Characterization of Complexes Formed in Fully Hydrated Dispersions of Dipalmitoyl Derivatives of Phosphatidylcholine and Diacylglycerol

Transcription

1 1374 Biophysial Journal Volume 68 April Charaterization of Complexes Formed in Fully Hydrated Dispersions of Dipalmitoyl Derivatives of Phosphatidylholine and Diaylglyerol Peter J. Quinn,* Hiroshi Takahashi,:!: and Ihiro Hatta:!: 'Department of Biohemistry, King's College London, Campdem Hill, London W8 7AH, United Kingdom, and *Department of Applied Physis, Shool of Engineering, Nagoya University, Nagoya Japan ABSTRACT The phase diagram of fully hydrated binary mixtures of dipalmitoylphosphatidylholine (DPPC) with 1,2dipalmitoylglyerol (DPG) published reently by L6pez-Garia et al. identifies regions where stoihiometri omplexes of 1:1 and 1:2 DPPC:DPG, respetively, are formed. In this study, the strutural parameters of the 1:1 omplex in the presene of pure DPPC was haraterized by synhrotron low angle and stati x-ray diffration methods. Strutural hanges upon transitions through phase boundaries were orrelated with enthalpy hanges observed by differential sanning alorimetry in mixtures of DPPC with 5, 7.5, 10, and 20 mol% DPG dispersed in exess water. Phase separation of a omplex in gel phase ould be deteted by alorimetry in the mixture ontaining 5 mol% DPG but was not detetable by synhrotron low angle x-ray diffration. Stati x-ray measurements show evidene of phase separation, partiularly in the refletions indexing hain paking. In the mixture ontaining 7.5 mol% DPG, two distint lamellar repeat spaings ould be seen in the temperature range from 25 to 34 C. The lamellar spaing of about 6.6 nm was assigned to pure gel phase DPPC beause the hange in the spaing orresponds with thermal transition of the pure phospholipid, and a longer repeat spaing of about 7.2 nm was assigned to domains of the 1:1 omplex of DPPC-DPG. In the temperature range from 34 to 42 C, I.e., in the region of oexistene of the ripple phase of DPPC and the gel phase of the omplex, a single, rather broad lamellar refletion appears beause of superposition of two refletions of DPPC and the omplex; the lamellar spaing of DPPC in the ripple phase is similar to that of the gel phase of omplex. In the oexistene region of the liquid-rystalline phase of DPPC and the gel phase of omplex (42-48 C), the lamellar refletions of the both phases are present. The fluidus boundary lies between the oexistene region and the fluid region. In the fluid region ( C), the gel state of omplex persists up to the fluidus boundary, whereupon the liquid-rystalline state of omplex replaes the gel state of the omplex. This indiates that the omplex is also immisible with DPPC even above the fluidus boundary at least in the temperature range lose to the phase boundary. For mixtures omprising 10 and 20 mol% DPG in DPPC, omplex formation is learly detetable in both the gel region and the oexistene region by x-ray diffration. Synhrotron x-ray measurements indiate phase separation between pure DPPC and liquid-rystalline omplex just above the fluidus boundary. Stati, wide angle x-ray measurements also suggest phase separations of the 1:1 omplex not only from the gel phase but also the liquid-rystalline phase of pure DPPC. Two distint diffration peaks were deteted for the mixture of DPPC with 5, 10, and 20 mol% DPG. One is due to the hain spaing of the omplex, and the other is due to that of the pure DPPC. In the oexistene region of the Iiquid-erystalline phase of DPPC and the gel phase of omplex, two kinds of diffration peaks of the hydroarbon hain of the gel phase omplex and the broad sattering profile for the hain melting of DPPC were observed in the wide angle region. Eletron density reonstruted from the lamellar refletions indiates that the thiknesses of both the bilayer and the water layer of the gel phase omplex are greater than those of the respetive thiknesses of gel phase DPPC. INTRODUCTION An important mehanism of signal transdution in biologial membranes is the prodution of hydrolyti produts of phospholipase ation, inluding diaylglyerols generated by phospholipase C-type enzymes ating on membrane phospholipid substrates. The ation of these enzymes is known to be augmented by numerous hormones, neurotransmitters, and growth fators, whih results in a transient aumulation of diaylglyerol in the membrane (Kikkawa and Nishizuka, 1986; Berridge, 1987). The produts of enzymati ation are believed to play an important role in signal transdution mehanisms (Dennis et ai., 1991). Reeivedfor publiation 16 May 1994 and in final form 21 September Address reprint requests to Dr. Peter J. Quinn, Division of Life Sienes, King's College London, Campdem Hill, London W8 7AH, U.K. Tel.: ; Fax: ; udb6oo@uk.a.kl..bay. Abbreviations used: DPPC, 1,2-dipalmitoyl-sn-glyero-3-phosphoholine; DPG, 1,2-dipalmitoylgiyerol; DSC, differential sanning alorimetry by the Biophysial Soiety /95/04/1374/09 $2.00 An examination of the moleular speies of 1,2diaylglyerol produts suggests that phosphatidylholine is the major substrate for phospholipase C ation over relatively long time sales (Peleh and Vane, 1989), whereas phosphatidylinositol 4,5-bisphosphate is the major substrate yielding diaylglyerol in an aute response (Nishizuka, 1984, 1992). Several onsequenes have been desribed when diaylglyerols are present in relatively small amounts in phospholipid bilayers and membranes. These inlude ativation of protein kinase C (Bell, 1986), potentiation of the ation of various phospholipases on the their respetive phospholipid substrates (Dawson et ai., 1983, 1984), and destabilization of phospholipid bilayer membranes so as to failitate fusion (Siegel et ai., 1989). The moleular mehanisms that underlie these different proesses are presently unknown, but there is general agreement that the presene of relatively small amounts of diaylglyerols an influene the phase behavior of membrane lipid bilayers. Aordingly, mixtures of diaylglyerols and phospholipids have been subjet to numerous studies to define the way that diaylglyerols modify the phase behavior

2 Quinn et al. Charaterization of DPPC-DPG Complexes of membrane phospholipids and thereby to gain a learer understanding of the mehanism of their biologial funtions. For example, studies of mixed monomoleular films of 1-palmitoyl-2-0Ieoyl-phosphatidylholine and varying proportions of 1,2-dioleoylglyerol have shown that stable omplexes of the two omponents are formed at a molar fration of 0.25 diglyeride (Samby and Brokman, 1985). Further evidene of the formation of stable omplexes was obtained by alorimetri studies of these mixtures in aqueous dispersions (Cunningham et ai., 1989), whih onfirmed onlusions based on earlier thermal and fluoresent probe studies of mixtures of various diaylglyerols with phospholipids (Ortiz et ai., 1988). Further evidene of phase immisibility and separations were obtained from partial phase diagrams onstruted from x-ray diffration studies of mixtures of various diaylglyerols with phospholipids (Das and Rand, 1984, 1984). More reently, the formation of omplexes between dimyristoyl derivatives (Heimburg et ai., 1992) and dipalmitoyl derivatives (LOpez-Garia et ai., 1994) of diaylglyerol and phosphatidylholine (PC) have been identified from phase diagrams of fully hydrated binary mixtures examined in the temperature range from 20 to 80 C. The omplexes appear to onsist of approximately 1 and 2 mol of DG per 1 mol of PC. Thermal evidene indiates that the 1:1 omplex oexists with pure phospholipid in the gel phase in mixtures ontaining less than 50 mol% 1,2-dipalmitoylglyerol (DPG). The essential features of the phase diagram of the dipalmitoylphosphatidylholine (DPPC)-DPG mixture for the DPG onentration range studied in this work are as follows. For the DPPC-DPG mixture with less than 12.5 mol% DPG, the pretransition of DPPC was observed in differential sanning alorimetry (DSC) traes. At temperatures below the main transition of DPPC, two regions an be defined. A lower temperature region where there is oexistene between the gel phase of pure DPPC and the gel phase of omplex. In the higher temperature region, the ripple phase of DPPC oexists with the gel phase of omplex. Furthermore, the liquid-rystalline phase of DPPC oexists with the gel phase of omplex at temperatures above the main transition of pure DPPC. LOpez-Garia et ai. (1994) have defined the fluidus boundary where the omplex undergoes a transition to the fluid phase but have found no evidene for any immisibility in the lamellar phase in this region of the phase diagram. Strutural evidene for the oexistene of stoihiometri omplexes with pure phospholipid is presently laking, and no haraterization of the ompounds formed in these mixtures has yet been performed. In the present study, we have used synhrotron low angle and stati x-ray diffration to examine the region of phase oexistene of phospholipid with stoihiometri omplex omprising of 1:1 DPG/DPPC in a molar fration. The resolution of the method is suffiient to distinguish the oexistene of the gel phase of omplex from the gel phase of DPPC and the gel phase of omplex with the liquid-rystalline phase of DPPC. The oexistene of the liquid-rystalline phase of omplex with the liquidrystalline phase of DPPC is detetable by synhrotron, low angle x-ray diffration for the region near the fluidus line MATERIALS AND METHODS Sample preparation DPPC was purhased from Avanti Polar Lipids (Birmingham, AL), and DPG was obtained from Sigma Chemial Co. (St Louis, MO). No hromatographi analyses were performed, but narrow (<0.2 C) main transition of the phospholipid and melting transition temperatures of the DPG indiated that the purity of the lipids was omparable with that of lipids used in previous studies and onsistent with the supplier's laim of 99% authenti ompound. DPPC and DPG were dissolved in hloroform and mixed in appropriate proportions to ahieve the desired molar frations. The solvent was evaporated under a stream of oxygen-free dry nitrogen and stored for 16 h in vauum to remove remaining traes of solvent. These onditions were known to result in minimization of the weight of lipid. Multilamellar suspensions of the lipid mixtures were obtained for alorimetry by adding 14 ml of bidistilled deionized (ph 5.3, ondutivity < 2 ms) per g of lipid and 1 ml per g of lipid for samples used in x-ray examination. The lipid mixtures were hydrated at 70 C for at least 1 h and dispersed by soniation in an ultrasoni bath. The lipid dispersions were stored at -()...4 C until required for examination. The method of preparation and storage gave reproduible phase behavior when samples prepared at different times were examined by alorimetry or x-ray diffration. Calorimetry DSC was performed using a DSC10 with SSC580 (SEIKO I & E, Tokyo, Japan) and with san rates of 0.8 C/min. Lipid dispersions (15 ml) were plaed in sample pans (PIN P/023C, SEIKO) and hermetially sealed. Transition enthalpies were determined by integration of the area under the transition urve, and the position of the peak maximum was used to determine the transition temperature. X-ray diffration Stati x-ray diffration patterns were reorded using a rotating anode generator (Rigaku RU2oo, Tokyo, Japan) produing CuKa with a Ni filter (A = nm). The sample was sandwihed between polyimide windows to give a path length of 1 mm and held in a water-jaketed sample holder through whih water was irulated at designated temperatures using a Thermomix (Braun Bioteh) irulating water bath. The sample temperature was monitored by a thermoouple plaed in the sample, and the stability was within ±0.1 C. The x-ray beam was foused using a Franks optial system, and the sample-to-detetor distane was 370 mm.low angle x-ray diffration profiles were deteted by a position-sensitive proportional ounter (PSPC, Rigaku) with a length of 100 mm divided into 2815 hannels. Counts were aumulated over h and were unorreted for bakground sattering. X-ray data inluding both low angle and wide angle (S = nm- 1) were also reorded on a two-dimensional imaging plate (Type BA-III, Fuji Photo Film Co. LTD., Japan) mounted at a distane of 187 mm from the sample. Eah pattern was averaged by azimuthally integrating twodimensional Debye-Sherrer patterns (Takahashi et ai., 1991). In these measurements, the exposure times were typially h. The x-ray diffration data used for the reonstrution of eletron density profiles were aumulated over h, and the intensities ofthese refletions were determined by fitting the observed refletion peaks to Lorentzian line shapes aompanied with a linear funtion (y = ax + b). These data used for the above analysis were orreted by subtrating the bakground sattering, the sattering for the sample ell ontaining no lipid. Synhrotron x-ray diffration studies were undertaken at station 15A of the Photon Fatory (Amemiya et ai., 1983). Monohromati x-rays (A = nm) produed by a germanium (111) rystal with a Frankuhen ut were foused horizontally and ollimated with two pairs of slits to give a beam dimension of 2.0 x 1.0 mm (width x height). The photon flux of the inident beam was monitored ontinuously by an ionization hamber loated in front of the sample holder and was typially in the range of 2-4 x 108 photons/so The photon flux was relatively invariant during individual temperature sans. The photon flux inident on the sample was redued by an estimated 30-fold amount

3 1376 Biophysial Journal Volume 68 April1995 Phase separation between DPPC and a omplex of DPG and DPPC is illustrated in thermal behavior of 5 mol% DPG in DPPC mixed aqueous dispersion shown in Fig. 1A. The main and pretransitions were observed at 41.5 and 34.5 C (peak temperatures), respetively, for pure DPPC as shown in (a) of Fig. 1 A. The main transition is seen in both heating and ooling sans of the mixtures shown in (b) and () of Fig. 1 A, respetively, but the transitions in the latter appear to be less ooperative as judged by a slight broadening of the peaks. The peak temperatures of the main transition of the mixtures are 41.6 and 40.4 C, for heating and ooling san, respetively. This is also true of the pretransition, whih is also broader in the mixture and takes plae at a slightly lower temperature (heating, 33SC; ooling, 30 C). The anomalies in heating and ooling sans near 48 and 47 C for heating and ooling sans, respetively, are due to an order-disorder transition of hydroarbon hains of a stoihiometri omplex of DPG and DPPC. It is of interest that this transition peak in ooling sans is sharper than that in heating sans. The enthalpy of the main transition of DPPC is the same as the enthalpy of DPPC plus DPG-DPPC reorded in the mixture (31.2 kj/mol). No differenes ould be deteted in lamellar repeat spaings obtained from stati or synhrotron x-ray diffration patterns reorded under idential onditions as those used to reord the alorimetri behavior, with the exeption of the oexistene region of the liquid-rystalline phase ofdppc and the gel phase ofthe omplex. The plots, (a) and (b), of Fig. 1 B show stati diffration patterns of DPPC and a 5 mol% DPG in DPPC mixture reorded at 42.5 C, respetively. The lamellar repeat spaing is inreased from 6.70 to 7.09 nm in the presene of 5 mol% DPG. In this mixture, there was no evidene of a phase separation in either the stati or synhrotron x-ray difby the water jaket in whih the sample was immersed, tbus giving a flux inident of photons/s on the sample. Diffration patterns were reorded using a one-dimensional, positionsensitive proportional ounter with 512 hannels (PSPC, Rigaku) linked to the station omputer. The detetor was positioned with tbe enter of the beam aimed at the midpoint of the detetor so that two symmetri profiles of the powder pattern on either side of the beam stop were reorded onurrently, eah ontaining tbe first- and seond-order Bragg refletions of the phospholipidbilayers. The ative length ofthe PSPC was 170 mm (462 hannels). Thesample-to-detetordistanewas1629mm, and the maximum angle of x-ray sattering that ould be deteted was 3. The hannel-tohanneldimension ofthe PSPCwas mm, givingan angular resolution of 0.013, equivalent to a resolution of approximately 0.07 nm for spaings in the range -6-7 nm. The sample was housed in a holder desribed in detail elsewhere (Tenhov et ai., 1989). The temperature of the sample was monitored ontinuously by a hromel-alumel thermoouple mounted adjaent to the sample. Linear heating and ooling temperature sans of0.8 C/min between 2 5 and 90 C were produed by eletronialiy ontrolied flow of water through the sample holder. The sample stage was translated vertially after eah san so that the maximum time of exposure to the beam was 45 min. This time of exposure produed no detetable hange in the diffration pattern or thermal phase behavior judged by alorimetry. Up to 90 suessive diffration patterns were reorded, eah of 30-s exposure. The patterns were subjeted to a smoothing routine, and Bragg spaings were determined from the position of the diffration maximum. Ẹ, -g w 10.0 I (a) _=--.-/'-1\'-- (b) r (Cl----. (a).l!l E C> Temperature (OC) A B 70.0 EXPERIMENTAL RESULTS (f) (b)! J! S (nm-1) FIGURE 1 Thermal and strutural parameters of aqueous dispersions of DPPC and 5 mol% DPG in DPPC. (A) Differential sanning alorimetri urves of(a) DPPC during heating at 0.8 C/min, (b) beating, and () ooling thermogramsofa mixture ofdpg and DPPC at temperature rates ofo.8 C/ min. The bar indiates 10 kj/mol deg. (B) Stati x-ray diffration profiles obtained from aqueous dispersions of (a) DPPC and (b) 5 mol% DPG in DPPC reorded at 42SC. fration patterns reorded over the temperature range C. No evidene for a the ripple repeat spaing was observed in diffration patterns reorded in the mixed aqueous dispersion. Clear evidene ofimmisibility of a omplex of DPG and DPPC was observed in mixtures ontaining 7.5 mol% DPG or more in DPPC. Immisibility between the omplex of DPG-DPPC in the gel phase and DPPC in the gel phase an be seen, for example, in the x-ray sattering patterns reorded at temperatures up to about 34 C in the heating san of 7.5 mol% DPG in DPPC (Fig. 2). A first-order lamellar repeat spaing of the omplex entered at 7.27 nm is seen together with the lamellar gel phase ofpure DPPC, whih has a repeat spaing of 6.57 nm. The lamellar repeat of the gel phase of pure phospholipid inreases approximately to that of the gel phase of omplex when the ripple struture ( C) is formed, and the lamellar spaing of the two phases annot be distinguished. Heating through the main transition of the phospholipid, however, again shows the struture ofthe omplex with an unhanged lamellar repeat spaing together with the more prominent lamellar repeat of the liquid-rystalline

4 Quinn et al. Charaterization of DPPC-DPG Complexes 1377 'iii OJ E Cl. 1ii (f) " "E "". Cl.:: C _ nm6.17nm _-54.5 _ S (nm-l ) FIGURE 2 Plots of suessive x-ray sattering intensity profiles versus reiproal spaing as a funtion of temperature of a fully hydrated binary mixture of DPPC and 7.5 mol% DPG in DPPC heated from 25 C at a rate of O.BoC/min. Inset shows the sequene of diffration patterns obtained for the first-order diffration maxima during a subsequent ooling san at O.BoC/min. Eah diffration pattern represents x-ray sattering aumulated during 30 s. phase of DPPC entered at 6.64 nm. The refletions of the omplex around 7.2 nm diminish above 48 C, whih delineates the fluidus boundary. The sattering profiles at higher temperatures indiate a major lamellar liquid-rystalline phase with a seond lamellar phase with a slightly shorter spaing that replaes the phase-separated DPG-DPPC omplex in the gel phase at temperatures below the phase boundary. This transition we assign, on the basis of orresponding enthalpy hanges observed in alorimetri sans (LOpez Garda et ai., 1994), to a gel-liquid-rystalline transition of the DPG-DPPC omplex. The transition an be seen more learly in the subsequent ooling san (Fig. 2). A shoulder in the wide angle side at 6.17 nm learly replaes the lamellar refletion around 7.23 nm upon heating through the fluidus boundary. This result indiates immisibility of the liquidrystalline phase of omplex from the liquid-rystalline phase of pure phospholipid. Phase separations between DPPC and the omplex of DPPC-DPG are more learly evident in the mixed dispersions ontaining 10 mol% DPG. Dynami x-ray diffration data reorded during heating and ooling after heating san ofthis sample showed that at the ommenement ofthe initial heating san (35 C) there were two lamellar repeat spaings; one struture is presumed to orrespond to the omplex with hydroarbon hains in the gel onfiguration, and the other lamellar struture is assigned to the pure DPPC either in an expanded lamellar-gel phase or a modified ripple phase. A similar pattern was obtained from a stati x-ray diffration profile reorded from a sample under equilibrium onditions at 25 C. A transition of the latter phase ours at about 42 C to liquid-rystalline phase with a spaing of 6.89 nm at 43 C dereasing to 6.60 nm at 52 C. This spaing is slightly longer than that obtained from dispersions of pure DPPC. The peak assigned to the omplex in the gel phase remains throughout the transition region, but the intensity greatly dereases. A third lamellar omponent appears at temperatures just above the transition of the pure DPPC, and the lattie spaing of this omponent inreases gradually to that idential with the omplex. The spaings of the lamellar liquidrystalline phase of pure DPPC and the gel phase of the omplex oexist up to the fluidus boundary, where the firstorder lamellar diffration peak for the omplex at 7.4 om disappears. The question of immisibility between pure DPPC and the liquid-rystalline phase of the DPPC-DPG omplex is problemati. Representative diffration patterns in the transition through the fluidus boundary are shown in Fig. 3. At temperatures above the fluidus line, the shape of the diffration bands are asymmetri, whih is partiularly evident in the seond-order refletion, with a shoulder on the wide angle side of the lamellar repeat spaings of the liquidrystalline phase of the pure phospholipid. It is possible that phase separations between phospholipid and omplex of DPPC-DPG persist at least in regions lose to the fluidus boundary. Changes in diffration patterns during heating and ooling a dispersion ontaining 20 mol% DPG in DPPC are illustrated in Fig. 4, A and B. The heating san shows persistene of the dominant x-ray sattering peak entered at 7.29 nm during transitions between gel and liquid-rystalline phases in the pure phospholipid, again providing strong evidene for phase separation of omplex in the gel phase. Heating the mixed dispersions to temperatures above the fluidus line auses broadening of the peaks. A stati diffration pattern of the 20 mol% DPG in DPPC mixture reorded at 70 C is ompared with pure DPPC in Fig. 4 C. A single lamellar spaing is observed in both ases. The only distinguishing feature between the two diffration patterns is an inrease in lamellar repeat spaing from 6.3 to 7.1 nm due to the presene of the DPG. Thus, there is no evidene of phase immisibility between the omplex and pure phospholipid at 70 C from the low angle stati diffration measurements. The gel phase immisibility between DPPC and the omplex of DPPC-DPG was also observed learly in the profiles 7.22 nm 6.64 om S (nm-l ) FIGURE 3 Plots of suessive x-ray sattering intensity profiles versus reiproal spaing as a funtion of temperature of a fully hydrated binary mixture of DPPC and 10 mol% DPG during ooling from 60 to 48 C at O.BoC/min through the fluidus line. Eah diffration pattern represents x-ray sattering intensity aumulated during 30 s.

5 1378 Biophysial Journal Volume 68 April mol% S (nrn-1) 20mol% >. -'iii -: 10mol% Ol : 'C -o () Cf) 5mol% Omol% 0.1 (a) 'iii : -: Ol : 'C -a () Cf) S (nrn-1) S (nm'1) FIGURE 5 Stati wide angle x-ray diffration patterns of pure DPPC and mixtures of DPPC with DPG at proportions of 5, 10, 20, and 50 rnol% reorded at 20 C. 5, 10, and 20 mol% DPG appear to be superpositions of the peaks of pure DPPC and the 1:1 omplex. The inrease in the intensity of a band entered at nm with inreasing proportion of DPG in the mixture is onsistent with the formation of 1:1 omplex that is phase separated from pure DPPC. Likewise, in the oexistene region of the liquidrystalline phase ofdppc the presene ofgel phase omplex an be observed from sattering peaks in the wide angle region. Thus, in Fig. 6 it an be seen that a broad diffration FIGURE 4 Plots of suessive x-ray sattering intensity profiles versus reiproal spaing reorded during heating (A) and ooling (B) an aqueous dispersion ofdppc ontaining 20 mol% DPG. Eah diffration pattern represents x-ray satteringaumulated during 30s. (C) Statix-ray diffration profiles of (a) pure DPPC and (b) 20 mol% DPG in DPPC reorded at 70 C. of the wide angle sattering region. Wide angle diffration provides information on the arrangement of the hydroarbon hains of lipid. Fig. 5 shows wide angle diffration profiles at 20 C for pure DPPC and the mixture of DPPC with 5, 10, and 20 mol% DPG and the 1:1 omplex of DPPC-DPG. The profile for pure DPPC onsists of a sharp peak at nm, and a broad shoulder on the higher angle side implies that the hydroarbon hains are paked in a quasi-hexagonal lattie and are tilted with respet to the bilayer plane (Tardieu et ai., 1973). On the other hand, a sharp symmetrial peak at nm for the 1:1 omplex of DPPC-DPG indiates that the hains are paked in the hexagonal lattie and are normal to the bilayer plane. The profiles for the mixture of DPPC with 'iii : E Ol : 'C 1il () Cf).2 E 'C ool o -.J mol% 5mol% S (nm- 1 ) FIGURE 6 Stati wide angle x-ray diffration profiles of mixtures of DPPC with 5, 10, and 20 mol% DPG reorded at 43, 45, and 50 C, respetively. These measurement temperatures are midway between the solidus and fluidus phase boundaries in the DPPC-DPG phase diagram (L6pez-Garia et ai., 1994). 3.0

6 Quinn et al. Charaterization of DPPC-DPG Complexes 1379 band from the hains of pure DPPC in the liquid-rystalline phase oexists with a sharp symmetrial peak entered at a spaing of about nm due to the gel phase omplex. At temperatures above the fluidus boundary, this sharp peak disappears and only a single diffuse wide angle sattering band is observed in all mixtures (data not shown). From the results, no onlusion an be drawn about the misibility of DPPC and DPG from the wide angle data at temperatures above the fluidus boundary. To estimate the thikness of the bilayer and the water layer, the eletron density profile of the gel phase 1:1 omplex of DPPC-DPG was alulated and was ompared with the eletron density profile of pure DPPC in the gel phase. The thikness of the water layer is established by the balane of attrative and repulsive fores between the neighboring surfaes of the bilayers, Le., the thikness of the water layer is refletion ofinterbilayer interations. The x-ray diffration patterns for the 1:1 omplex and pure DPPC at 20 C are shown in Fig. 7 A. The DPPC-DPG omplex gives rise to five orders of the lamellar refletion. On the other hand, the higher order lamellar refletions up to the eighth order were observed for pure DPPC; however, the intensities ofseventhand eighth-order refletions were very weak, and the sixth refletion was not deteted. Fig. 7 B shows the onedimensional eletron density profiles for pure DPPC and the 1:1 omplex of DPPC-DPG in the gel phase. To obtain omparable resolution for the two samples, the five peak intensities were used to alulate the Fourier reonstrution ofthe eletron density profiles. These profiles were alulated using the phase ombination -,-,+,-,-. This phase ombination has been reported onsistently for the DPPC in the gel phase using the water swelling method (Torbet and Wilkins, 1976; Franks and Lieb, 1979; MIntosh and Simon, 1986). The same phase ombination has been applied in the Fourier reonstrution of the 1:1 omplex. The alulated eletron density profile is onsistent with that of a hydrated lipid bilayer (see Appendix). The eletron density profiles show that the thikness of the bilayer and the water layer of the 1:1 omplex are greater than those obtained for pure DPPC. From the eletron density profiles, we estimate the thikness of the bilayer (Le., the distane of between the headgroups) ofpure DPPC and the 1:1 omplex are 4.11 and 4.54 nm, respetively. The differene ofthe bilayer thikness is 0.43 nm. This value may be interpreted by onsidering the tilt angle of hydroarbon hain. The length of hydroarbon hain of DPPC in the gel phase was estimated by Nagle and Wiener (1988) to be 2.03 nm. Ifthe hains ofthe 1:1 omplex are not tilted, the tilt angle ofthe hains ofdppc is alulated to be about 27 (os- 1 «2 X )/2 X 2.03) = 26.6 ). This value is in agreement with reported values of hain tilt (Tardieu et ai., 1973; Tristram-Nagle et ai., 1993). The differene ofthe thikness of the water layer would indiate the differene of the interbilayer interation between pure lipid and the omplex. It may not be unexpeted that inserting DPG between moleules ofdppc alters the balane offores between adjaent polar headgroups in the surfae layer thereby altering the interbilayer interating fores.... >. '(i) C... <l> C 0> C.;:: <l> () C/) -6 DISCUSSION DPPC-DPG 1:1 DPPC DPPC-DPG 1 : X (nm) 4 x10 x S (nm ) FIGURE 7 (A) Stati x-ray diffration profiles obtained from aqueous dispersions of DPPC and the 1:1 omplex ofdppc-dpg reorded at ZOoC. For S > 0.3 om-i, the plots with a 10-fold expanded sale are also shown. (B) One-dimensional eletron density profiles onstruted from the x-ray diffration patterns for the gel phases of pure DPPC and the 1:1 omplex of DPPC-DPG. The phase diagram for the fully hydrated DPPC-DPG mixtures has already reported by LOpez-Garda et a1. (1994). The formation of omplexes has been suggested from the phase diagram. In mixtures ontaining less than 50 mol% DPG, the omplex omposition is 1:1 (mol:mol) for DPPC:DPG. In temperature regions below the main transition of DPPC, immisibility between pure DPPC and stoihiometri B A 6

7 1380 Biophysial Journal omplex of DPPC-DPG has been observed using the freeze-frature eletron mirosopy (Lopez-Garia et ai., 1994). Cunningham et ai. (1989) and Ortiz et ai. (1988) have also pointed out the immisibility of omplex and PC for diaylglyerol-pc mixtures from the results of DSC. In the temperature region below the main transition of pure DPPC, two distint sets of lamellar refletions are present in the low angle sattering region for the mixtures. The oexistene of two diffration peaks was also observed in the wide angle region for the mixtures. These results onfirm that DPPC in either the gel or ripple phase is immisible with the gel phase of DPPC-DPG omplex. Furthermore, two different sets of lamellar spaings arising from pure DPPC and DPPC-DPG omplex were also observed in the temperature region above the main transition of pure DPPC. The diffration peak in the wide angle region orresponding to the paking of the hydroarbon hains of the omplex appears together with the board-sattering profile, indiating the hain-melting of pure DPPC. Therefore, the DPPC-DPG omplex is immisible with the DPPC bilayer in both the gel and the liquid-rystalline phases. The most likely explanation of why the phase separation ould not be deteted for the mixture with 5 mol% DPG in DPPC is that the amount of the omplex present is not suffiient to produe a detetable diffration peak. In the previous work of LOpez-Garia et ai. (1994), the pretransition of DPPC was deteted learly in thermograms of mixtures ontaining up to 12.5 mol% DPG. In the temperature region between the pretransition and the main transition, we observed a large lamellar spaing that is harateristi of the ripple phase (see, for example, the diffration patterns from a mixture ontaining 7.5 mol% DPG (Fig. 2». However, we ould not detet the refletion orresponding to the ripple repeat periodiity. This might be beause the lateral domain size of DPPC in the mixtures is not large enough to give detetable sattering intensity and also that the ripple struture is not well ordered. The pure DPPC domain might also inlude small amounts of DPG that perturb the formation of a oherent ripple struture. There is evidene, however, that the harateristially large lamellar repeat of the ripple phase is stabilized by DPG. This is seen in x-ray diffration patterns as a differene in heating and ooling sans for the mixture with 20 mol% DPG. The lamellar refletion has a shoulder on the wide angle side of the diffration peak around 30 C in heating san. By ontrast, there is only a single symmetrial peak orresponding to the lamellar repeat spaing in the ooling san (f. Fig. 4, A and B). As seen in the thermal data (Fig. 1 A), the pretransition has a hysteresis in the ooling san. The pretransition in ooling mode ours at a lower temperature than heating san; therefore, the DPPC domain in the mixture of 20 mol% DPG still remains in the ripple phase even at 30 C. Aordingly, only a single peak ould be observed in the ooling san at low temperature. The lamellar repeat spaing of the omplex is longer than that of pure DPPC in the gel phase. A similar finding has Volume 68 April 1995 already been reported for phospholipid-diaylglyerol mixtures (Das and Rand, 1986; LOpez-Garia et ai., 1994). The lamellar repeat spaing is the sum of the thikness of the bilayer and the water layer. The Fourier reonstrutions of the eletron density profiles show that both of the bilayer thikness and thikness of water layer inrease. This result implies that DPG moleules affet the paking and the tilt of hydroarbon hains of DPPC in the omplex. In this ase, the presene of gel phase DPG alters the tilt angle of the hains of DPPC with respet to the bilayer normal (Tardieu et ai., 1973; Tristram-Nagle et ai., 1993). Suh an effet has also been observed in binary mixtures of DPPC and palmiti aid (Koynova et ai., 1988). One of the most signifiant results is that from the x-ray diffration patterns of samples ontaining of 7.5 mol% or more of DPG where there is diret evidene of immisibility between pure phospholipid in the liquid-rystalline phase and the stoihiometri omplex. Immisibility above the fluidus boundary remains unresolved in mixtures ontaining DPG, but immisibility has been reported for binary mixtures of dielaidoylphosphatidylholine and dipalmitoylphosphatidylethanolamine above the fluidus boundary using spin-label eletron spin resonane spetrosopy (Wu and MConnell, 1975). The spin-label method, however, is sensitive to the presene of relatively small lateral domains within the bilayer. low angle x-ray diffration, however, demands orrelation between domains in suessive bilayers. Judging from two different lamellar spaings, the domains of the DPPC-rih region and the DPPC-DPG omplex must separate three-dimensionally as a onsequene of lateral twodimensional phase separation brought about by intermoleular interations between DPPC and DPG. This suggests that the free energy assoiated with the interation between bilayers in the ase of the surfae of the omplex juxtaposed with a surfae of DPPC bilayer is high ompared with that existing between the same neighboring surfaes. In fat, the eletron density profiles (Fig. 7 B) show that the water layer, whih reflets the interation between bilayers, is different between pure DPPC and the omplex. The formation of omplexes and domain strutures of DPPC and DPG, whih are detetable even at low molar ratios of DPG, may underlie the biologial funtion of diaylglyerol in ell membranes. Thus, the reation of domains rih in diaylglyerol in the fluid bilayer matrix ould give rise to strutural features responsible for ativation of various enzymes related by signal transdution mehanisms in biomembranes (Dennis et ai., 1991; Dawson et ai., 1984). The authors are grateful to Dr. Y. Amemiya for helpful advie in setting up the instrumentation for synhrotron x-ray diffration experiments. The authors thank the Computation Center of Nagoya University and the Photon Fatory of National Laboratory for High Energy Physis. J. C. Gomez-Fernandez generously provided haraterized lipid for use in the study. This work was aided from funds provided by the Siene and Engineering Researh Counil (U.K.). P. J. Quinn was supported by a British Counil/Monbu-sho (Ministry of Eduation, Siene and Culture, Japan) Visiting Professorship.

8 Charaterization of DPPC-DPG Complexes Quinn et al. APPENDIX The phase determination required to alulate the eletron density profile of the DPPC-DPG omplex (1:1) at 20 C was done in two ways. The bilayer arrangement is onsidered to be a entrosymmetri struture. In this ase, the values of the strutural fators are real numbers, Le., the phase angles are limited to 0 or 'Fr. For the DPPC-DPG omplex, five lamellar refletions are observed; therefore, the variety of the phase ombinations is expeted to be 32. All possibilities of the phase ombination were examined. Only one set, (-, -, +, -, -, i.e., 'Fr, 'Fr, 0, 'Fr, 'Fr), gave an eletron density profile that was not inonsistent with a hydrated lipid bilayer. All other phase sets gave anomalous results, suh as large positive eletron density peaks in the enter of the bilayer. In a seond approah, we interpreted the intensities of lamellar refletion using a strip model. Strip models have been used to analyze the struture of lipid bilayers (Worthington, 1969; Furuya et ai., 1976; Wiener et ai., 1989). The strip model represents the eletron density of bilayer strutures as a sequene of onstant eletron density. We onsidered a four-strip model for the eletron density of the unit ell of multilamellar bilayer. The bilayer is divided into by four parts with onstant eletron density, namely, terminal methyl, methylene, headgroup, and interlamellar water (Fig. 8 A). In the strip model, the struture fator an be alulated analytially (Worthington, 1969). A four-strip model has seven parameters, Le., the value of eletron density of eah of four parts and three values of boundary position of eah "E 400 w 350 :! e 250 [jj methylene "- terminal methyl N E i:i:" 400 iji aj q; (1) where, FC<h) is the struture fator alulated from the four-strip model and I F m( h) I is the absolute value of the measured normalized struture fator. This value is the square root of the integrated intensity and is normalized as I F m(1) I = h is the diffration order. Beause a one-dimensional system is onsidered in this study, the unit of the struture fator is e/nm 2 I is sum of h. K is a sale fator. The sale fator K is also one of the parameters. Nonlinear least-squares fitting was performed to establish the minimum of the residue R. The optimal value obtained for the eletron density of the headgroup was 445 e/nm3 The positions of the methyleneheadgroup boundary and the headgroup-water boundary were 1.72 and 2.96 nm, respetively. Fig. 8 B shows a omparison of measured and alulated struture fators. The agreement of both fators indiates that the four-strip model is a reasonable model of the 1:1 omplex of DPPC-DPG. The resulting phase set derived from this analysis of the omplex (-, -, +, -, -) is the same as that obtained from examination of all possible phase sets. REFERENCES 1000 strip. The eletron density of the interlamellar water is assigned a value of 335 e/nm3 The eletron density of the methylene part was estimated by same method used by Wiener et al. (1989). The projeted separation of methylenes along an all-trans hain is nm. The ross setional area of the hains an be alulated from the Bragg spaing of the wide angle diffration maxima. Furthermore, judging from the shape of the wide angle diffration, the hydroarbon hains of the 1:1 omplex of DPPC-DPG are oriented normal to the bilayer plane. From these parameters, the volume of eah methylene residue of the 1:1 omplex is estimated to be 2.48 X 10-2 nm 3. Beause methylene has eight eletrons, the eletron density of methylene is estimated to be 323 e/nm3 It is assumed that the position of the boundary between terminal methyl and methylene and the eletron density of terminal methyl part are the same for DPPC and the DPPC-DPG omplex. These values were estimated to be 0.47 nm and 210 e/nm3, respetively, for DPPC at 20 C from the analysis of the four-strip model using our low angle intensity data. The origin is the enter of the bilayer. Similar values have been estimated for parameters (see Fig. 6 in Wiener et ai., 1989). The optimal values of three remaining parameters (the value of eletron density of headgroup, and the positions of the methylene-headgroup boundary and the headgroup-water boundary) were determined by searhing the minimum of the residue R defined by 4 X (nm) h FIGURE 8 (A) Eletron density profiles and the four-strip model of the 1:1 omplex of DPPC-DPG at 20 C. The thin solid line shows the best fit of the four-strip model of the 1:1 omplex. The thik solid line indiates the eletron density profile onstruted from the measured struture fators multiplied by the sale fator K at best fit. The phase set ( -, -, +, -, -) and value of F(O) were derived from the analysis of the four-strip model for the 1:1 omplex of DPPC-DPG at 20 C. The dashed line shows the model eletron density profile with same resolution, alulated from the four-strip model. (B) Comparison of the alulated struture fator F(h) from the four-strip model (--) and the measured struture fator Fm(h) (e). For both fators, the absolute values are shown. The phases assigned to the different regions are indiated by + and - on the figure. Arnemiya, Y., K. Wakabayashi, T. Hamanaka, T. Wakabayashi, T. Matsushita, and H. Hashizume Design of small-angle x-ray diffratometer using synhrotron radiation at the Photon Fatory. Nul. Instrum. Methods. 208: Berridge, M. J Inositol-triphosphate and diaylglyerol: two interating seond messengers. Annu. Rev. Biohem. 56: Bell, R. M Protein kinase C ativation by diaylglyerol seond messengers. Cell. 45: Cunningham, B. A., T. Tsujita, and H. L. Brokman Enzymati and physial haraterization of diaylglyerol-phosphatidylholine interations in bilayers and monolayers. Biohemistry. 28: Das, S., and R. P. Rand Diaylglyerol auses major strutural transitions in phospholipid bilayer membranes. Biohem. Biophys. Res. Commun. 124: Das, S., and R. P. Rand Modifiation by diaylglyerol of the struture and interation of various phospholipid bilayer membranes. Biohemistry. 25: Dawson, R. M. C., N. L. Hemington, and R. F. Irvine Diaylglyerol potentiates phospholipase attak upon phospholipid bilayers: possible onnetion with ell stimulation. Biohem. Biophys. Res. Commun. 117: Dawson, R. M.., R. F. Irvine, J. Bray, and P. J. Quinn Long-hain diaylglyerols ause perturbation in the struture of phospholipid bilayers rendering them more suseptible to phospholipase attak. Biohem. Biophys. Res. Commun. 125: Dennis, E. A., S. G. Rhee, M. M. BiIlah, and Y. A. Hannun Role of phospholipase in generating lipid seond messengers in signal transdution. FASEB J. 5:

9 1382 Biophysial Journal Volume 68 April 1995 Franks, N. P., and W. R. Lieb The struture of lipid bilayers and the effets of general anaesthetis: an x-ray and neutron diffration study. J. Mol. BioI. 133: Furuya, K., T. Yamaguhi, Y. Inoko, and T. Mitsui Strutures of uranyl-deorated leithin and leithin-holesterol bilayers. Ata Cryst. B32: Heimburg, T., U. Wiirz, and D. Marsh Binary phase diagram of hydrated dimyristoylglyerol-dimyristoylphosphatidylholine mixtures. Biophys. J. 63: Kikkawa, U., Y. Nishizuka The role of protein kinase C in transmembrane signalling. Annu. Rev. Cell BioI. 2: Koynova, R. D., B. G. Tenhov, P. J. Quinn, and P. Laggner, Struture and phase behavior of hydrated mixtures of L-dipalmitoylphosphatidylholine and palmiti aid. Correlations between strutural rearrangements, speifi volume hanges and endothermi events. Chem. Phys. Lipids. 48: LOpez-Garia, F., 1. Villalafn,1. C. GOmez-Fernandez, and P. 1. Quinn The phase behavior of mixed aqueous dispersions ofdipalmitoyl derivatives of phosphatidylholine and diaylglyerol. Biophys. J. 66: Nagle, J. F., and M. C. Wiener Struture of fully hydrated bilayer dispersion. Biohim. Biophys. Ata. 942:1-10. Nishizuka, Y The role of protein kinase C in ell surfae signal transdution in tumour promotion. Nature. 308: Nishizuka, Y Intraellular signaling by hydrolysis of phospholipids and ativation of protein kinase C. Siene. 258:607--{i13. Ortiz, A., J. Villalain, and J. C. Gomez-Fernandez Interation of diaylglyerols with phosphatidylholine vesiles studied by differential sanning alorimetry and fluoresene probe depolarization. Biohemistry. 27: Peleh, S., and D. E. Vane Signal transdution via phosphatidylholine yles. Trans. Biohem. So. 14: Samby, J. M., and H. L. Brokman Misibility, hain paking, and hydration of I-palmitoyl-2-oleoyl phosphatidylholine and other lipids in surfae phases. Biophys. J. 48: Siegel, D., J. Banshbah, A. Alford, H. Ellens, L. J. Lis, P. J. Quinn, P. L. Yeagle, and J. Bentz Physiologial levels of diaylglyerols in phospholipid membranes indue membrane fusion and stabilize inverted phases. Biohemistry. 28: Takahashi, H., S. Matuoka, S. Kato, K. Ohki, and I. Halla Eletrostati interation of poly(l-lysine) with dipalmitoylphosphatidi aid studied by x-ray diffration. Biohim. Biophys. Ata. 1069: Tardieu, A., V. Luzzati, and F. C. Reman Struture and polymorphism of hydrated hains of lipids: a study of leithin-water phases. J. Mol. BioI. 75: Tenhov, B. G., H. Yao, and I. Halla Time-resolved x-ray diffration and alorimetri studies at low san rates. I. Fully hydrated dipalmitoylphosphatidylholine (DPPC) and DPPC/water/ethanol phases. Biophys. J. 56: Torbet, J., and M. H. F. Wilkins X-ray diffration studies of leithin bilayers. J. Theor. BioI. 62: Tristram-Nagle, S., R. Zhang, R. M. Suter, C. R. Worthington, W.-J. Sun, and J. F. Nagle Measurement of hain tilt angle in fully hydrated bilayers of gel phase leithins. Biophys. J. 64: Wiener, M.., R. M. Suter, and J. F. Nagle Struture of the fully hydrated gel phase of dipalmitoylphosphatidylholine. Biophys. J. 55: Worthington, C. R The interpretation of low-angle x-ray data from planar and onentri multilayered strutures: the use ofone-dimensional eletron density strip models. Biophys. J. 9: Wu, S. H.-W., and H. M. MConnell Phase separations in phospholipid membranes. Biohemistry. 14: