversion of lysophospholipids to cyclic lysophosphatidic acid by phospholipase D. J. Biol. Chem. 271, 953–957. 14. Geng, D., Baker, D.P., Foley, S.F., Zhou, C., Stieglitz, K.A. & Roberts, M.F. (1999) A 20-kDa domain is required for phos- phatidic acid-induced allosteric activation of phospholipase D from Streptomyces chromofuscus. Biochim. Biophys. Acta 1430, 234–244. 15. Dumaual, A.C., Jenski, L.J. & Stillwell, W. (2000) Liquid crys- talline/gel state phase separation in docosahexaenoic acid-con- taining bilayers and monolayers. Biochim. Biophys. Acta 1463, 395–406. 16. Li, X.M., Momsen, M.M., Smaby, J.M., Brockman, H.L. & Brown R.E. (2001) Cholesterol decreases the interfacial elasticity and detergent solubility of sphingomyelins. Biochemistry 40, 5954– 5963. 17. Bechoua, S., Dubois, M., Ne ´ moz, G., Lagarde, M. & Prigent, A F. (1998) Docosahexaenoic acid lowers phosphatidate level in human activated lymphocytes despite phospholipase D activation. J. Lipid Res. 39, 873–883. 18. Lee, Y.C., Zheng, Y.O., Taraschi, T.F. & Janes, N. (1996) Hydrophobic alkyl headgroups strongly promote membrane curvature and violate the headgroup volume correlation due to ÔheadgroupÕ insertion. Biochemistry 35, 3677–3684. 19. Marsh, D. (1996) Lateral pressure in membranes. Biochim. Biophys. Acta 1286, 183–223. 20. Silvius, J.R. (2003) Role of cholesterol in lipid raft formation: lessons from lipid model systems. Biochim. Biophys. Acta 1610, 174–183. 21. van Blitterswijk, W.J. & Hilkmann, H. (1993) Rapid attenuation of receptor-induced diacylglycerol and phosphatidic acid by phospholipase D-mediated transphosphatidylation: formation of bisphosphatidic acid. EMBO J. 12, 2655–2662. 22. Jimenez-Monreal, A.M., Villalain, J., Aranda, F.J. & Gomez- Fernandez, J.C. (1998) The phase behavior of aqueous dispersions of unsaturated mixtures of diacylglycerols and phospholipids. Biochim. Biophys. Acta. 1373, 209–219. 23. Rebecchi, M., Boguslavsky, V., Boguslavsky, L. & McLaughlin, S. (1992) Phosphoinositide-specific phospholipase C-delta 1: effect of monolayer surface pressure and electrostatic surface potentials on activity. Biochemistry 31, 12748–12753. 24. Guichardant, M., Taibi-Tronche, P., Fay, L.B. & Lagarde, M. (1998) Covalent modifications of aminophospholipids by 4-hydroxynonenal. Free Rad. Biol. Med. 25, 1049–1056. 25. McNamara, P.J., Cuevas, W.A. & Songer, J.G. (1995) Toxic phospholipases D of Corynebacterium pseudotuberculosis, C. ulcerans and Arcanobacterium haemolyticum: cloning and sequence homology. Gene 156, 113–118. 4530 K. El Kirat et al. (Eur. J. Biochem. 270) Ó FEBS 2003 . transphosphati- dylation reaction in the presence of diacylglycerol. In order to detect bis(PA) production in biomimetic membranes incubated with S. chromofuscus PLD, we prepared vesi- cles, containing different molar. phosphatidylalcohol produced can be directly hydrolyzed by PLD. We also focused on the surface pressure dependency of PLD-catalyzed hydrolysis of phospholipids. These experiments provided new informa- tion. the hydrophobicity of the phosphatidylalcohol headgroup. Indeed, several physical studies with liposomes showed the tendency of such types of phospholipid headgroup to insert into the membranes, leading to