llc‘f‘“ 4 tit-'2‘. {‘I',‘ .‘l SYNTHESIS OF LIPIDS CONTAINING PHYTANIC ACYL SUBSTITUENTS AND PHYSICOCHEMICAL CHARACTERIZATION OF I,2-DIPHYTANOYL-§anLYCERO-3-PHOSPHOCHOLINE IN HYDRATEO SYSTEMS By HILDEGARDE LINDSEY A DISSERTATION Submitted to Michigan State University in partial fulfiilment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1978 ABSTRACT SYNTHESIS OF LIPIDS CONTAINING PHYTANIC ACYL SUBSTITUENTS AND PHYSICOCHEMICAL CHARACTERIZATION OF I,2-DIPHYTANOYL-§Q:GLYCERO-3-PHOSPHOCHOLINE IN HYDRATED SYSTEMS By Hildegarde Lindsey Diphytanoyl phosphatidylcholine (DOPC) was synthesized following a published procedure. This synthesis also served as a pattern for the preparation of two other phospholipids and a glycolipid. A series of studies were undertaken aimed at characterization of the structural and dynamic properties of the synthetic lipid, diphytanoyl phosphatidylcholine, in multilamellar dispersions and vesicle suspensions. This lipid exhibited no detectable gel to liquid-crystalline phase transition over the temperature range studies (-lZOOC to +lZOOC). The effect of diphytanoyl phosphatidylcholine on the gel to liquid-crystalline phase transition of dipalmitoyl phosphatidyl- choline in mixtures of the two lipids resembled qualitatively that produced by the addition of up to 30 mole % of cholesterol to phosphatidylcholines. Hildegarde Lindsey Examination of proton nuclear magnetic resonance (NMR) free induction decay obtained from multilayer dispersions of diphytanoyl phosphatidylcholine provided an estimate of the methylene proton order parameter. The estimated value of -O.2l is comparable to those determined for other phospholipids. Sonication of aqueous dispersions of diphytanoyl phosphatidyl- phosphatidylcholine led to formation of bilayer vesicles as determined by the measurement of the ratio of the outer/inner choline methyl proton resonances, vesicle sizes in electron micrographs and analytical ultracentrifuge, and comparison of proton NMR linewidths between multilayer and sonicated dispersions. Ultracentrifugation of DOPC vesicles in H20 and 2H20 media yielded a value of l.Ol3 i 0.026 ml/g for the partial specific volume of this lipid. Spin-lattice relaxation rates were measured for the methyl and methylene-methyne protons of the hydrocarbon chains of diphytanoyl phosphatidylcholine in bilayer vesicles over a range of temperatures and at two nuclear magnetic resonance frequencies (lOO and 220 MHz). The observed relaxation rates for the methylene-methyne protons in this system were approximately twice those previously reported for dipalmitoyl phosphatidylcholine at comparable temperatures and resonance frequencies, whereas the relaxation rates measured for the methyl protons were greater than those of the straight chain lipid by roughly an order of magnitude. Measurement of the spin-lattice relaxation rates of the hydrocarbon protons of the diphytanoyl phosphatidylcholine in a Hildegarde Lindsey system of lO mole % of the branched chain lipid in a deuterated host lipid, diperdeuteropalmitoyl phosphatidylcholine, produced the following results: In the region of the gel to liquid-crystalline phase transition temperature of the deuterated dipalmitoyl phos- phatidylcholine host lipid, a discontinuity in the temperature dependence of the proton NMR longitudinal relaxation rates of the branched chain lipid was observed. This result may be taken as evidence of laternal phase separation of a liquid-crystalline phase enriched in diphytanoyl phosphatidylcholine from a gel phase enriched in diperdeuteropalmitoyl phosphatidycholine at temperatures below the phase transition temperature of the deuterated host lipid. This conclusion is supported by the observation of an abrupt change in the hydrocarbon methylene linewidth (at 100 MHz) of 10 mole % diphytanoyl phosphatidylcholine in diperdeuteropalmitoyl phosphalidyl— choline over the temperature range where lateral phase separation is taking place according to the differential thermograms. Permeation studies involving the rate of escape of glucose from diphytanoyl and dipalmitoyl phosphatidylcholine vesicles were initiated, but no definitive conclusions could be drawn from the limited data. Preliminary reconstitution studies of cytochrome g_oxidase in diphytanoyl phosphatidylcholine indicate that this lipid approxi- mates natural source lipids in its ability to restore enzymatic activity to dilipidated membrane proteins. ACKNOWLEDGEMENT I should like to thank Dr. Donald G. Farnum for possessing sufficient intellectual curiosity and flexibility to take me on as a student and for providing moral support and encouragement which helped sustain me over many difficult times. I am deeply grateful for the opportunity to join Dr. Sunney I. Chan's group at Caltech and for the impetus which this ex- perience provided to my growth as a person and a scientist. My thanks are extended to Dr. Nils O. Petersen and Dr. Valerie Hu for help on the work described herein and for their gracious permission to reproduce material included in their theses in mine. To Caltech, where salus populi suprema lex est, I offer to all within your walls gratitude for the gentleness, courtesy, tolerance, kindness, intellectual stimulation, supportive atmosphere, and justice which I experienced during my tenure there. I would like to acknowledge the financial support provided by Michigan State University, the Lucille K. Millar Scholarship, the Research Corporation, and Dr. S. 1. Chan, without which I could not have completed my work. ii TABLE OF CONTENTS Page INTRODUCTION .......................... 1 Statement of Problem ..................... 1 PART I ............................. 12 Synthesis of Lipids Containing Phytanic Acyl Substituents - . 12 EXPERIMENTAL I ......................... 24 l,2-Diphytanoyl-§n:glycero-3-phosphocholine (DDPC) ----- 24 l,2-Isoprophylidene—§Q:glycero-3-benzylcarbonate ,,,,,, 24 3(R,S),7(R),ll(R),lS-Tetramethylhexadecanoic acid (phytanic acid) ..................... 25 l,2-Diphytanoyl~§grglycerol ................ 25 2(S)-benzyl-N,N,N',N'-tetrabenzyl-2,6-diaminohexanoate, , . 27 2(S)-N,N,N',N'-tetrabenzyl-2,6-diaminohexanol . . ..... 23 l,Z-Diphytanoyl-§n:glycero-3-phosphodichloride ,,,,,,, 28 l,2-Diphytanoyl-§g:glycero-3-phospho-[2(S)-N,N,N',N'- tetrabenzyl-2,6-diaminohexanol] (DDPL) .......... 29 l,Z-diphytanoyl-§g¢glycero—3~phosphoglucosetetra- acetate . ......... ’. .‘. . ........... 30 B-D-Glucopyranosyl-(l,2-diphytanoyl-§g:glycerol)- 2,3,4,6-tetraacetate .................. 33 PART 11 ................ , ............ 35 Physicochemical Characterization of l,2-Diphytanoyl-§g: glycero-B-phosphocholine in Hydrated Systems ....... 35 Characterization of Unsonicated Aqueous Suspensions of DDPC .......................... 37 Establishment of phase formed by fully hydrated DDPC . . 37 Determination of gel to liquid-crystalline phase transition temperature for DDPC ............ 33 iii Page Examination of properties of lipid mixtures containing D¢PC .......................... 39 Determination of the order parameter for the hydrocarbon chain of D¢PC ..................... 46 Characterization of Sonicated Suspensions of D¢PC ,,,,, 54 Confirmation of presence of and determination of average size of D¢PC bilayer vesicles --------- 54 Determination of the heterogeneity of DDPC vesicle samples ........................ 59 Determination of partial specific volume of DOPC in vesicles ........................ 64 Examination of pmr spectral characteristics of DDPC in various media .................... 69 Verification of the average size of DDPC vesicles . . . . 3] Investigation of integrity of DDPC vesicles as measured by ion permeation ,,,,,,,,,,,,,, 86 Interpretation of spin-lattice relaxation rates of DOPC in bilayer vesicles in terms of kink diffusion and chain reorientation rates ............. 88 Detection of lateral phase separation in vesicles prepared from lipid mixtures containing DDPC ,,,,, 105 Characterization of Permeability of DDPC Vesicles ,,,,, 109 EXPERIMENTAL II ........................ ll5 Materials .......................... llS Sample Preparation ..................... ll5 Instrumentation and Methods ................. 117 PART III ........................... l20 Reconstitution of Cytochrome g_Oxidase in l,2—Diphytanoyl- .sn-glycero-3-phosphocholine ................ l20 EXPERIMENTAL III ....................... 124 Materials .......................... l24 Sample Preparation ..................... l24 Instrumentation and Methods ................. 125 iv Page SUMMARY AND CONCLUSIONS .................... 127 BIBLIOGRAPHY ......................... 132 APPENDIX ........................... 135 Unsuccessful Synthetic and Purification Steps ........ l36 LIST OF TABLES Table Page I Average Size of Vesicles Produced by Continuous Sonication ....................... 60 II Apparent Sedimentation Rates of Bilayer Vesicles ..... 68 III Activity of Cytochrome c Oxidase Reconstituted with Various Lipids [Lipid/Protein = 20/1 w/w] ........ 122 yi LIST OF FIGURES Figure Page 1 Photographs of space-filling models of DDPC showing the primary (left-hand) phytanic acyl chain in various conformations: (a) all-staggered conformation; (b) conformation containing one gauche chain-chain inter- action; (c) Kinked conformation. The insert (lower- right) shows schematically the effect of the intro- duction of a kink on the length of the chain, the chain cross-sectional area, and the disposition of the carbon backbone of the chain .................. 5 2 Potential energy as a function of rotation angle about the carbon-carbon bond illustrating: (a) phytanic acyl moiet , rotation of 1 120° involving a methyl branch; (b) palmitic acyl moiety, any internal rota- tion of ilZO° ...................... 7 3 Synthetic scheme for l,Z-diphytanoyl-§g;glycero-3- phosphocholine. {No stereochemistry of the glycerol is implied in the structures in Figures 3 through 5.}. . l4 4 Synthetic scheme for 1,2-diphytanoyl-sn-glycero-3- phospho-[2(S)-N,N,N',N',-tetrabenzyl42:6-diamino- hexanol] ........................ 18 5 a) Synthetic scheme for l,2-diphytanoyl-3-B-D- gl ucosyl -6-sulfonic acid-sn-glycerol . ........ 20 b) Synthetic scheme for l,2-diphytanoyl-§g:glycero- 3-phosphoglucose ................... 21 6 NMR and IR of l,2-diphytanoyl-§g:glycero-3- phosphocholine ..................... 25 7 NMR and IR of l,2-diphytanoy1-§n:glycero-3- phospho-[2(S)-N,N,N',N'-tetrabenzyl-2,6- diaminohexanol] ..................... 31 8 NMR and IR of l,2-diphytano l-3-B-D- [2,3,4,6-tetraacetoglucosyl -§g:glycerol ........ 34 vii Figure Page 10 ll 12 l3 l4 Differential thermograms of mixtures of diphytanoyl phosphatidylcholine in dipalmitoyl phosphatidyl- choline. Diphytanoyl phOSphatidylcholine: (a) 1 mole %; (b) 4 mole %; (c) 7 mole %; (d) 11 mole %; (e) 18 mole %; (f) 22 mole %; (g) 46 mole %. The sample consisted of 5 mg lipid in 5 ul of water ..... 41 Partial phase diagram of mixtures of diphytanoyl phosphatidylcholine in dipalmitoyl phosphatidyl- choline. Region A consists of a homogeneous liquid-crystalline phase 10 mole % DDPC in DPPC. Region B consists of a liquid-crystalline phase enriched enriched with respect to diphytanoyl phosphatidyl- choline in equilibrium with a gel phase enriched with respect to dipalmitoyl phosphatidylcholine. Region C consists of nearly pure gel phase dipalmitoyl phosphatidylcholine ................... 45 Illustration of vectors and angles relevant to lipid hydrocarbon chain motional model. H0 is the direction of the applied magnetic field; d is the director which is normal to the lipid bilayer interface; the angles are described in the text. Although in the figure all the vectors appear co-planar, generally the rela- tionship between the various vectors is three dimensional. {This figure is a partial repro- duction of a figure from the Ph.D. thesis of Dr. N. O. Petersen and is used here with the permission of the author.} ............... 47 (a) A representative FID taken at 57.4 MHz at 20°C. Total time elapsed is 500 usec, with each point corresponding to 0.5 usec. (b) A plot of FID intensity v§_time ........... 51 The range of values of Pt, the probability of a trans orientation, for given values of Au, the limit of the chain reorientation, which yield a proton order parameter S = - 0.17 i 0.04 {This figure is a partial reerHuction of a figure from the Ph.D. thesis of Dr. N. O. Petersen and is used here with the permission of the author.} ...... 53 A representative electron micrograph of DDPC vesicles produced by continuous sonication in 2HZO-Ringer ....................... 55 viii Figure Page 15 Photographs of schlieren patterns from representative sedimentation velocity experiments: (a) D¢PC vesicles formed by continuous sonication in 2HZO-Ringer; (b) DPPC vesicles formed by continuous sonication in H20- Ringer. Insert (a) Photograph of a schlieren pattern taken early in a sedimentation velocity experiment on DDPC vesicles formed by cycling sonication. Insert (b) Photograph of a schlieren pattern taken early in a sedimentation velocity experiment on DPPC vesicles formed by cycling sonication. Note the presence of a series of small "peaks" on the leading edge of the sedimentation boundary. Insert (c) A photograph of a schlieren pattern taken at a later time of the same DPPC vesicle sample. Note the appearance of a shoulder on the main vesicle peak ............ 63 16 PMR spectra. (a) 220 MHz spectrum of diphytanoyl phosphatidylcholine in C0013 at 20°C (sweep width 2500 Hz); (b) 100 MHz spectrum of small sonicated diphytanoyl phosphatidylcholine bilayer vesicles in 5 mM La(N03)3 at 55 C (sweep width 1000 Hz); (c) 100 MHz spectrum of sonicated diphytanoyl phospha- tidylcholine vesicles in 2H20-Ringer at 55°C (sweep width 1000 Hz); (d) 220 MHz spectrum of diphytanoyl phosphatidylcholine multilayer sus ension in 2H20- Ringer at 65°C (sweep width 20 KHz ; (e) 220 MHz spectrum of dipalmitoyl phosphatidylcholine in CDC13 at 20°C (sweep width 2500 Hz); (f) 100 MHz spectrum of small sonicated dipalmitoyl phosphatidylcholine vesicles in 5 mM La(N03)3 at 55°C (sweep width 1000 Hz); (9) 100 MHz spectrum of sonicated dipalmitoyl phosphatidylcholine vesicles in 2HZO-Ringer at 55°C (sweep width 1000 Hz); (h) 220 MHz spectrum of dipalmitoyl phosphatidylcholine multilayer suspen- sion in 2H20-Ringer at 65°C (sweep width 20 KHz) . . . . 72 17 Time course of paramagnetically induced choline resonance splitting. The figure shows a series of spectra taken at 220 MHz of DDPC vesicles in 5 mM La(N03)3 at 55°C. (a) standard; (b) im- mediately after addition of an equal volume of 5 mM Eu(N03)3; (c) 1 hr. 27 min. after addition; (d) 2 hr. 23 min.; (e) 23 hr. 4 min.; (f) 51 hr. 28 min.; (9) 74 hr. 47 min. The paramagnetically induced separations of the inner and outer choline resonances are given on the figure ........... 85 ix Figure Page 18 The predicted behavior of (1/T1) (dotted lines) 19 at 100 and 220 MHz calculated according to the motional model where the observed spin-lattice relaxation rates arise from the summed contri- butions of two correlation times according to the equation 1/T1 = AT + B/wozT H J. {For the sake of simplicity it has been assumed that A = B.} For DOPC: _8 _8 (a) Ill = 10 sec, I = 5 x 10 sec, Ea for both motions -l.5 KcalImole (b) T" = 10""8 sec, I = 5 x 10'8 sec, Ea for I = 2.0 Kcal/mol, J. ‘1 -8 Ea for I = 1.5 Kcal/mol; (c) I 10 sec, I = 5 x'J'TO'8 sec, Ea for I =‘2.5 Kcal/mole, E: for I = 1.5 Kcal/mole; (a) I = I = 10'8 sec, Ea +or IN = 3.0 Kcal/mole IIEa for I = 1.0 Kcallmole ........... For DPPC: -e (e) T - 10 sec, I” = 10-10 J_ sec, Ea for T” = Ea r = 3.0 Kcal/mole; (f) T = 10‘11 u sec, Ea fo'F'Ill = Ea for Ii.= 3.0 Kcal/mole ...... 92 Spin-lattice relaxation rates (l/Tl) of the hydrocarbon methylene protons in sonicated bilayer vesicles as a function of reciprocal temperature and at two nmr frequencies: (a) 100 MHz, (b) 200 MHz. (3 diphytanoyl phospha- tidylcholine; O 10 mole % diphytanoyl phos- phatidylcholine in dipalmitoyl phosphatidyl- choline-d62; [J dipalmitoyl phosphatidylcholine; II -lO% dipalmitoyl phosphatidylcholine dis- persed in 90% dipalmitoyl phosphatidylcholine- d62 ........................... 95 Figure Page 20 Spin-lattice relaxation rates (l/T1) of the hydrocarbon methyl protons in sonicated bilayer vesicles as a function of reciprocal temperature and at two NMR frequencies (100 and 220 MHz). 0 diphytanoyl phosphatidylcholine; Q 10 mole % diphytanoyl phosphatidylcholine dis- persed in dipalmitoyl phosphatidylcholine-d62 ...... 100 21 Observed pmr linewidths of the hydrocarbon chain methylene protons at 220 MHz. ( 0) small soni— cated dipalmitoyl phosphatidylcholine vesicles (measurements below the thermal phase transition were obtained using 20 Khz sweep width); ([3 ) vesicles prepared by sonicating a mixture of 10 mole % diphytanoyl phosphatidylcholine in dipalmitoyl phosphatidylcholine-d62; and (13.) sonicated diphytanoyl phosphatidylcholine vesicles as a function of temperature .............. 108 xi INTRODUCTION Statement of Problem l,2-Diphytanoyl-§g:glycero-3-phosphocholine (DOPC) was originally synthesized by Neisbach and co-workers [1], and its electrical properties in "black" lipid membranes were investigated. The author's interest in this molecule was motivated by a desire to find a modification in the acyl portion of a synthetic phospholipid which might be expected to mimic the presence of unsaturation in the hydrocarbon chain but would avoid the instability towards oxidation of the double bond which is so troublesome in unsaturated lipids. Unsaturation or properties similar to those conferred by the presence of unsaturation are desirable in a synthetic lipid which is to be used in model membrane studies for the following reasons. Naturally occurring lipids generally contain at least one unsaturated site in the acyl chains, and evidence is accumulating in enzyme recon- stitution studies that unsaturated sites in lipids are necessary to the proper functioning of those enzymes which are normally membrane bound [2]. Therefore,since chain branching was expected to mimic unsaturation, the synthesis of DDPC was undertaken according to the published synthesis mentioned above. In addition, the author proposed to synthesize two other lipids bearing the same acyl chain but with polar head groups which would carry either a positive or a negative charge at physiological pH. 1 The permeability of the three lipids in vesicular suspensions toward neutral molecules was to have been investigated under conditions where the pH and ionic strength of the aqueous media were varied. The assumption that such a study would be feasible has since been proved to be naive, at least in the case of the two charged lipids. It has been shown that sonicated suspensions of charged lipids are unstable toward fusion and multilayer formation [3]. Moreover, the author has found that lipids containing the phytanic acyl moiety behave in a complex and poorly understood manner in the presence of salts. The author wished,by measuring the permeation rate of neutral molecules through bilayer model membranes,to test a model for permeation which was based on the premise that passive diffusion in membranes takes place via "kink" propagation [4]. The originator of this model, H. Trauble, states that kinks in the hydrocarbon region of the membrane create "free volume" in this phase into which such molecules as water, glucose and hydrated ions can dissolve. Then the solute molecule is thought of as crossing the membrane by jumping from one kink site to another in a random walk or by being carried by the kink as it propagates along the hydrocarbon chain. A kink is formed when an extended hydrocarbon chain in the all-trans conformation in the membrane undergoes a pair of B-coupled gauche rotations {one rotation in the clockwise sense (9+), the other in a counterclockwise sense (9-3 [4] (see insert below). If the two gauche bonding interactions are separated by more than one carbon-carbon bond in a staggered conformation the chain is said to contain a "jog". The intrinsic probability of the presence of a kink or jog in a hydrocarbon chain is a function of the number of sites available in the chain. Therefore, since the number of carbons involved in a kink is four, whereas the number involved in a jog is six or more, in a chain of n carbons, the formation of a kink is considered to be a more probable event than the forma- tion of a jog. The physical consequence of the introduction of a kink or jog into a hydrocarbon chain, aside from the obvious conformational change, is to shorten the chain hY’b].252 and to increase the chain cross-section area very slightly. This effect is illustrated by the photographs of the space filling model of DDPC in Figure l. The methyl branches of the phytanic acid chain introduce steric barriers to kink formation and propagation in lipids containing this molecule which do not exist in saturated unbranched acyl chains inlipdds(see Figure 2). The requirement that four carbons be involved in the formation of a kink leads to the conclusion that the formative event in DOPC must involve a methyl branch point. The absolute configuration at these branch points (see Figure 2) is Figure 1 Photographs of space-filling models of DDPC showing the primary (left—hand) phytanic acyl chain in various conformations: (a) all-staggered conformation; (b) conformation containing one gauche chain-chain interaction; (c) Kinked conformation. The insert (lower-right) shows schematically the effect of the introduction of a kink on the length of the chain, the chain cross-sectional area, and the disposition of the carbon backbone of the chain. FIGURE 1 b flaw—TA: all—t rrrrrr ham kmk c insert Figure 2 Potential energy as a function of rotation angle about the carbon- carbon bond illustrating: (a) phytanic acyl moiety, rotation of i 120° involving a methyl branch; (b) palmitic acyl moiety, any internal rotation of 1120“. O 0 (K) (R) R,s } PHYTANIC ACYL GROUP 120° '\ kcol JAE wlmole fv I6” 0 i PALMITIC ACYL GROUP '5 120° 60 0- oo :20 II a mo kcal -" kcal AEm2mole lAE 012 mole FIGURE 2 such that the methyl groups are disposed in one direction with respect to the chain in the all-trans conformation. Consequently, when a gauche rotation takes place which involves a methyl branch, the rotational potential energy function is markedly asymmetric, and this asymmetry to rotation has the same sense for all methyl groups in the chain. As illustrated in the bottom portion of Figure 2, a chain which contains only methylene groups does not experience any asymmetry in rotational potential energy. A jog in the phytanic acyl chain which resulted from a pair of v-coupled g: rotations would encounter two types of rotational energy barriers. There are only three positions (the starred carbons) in the phytanic acyl chain about which the chain can form a y-coupled jog without participation of a methyl branch. The introduction of a y-coupled jog by rotation about any other carbon-carbon bonds in the phytanic acid chain must involve two methyl branch points, and one of these rotations must take the high energy path to the gauche conformer. From symmetry considerations arcoupled jogs should have the same potential energy profile as the kink. However, the population of 6-coupled jogs is predicted to be quite low due to the finite length of the phytanic acid chain. On the basis of the different ground state energies in the two systems the population of kinked conformers in the phytanic acyl chain is expected to be different from that found in a straight chain of comparable chain length, e.g. palmitic acid (see Figure 2). This energy situation arises in the following manner. In the phytanic acyl moiety rotation about a carbon-carbon bond at a methyl branch point through the lower energy transition state merely exchanges a methyl-chain gauche interaction for a chain-chain gauche interaction (see the g- rotation in Figure 2a). However, the rotation through the higher energy transition state (g+, Figure 2a) introduces an additional gauche interaction as well as interchanging chain-methyl and chain-chain gauche interactions. Thus, it is evident that the gauche conformer which results from the low energy rotation has approximately the same ground state energy as that of the all-trans form, while that resulting from the high energy rotation possesses a higher ground state energy as a result of the additional gauche interaction. However, gauche conformers must occur in pairs in lipids in the lamellar phase. The other half of the kink in the phytanic acyl chain does not involve a methyl branch, and thus the potential energy barrier is symmetrical toward rotation in the 9+ and 9- directions, and either rotation introduces another chain-chain gauche interaction. Therefore, the net change in ground state energy over the all-trans form for the introduction of a kink into the phytanic acid chain should be'tl Kcal/mole if rotation involving the methyl group passes over the lower energy barrier and m2 Kcal/mole if the rotation involving the methyl group must pass over the high energy barrier. The same analysis of the straight chain analogue of phytanic acid (i.e. palmitic acid) yields a totally symmetrical rotational energy profile for the introduction of a kink. Thus the difference in the ground state energies between the all-trans form and the conformer con- taining a kink is expected to be'mz Kcal/mole. It follows that 10 the population of kinked conformers is predicted to be greater in phytanic acid than in palmitic acid at the same temperature. However, if kinks act as sites and/or carriers for passive diffusion in membranes, then not only is the population of kinks important, but also the rate of kink propagation. The frequency with which a kink or a jog appears at a particular position in an unbranched hydrocarbon chain in a lipid is a straightforward function of the jump rate for the kink and the mean square distance travelled per jump. In the case of a phytanic acyl substituent this frequency can be estimated only by summing the products of the intrinsic probabilities for the various kinks and jogs and the jump rate associated with each. Therefbre, the author concluded that an attempt to correlate kink populations and kink diffusion rate with measured permeation rates for a neutral molecule through a bilayer membrane composed of DDPC would not be straightforward. No naturally occuring lipid containing phytanic acid has yet been reported. However, dihydrophytol, the reduced form of phytanic acid, has been shown to be the only detectible hydrocarbon chain present in the diether analogues of the glycerol derived lipids found in the membranes of certain halobacteria [5]. The protein, bacteriorhodopsin, which consitutes the "purple membrane" is found in species of these bacteria [6]. The purple membrane has been shown to be capable of performing "photosynthesis" j_n_ gjyg_and jg_vj§rg_[7, 8]. A proton gradient is generated in the presence of light, and, concurrently, ATP is synthesized. There- fore, the purple membrane in reconstituted systems is considered 11 to be an ideal model system for testing Peter Mitchell's chemiosmotic hypothesis which states that a proton gradient generated in oxidative phosphorylation (and photosynthesis) is discharged via the production of ATP. The physiological significance of the absence of hydrocarbon moieties other than dihydrophytol in the lipids associated with the purple membrane is unclear. It is possible that these lipids confer special permeation properties on the halobacterial membranes. {The growth medium for these halobacteria is 3.4 M NaCl, 0.10 M 3 MgCl 0.03 M KCl and 1.4 x 10' M CaClz; lysis is induced simply 2, by lowering the salt concentration to 1M NaCl [9].} Moreover, it may be found that bacteriorhodopsin does not remain functional in the absence of branched chain lipids. With the intent of clarifying the structure-function relationship of this particular acyclic diterepenoid moiety in membranes, the author has undertaken a study to elucidate the structural and dynamic properties of model membrane systems containing synthetic phospholipids with phytanic acid as the acyl protion of the lipid. PART I Synthesis of Lipids Containing Phytanic Acyl Substituents The view that lipid bilayers constitute the structural matrix of biological membranes has provided stimulation for the study of model membrane systems such as lipid monolayers, single lipid bilayers (vesicles and "black" lipid membranes), and multilamellar lipid syspensions (frequently called liposomes). The bulk of this work has been carried out on lipids isolated from natural sources. However, there are several objections which can be raised to the use of lipids isolated from natural sources in studies aimed at the physiochemical characterization of lipids. The isolation procedures utilized for the separation and purification of natural lipids are chromatographic, and the elution patterns are based on differences in the polarity, usually of the polar head group, of the molecules. Therefore, each lipid isolated in this manner actually consists of a complex mixture of compounds with similar physical characteristics but widely different acyl substituents. Moreover, the acyl substituents generally contain at least one unsaturated site which renders these lipids susceptible to oxidation. Chemically stable phospholipids have been synthesized with saturated straight chain acyl groups. However;when one or both 12 T3 of the acyl chains are shorter than fourteen carbons,the molecule exhibits appreciable water solubility which destabilizes bilayers formed from such a lipid. 0n the other hand, when the acyl sub- stituents have chain lengths greater than 16 carbons, the lipid possesses a gel to liquid-crystalline phase transition temperature which exceeds 500 C [1]. These lipids can be utilized in studies aimed at understanding the nature of lipid-lipid interactions but are quite unsuitable for enzyme reconstitution studies on enzymes isolated from mammalian or bacterial sources. The author had reason to believe that the synthetic lipid l,2-diphytanoyl-3-§g;phosphatidylcholine (DDPC) would possess physical properties which would avoid the experimental drawbacks encountered with both the unsaturated natural lipids and commercially available saturated phospholipids. It was known that phytanic acid was a liquid at room temperature. Therefore, DOPC would be expected to possess a gel to liquid-crystalline phase transition temperature which would more closely resemble those found for natural phosphatidylcholines (i.e., O-SOC). At the same time, it was predicted that in the case of this synthetic lipid both the water solubility brought on by short chain length and the chemical instability resulting from the presence of unsaturated sites in the chain would be avoided. Since DDPC showed promise of great versatility in model membrane studies, the author proposed to synthesize this lipid following a published procedure [1]. The reaction scheme for the synthesis is shown in Figure 3. Certain protions of the published I4 a£fi¥§ new“? EWWE 3.3%.; Nngwgwwv— + —0 3:20 I z 4 fl 8 85:8»: .2. 3.3 Nun-.20 0:0—0 .oFo—m $00 Theo—E: : Saw-o 433 «EN :9 63:: ”Sea a £52 :4. 5,38 83: 6: a in— :5. a.“ .6 A :9 ¢0 a: v 0.35 .5».— 3650 .N 2% Ne... a: .— 223 FIGURE 3 ,2 -phosphocholine. the structures -glycero-3 implied in -diphytanoyl-sn y of the glycerol 15 {No stereochemistr in Figures 3 through 5.} _Synthetic scheme for l I5 synthesis were particularly problematic, and these sections will be elaborated upon. The acylation step in this synthesis is very inefficient. Therefore, the phytanic acid should be carefully purified in order not to add to the multitude of side products. The phytanic acid used in the acylation was synthesized by reduction and oxidation of phytol. Even the best commercially available phytol (E. M. Merck, Darmstadt) contains a significant amount of impurities. The author has identified indane, substituted pyrroles, a C15 branched hydrocarbon, a 019 a-Me alcohol, and small amounts of highly colored tetrapyrrole fragments via GC-mass spectro- metry. {Since chlorophyll is the most probable source for com- mercial phytol, these impurities seem quite reasonable.} However, neither the phytol nor phytanic acid synthesized from it was responsive to purification attempts via high pressure liquid chromatography, various silica gel and alumina columns, reverse- phase liquid chromatography, preparative gas chromatography, nor preparative layer chromatography. Apparently this acyclic diter- penoid molecule is amphoteric and forms micelles in organic solvents. {Since micelles cause solubilization of other molecules which resemble physically the major component of the micelle, purification of molecules which form micelles via conventional methods is often unsuccessful.} The author found, however, that the methyl ester of phytanic acid could be rendered > 99% pure by distillation on a spinning—band column. 16 In the published synthesis trichloroethylchloroformate was used to block the hydroxyl at the 3 position of the glycerol. Dr. F. R. Pfeiffer informed the author that this blocking group was extremely susceptible to cyclic carbonate formation once the isopropylidene blocking group had been removed from the l and 2 hydroxyls of the glycerol. Therefore, the author elected to make a minor modification to the synthesis at this point and use benzylcloroformate to block the hydroxyl at the 3 position of the gylcerol. This modification proved to be futile. No im- provement over the reported yields in the acylation reaction was realized. The poor yield for the acylation reaction appears to result from the steric hindrance encountered in esterification of the secondary (2°) hydroxyl of glycerol. Acylation of the primary hydroxyl appears to take place with relative ease, whereupon the bulky phytanic acyl group blocks the approach to the secondary hydroxyl. A competition is established between acylation of the secondary hydroxyl via transesterification and the attack of the secondary hydroxyl on the benzylcarbonate to form the cyclic carbonate and benzyl alcohol. Therefore, this competition must be eliminated in order fOr the acylation to proceed with reasonable yield. The author had also proposed to synthesize two other lipids. The hydrocarbon portion of the lipids was to be invariant, i.e. phytanic acid, while the polar portion of the lipid was to be modified to bear either a net positive or negative charge at I7 physiological pH. The precursor of the positively charged polar head group was prepared according to the scheme shown in Figure 4. The author had originally tried to use benzylchloroformate to block the amine groups of lysine while the carboxylate function was reduced to an alcohol. But the carbamate proved much less stable to reduction than the acid,and that approach had to be abandoned. Benzyl was settled upon as the blocking group of choice because of its stability towards common reducing agents which would be used to reduce the acid, and the ease with which it is removed by catalytic reduction. Under the conditions of benzylation the carboxylic acid group was converted to the benzyl ester. {It is possible that racemization of the L(-)-lysine was affected to a small extent under the benzylation conditions. See Experimental.} The author was hesitant to subject the blocked lysine benzyl ester to a vigorous reducing agent such as LiAlH4. Therefore diborane was selected as a possible reducing agent for the ester. The ester was reduced, but the diborane also displaced the benzyl blocking groups from the amine functions. Therefore, a milder reducing agent had to be sought. LiBH4, which is known to be a selective reducing agent, is effective in reducing ester functions to alcohols and proved successful [10]. It should be noted that the reduced blocked lysine appears to be especially suceptible to air oxidation. The nature of this reaction is not known, but the product must be stored over P205 under N2 or Ar at -20°C. In this reaction scheme the coupling of the reduced blocked lysine with l,2-diphytanoyl-sg; glycero-3-phosphodichloride went without the multitude of side 18 l I «an a 692.2,..98C38 02320 82A Iv «Anny: «=20 wanxa—o ohhfl . . w 3:20 1+: «:91 =o,=o-u-.§ez€e $8 05:»? .53 .=m: N as a $5.. am: .3. S§éu§e=fe II $98.39,: _. as 1. FIGURE 4 Synthetic scheme for l, ycero-3-phospho-[2(S)- gl xanol]. s_n- 2-diphytanoyl- N,N,N',N',-tetrabenzyl-2,6-diaminohe I9 reactions which were encountered in the phosphatidylcholine reactions. The introduction of the negatively charged polar head group was attempted via several approaches. The scheme shown in Figure 5a was originally proposed. However, the difficulties encountered in the introduction of the glucoside linkage between the lipid and the sugar proved very time consuming and the yields were consistently low. Therefore, since the characterization of the lipids in model membrane systems seemed more important than the pursuit of a solution to this synthetic problem, this approach was abandoned after preparation of the blocked glucoside. As an alternate approach to the introduction of a negatively charged polar head group, the author decided to attempt the synthesis of phosphatidylglucose. The existing literature indicated that the method of choice for coupling phosphatidic acid with glucose involved the attack of a silver salt of phos- phatidic acid (or other suitably activated phosphate) on the a-bromoglucose pentaacetate. The author originally attempted to generate the silver salt of diphytanoyl phosphatidic acid by quenching the phosphonyl didhloride with saturated AgNO3 solution. This salt was isolated and the coupling was attempted. It was unsuccessful. When phosphatidic acid was reacted with the acetobromoglucose in the presence of AgOSOZCF3, the desired reaction did take place to a small extent (see Figure 5b). Through a series of unfortunate circumstances the blocked lipid was exposed to moisture and kept at room temperature for an extended period 20 .pogmuapmumm-u_um u_=ompamlml_»mou3—mlolmumnpzdcmuazawuuw.p so; msmgum ovumsu=>m on #50: \e: I \2. nmfifiéfi 1 fig... I ”manila“: \2. a . \2_ whiz... u”. I .mulizle. a . a. .4 e ._ a iaéIefiq l... + .mu. an 23.....K 23R .3 a 21 ~/(:II,0Ac Ech11 + Ac . H0 AcO/~ A Ag08020F3 anh. HCH3 ‘Vflfib FIGURE 5 b R: 019"” Synthetic scheme for l,Z-diphytanoyl-§g;glycero-3-phosphoglucose. 22 of time. The product decomposed. An NMR and IR spectrum were taken of the crude product before decompositionand the interpretation of these spectra are consistent with the expected product. However, the evidence is insufficient to state conclusively that the desired product was isolated. Having carried out the above syntheses based on existing literature in the field, the author would like to suggest the following modifications to the cited procedures. l) In the synthesis of the l,2-diacylated glycerol the author would suggest the use of the benzyl blocking group at the 3 position on the glycerol. This group would go on under mild basic conditions which would not disturb the isopropylidene blocking group and could be easily removed via hydrogenolysis. In addition, it should be completely resistant to attack by the 2° hydroxyl of glycerol and thus the yields of the l,2-diacyl glycerol should be greatly improved. 2) The use of the acid chloride for acylation may not be necessary. It has been reported that acylations of this type can be accomplished in greater than 90% yield via the use of carbonyl- dimide . as a dehydrating agent [ll]. In the case of phytanic acid this would be of great aid since conversion to the acid chloride produces a side reaction which results in the production of a highly colored product of complex nature which is difficult to remove from the desired product. 3) The glucoside coupling reaction might proceed with higher yields in the phosphatidic acid were treated with carbonyldimidazole to activate the phosphate [12]. 23 4) If the B linkage in phosphatidylglucose is not desired or considered necessary, then the sugar lipid should be available in good yield simply by addition of the a-D-glucosepentaacetate to the phosphodichloride intermediate. EXPERIMENTAL I l,Z-Diphytanoyl-§g:glycero-3-phosphocholine (DQPC). A waxy solid having an Rf value of 0.33 to 0.35 on Merck silica gel 6 analytical plates eluted with 65:25:4 (v/v) chloroform-methanol- water, was prepared in 10-12% yield from l,2-diphytanoy1gsggglycerol by the procedure of Redwood gt_al. [l] with the following modifi- cations. (See Figure 6 for the IR and NMR of the product.) 1,2-Isopropylidene-§E:glycero—3-benzy1carbonate. To 1,2-isopropylidene-sg:glycerol(35.4Sh 0.268 mol) and pyridine (21.6 ml, 0.268 mol) in dry chloroform at 5°C benzylchloroformate (38.8 ml, 0.268 mol) was added drop wise with stirring. The re- action mixture was then allowed to come to room temperature,and stirring was continued for 24 hours. Then ether (500 ml) was added to the reaction mixture, and the ether-chloroform fraction was washed with 1% hydrochloric acid, water, and 5% sodium bicarbonate followed by a final water wash. The organic phase was dried over anhydrous sodium sulfate. The solvents were removed by rotary evaporation. The product was purified first by vacuum distillation (135°C, 1mm) to remove volatile contaminants, followed by filtration through Florisil (Baker) to remove polymeric products. 0ily l,2- isopropylidene-ggrgycero-B-benzylcarbonate (65-709, 90%) was 24 .mcmpaguocamo:a-muocmua_mnmw-pxocuuazamn-~.— do «H van «:2 m!» no.0 h... «0.. 00d 0m¢30i m2: m: _ _ _ _ . omamlmmmL 005 000 08000. 000. .kcu 26 recovered. NMR (00013) 6 1.3(d,6H), 3.3-4.4 (cmplx m, 6H), 5.0 (5,2H) 7.2 ($.5H). 3(R,S),7(R),11(R),15-Tetramethy1hexadecanoic acid (phytanic acid). Phytol (>95% pure) was purchased from E. M. Merck, Darmstadt. The alcohol was used as purchased and was hydrogenated and oxidized according to a published procedure [1]. The acid was then converted to the methyl ester via conventional procedures. This methyl ester of phytanic acid was subjected to distillation on a spinning band column with 1000 theoretical plates. The fractions coming over at 1-2 mm pressure and from 142.5 to 146.500 were fbund to contain the methyl ester of phytanic acid with less than 1% impurities detectable by GC-mass spectroscopy. The methyl ester was saponified, and pure phytanic acid was recovered in an overall yield of'h50% based on the crude acid. The acid had: IR (neat) 2880, 2640, 1790, 1450, 1185, 935 cm"; NMR (00013) a 0.7-0.96 (m, 15H), 0.96-2.4 (br m, 24H), 11.0 (s, 1H). 1,2-Diphytanoyl-§g;glycerol. 1,2-Diphytanoyl-§n:glycero-3-benzy1carbonate (10.53 g, 0.013 mol) was dissolved in 75-100 ml of 95% ethanol-ether (sufficient ether was added to render the solution homogeneous) and 5% Pd on charcoal (1.59 g) was added. The reaction mixture was flushed with hydrogen gas and then allowed to react under 1 atm. of hydrogen for approxi- mately 1 hour (a trace of H+ increases the rate). The solution was then filtered to remove the catalyst. The catalyst was washed with 27 1:1(v/v) ether chloroform (75 ml). The combined filtrate was washed with water (3x50 m1 ), 5% sodium bicarbonate,and saturated sodium chloride solution and dried over anhydrous magnesium sulfate. Removal of the solvents at room temperature {to avoid the conversion of the l,2-diacy1 product to the 1,3 diacyl compound} on a rotary evaporator yielded l,2-diphytanoy1-§g;glycer01 (8.06 g, 0.012 mol). The product had: IR (neat) 3550, 2940, 1755, 1470, 1390, 1270, 1172, 793, 740, 700 cm"; NMR (00013) 3 0.70-0.82 (m, 30H), 0.32- 2.58 (br cmplx m, 49H), 3.69 (d, 2H), 4.26 (d-q, 2 H), 5.06 (t, 1H). 2(S)-benzyl-N,N,N',N'-tetrabenzy1-2,6-diaminohexanoate. L(-)-1ysine(7.39, 50 1111101) was dissolved in 2:1 (v/v) ethanol- water (~75m1) to which 7N KOH (40 ml) was added. The solution was brought to a boil and benzylbromide (36 m1, 300 mmol) was added drOpwise with stirring. After the addition of the benzylbromide was complete, the reaction mixture was refluxed for 1 hour. The reaction mixture was cooled in an ice bath while it was acidified with acetic acid. The acidified reaction mixture was then ex- tracted with chloroform. The chloroform fractions were washed with water and saturated sodium chloride. The solvent was removed on a rotary evaporator, and the resultant glass was chromatographed on Florisil (eluted with chloroform). The solvent was removed,and the clear glassy product was recovered (14.5 g, 24 moles, 50%). The compound showed: IR (neat) 3080, 3060, 3020, 2940, 2800, 1720, 1490, 1450, 1360, 1130, 740, 690 cm"; NMR (00013)5 1.07-1.82 (cmplx br. m, 6H), 2.32 (t, 2H), 3.30 (t, l H), 3.45 (s, 4H), 3.7 (q, 4H), 5.17 (q, 2 H), 7.3 (m, 25H). 28 2(S)-N,N,N',N'-tetrabenzy1-2,6-diaminohexanol. Pentabenzyl-L(-)-1ysine(7.5 g, 13 mmol) was dissolved in diglyme(20 ml) which had been distilled from CaH. To this solution 1M L18H4(20 ml) was added,and the reaction mixture (protected from water) was refluxed on a steam bath for 3 hours. The reaction mixture was then poured over ice (0400 g) to which concentrated hydrochloric acid (10 m1) had been added. The product was extracted from the water phase into chloroform. The combined chloroform fractions were washed with 5% sodium bicarbonate, waten,and saturated sodium chloride and dried over anhydrous magnesium sulfate. The solvent was removed,and 5.4 g of product was recovered which was 60% reduced. This compound was used without further purification in the next reaction. IR (neat) 3430 cm']; NMR (CDC13) 6 1.27 (s, 1 H) represent those portions of the spectra which were not identical to that recorded above. 1,2-Diphytanoy1-§n:glycero-3—phosphodichloride. l,2-Diphytanoy1-§nfglycer01 (7.58 g, 9.3 mmol) dissolved in a minimum of dry chloroform was added dropwiSe to phosphorus oxychloride (0.9 ml, 10 mmol) in 35 m1 of dry pyridine at 0°C with stirring over a period of half an hour. After the addition was complete, the reaction mixture was stirred for 1 hour at 0°C and then allowed to come to room temperature. The reaction mixture was stirred for 2 hours and then used immediately. 29 1,2-Diphytanoy1-§n;glycero-3-phospho-[2(S)-N,N,N',N'-tetrabenzy1- 2-6-diaminohexanol] (DQPL). To half the solution of 1,2-diphytanoy1-§g;glycero-3-phosphodi- chloride in pyridine at 0°C was added the 60% reduced 2(S)-benzyl- N,N,N',N'-tetrabenzyl-2,6-diaminohexanoate (5.4g) in dry chloroform dropwise over a period of 1/2 hour with stirring. After the addition was complete the solution was allowed to come to room temperature. The reaction mixture was stirred at room temperature for 75 hours while the reaction was followed via TLC. (Merck silica gel 0 analytical plates developed in 65:25:4 (v/v) chloro- form-methanol-water; revelatory agent molybdenum blue [13].) At the end of that time the reaction mixture was cooled in an ice bath while 0.1N KCl (20 ml) was added with vigorous stirring. Then Me0H (20 ml) was added, and the pH of the solution was ad- justed to 3 with concentrated hydrochloric acid. A small amount of chloroform was used to extract the organic phase, and the combined chloroform fractions were stripped of solvent, and the residue was dried over P205 under N2 at -20°C. The dried residue was applied to the Unisil column,and the column was eluted with chloroform containing methanol (0-50%‘v/v). The 5% methanol in chloroform fraction contained the product slightly contaminated with side products. Repeated chromatography on preparative layer plates (Merck silica gel G developed sequentially in chloroform and 5% methanol in chloroform, fractions eluted with 50% methanol in chloroform) yielded 0.88 g (0.71 mnol, 15%) of the product. {It should be noted that a significant amount of this product was lost 30 in a shipping accident and a subsequent freezer breakdown.} The IR and NMR spectra of the product are shown in Figure 7. 1,2-diphytanoyl—§n;glycero-3-phosphoglucosetetraacetate. 1,2-Diphytanoy1-§n_-glycero-3-phosphate (3.79 g, 4.7 mol) was dissolved in ~50 m1 of anhydrous toluene and pyridine (3-5 ml, >40 111n01). Acetobromoglucose (1.85 g, 4.5 1111101) in a minimum of anhydrous toluene was added to the above solution in the dark. Silver triflate (AgOSOZCF3) [13] (049, > 3 mmol) was added to initiate the reaction. The reaction mixture was stirred in the dark for'v40 hours and followed on TLC (Merck silica gel G—analytical plates eluted with 65:25:4 (v/v) chloroform-methanol-water; revela- tory agent molybdenum blue). The solvents were removed on the rotary evaporator and under vacuum (0.01mm). The excess sugar was removed by repeated extractions of the glassy reaction mixture with petroleum ether followed by dissolution of the petroleum ether insoluble sugar in acetone and reprecipitation with petroleum ether. The petroleum ether precipitations were continued until the sugar precipitate was white and fluffy (the petroleum ether fractions were yellowish brown). The petroleum ether fractions were combined and treated with 0.05 M NaI in acetone to remove the excess silver. A visible reaction took place which resulted in the decolorization of the petroleum ether fraction, and a white precipitate formed. The precipitate was removed by filtra- tion on a sintered glass funnel and washed with petroleum ether. The precipitate was slightly soluble in acetone. An NMR of the .Hpocostocwsmwu1o.ml_»~=maacumu -.z..z.z.z-amuNa-ogamoga-m-ocmuspmnmm-.aceauazawu-~._ co as new «:2 30 8.3.5468 ow." cow ._3 3W «0% new. a .msFJJJ _ 4? ~=Z7_ 31 _ n $50.”. 7:8 Eggs: com 80. com. 8203. 89.008003003003003 — _ _ _ _ q — _ _ _ m. 32 solid showed two singlets (6 2.09, 3.23). The solvents were stripped from the petroleum ether fraction, and the residue was applied to a Florisil (Baker) column which was eluted with petroleum ether, petroleum ether-chloroform, chloroform, and methanol in chloroform. The fractions which were eluted with 2-5% methanol in chloroform were found to contain the crude product (0.06g, 5.25x10'2 mmol, 1%). The crude product was never purified further. Through a series of unfortunate circumstances it was exposed to moisture at room temperature for a prolonged period. This treatment apparently caused it to decompose into the sugar and phosphatidic acid. NMR of the waxy freshly isolated crude product failed to resolve the acetate singlets, but some NMR intensity was present in the appropriate range,162.12-2.96. IR of the same material showed no hydroxyl; absorption in the region 1 to l460cm'1 was very strong and complex. The de- from 900 cm- composed product consisted of a solid which had low chloroform solubility, and a waxy substance which dissolved readily in chloro- form. An IR of the mixture showed a strong broad absorption at l 1 l W3400cm' , and the intensity in the region from 900 cm' to 1460 cm' had diminished greatly and the absorptions were easily identified. TLC on silica gel G analytical plates (Merck) in 65:25:4 (v/v) chloroform-methanol-water of the freshly isolated crude product gave a Rf of 0.65 for the major component. TLC of the hydrolyzed product under the same conditions yielded a Rf of 0.23 and 0.05 for the two major components. 33 B-D-Glucopyranosyl-(l,2—diphytanoy1~§fl:glycerol)-2,3,4,6- tetraacetate. 1,2-Diphytanoy1-§fl-g1ycerol (0.1 g, 6 mnol ), acetobromoglucose (2.5 g, 6 mmol), and anhydrous potassium carbonate (0.88 g, 6 mmol), were taken up in anhydrous acetonitrile. Silver triflate [13] (1.6 g, 6 mmol) dissolved in a minimum of anhydrous acetonitrile was added dropwise in the dark with stirring to initiate the reaction. The course of the reaction was followed by disappearance of the starting materials via TLC (conditions given previously). The reaction was allowed to continue at room temperature with stirring in the dark for “54 hours. The reaction mixture was then filtered to remove solids, and the filter was washed with aceton- itrile. The combined filtrates were stripped of solvent on a rotary evaporator. The unreacted sugar was precipitated by taking the residue up in ethyl ether. The sugar was removed by filtration, and the filtrate was rotary evaporated to remove the ether. The residue was applied to a Florisil column which was eluted with petroleum ether, petroleum ether-ethyl ether (7:1; 5:1v/v), and petroleum ether-ether-acetone (4:1:0.5 v/v). The desired product was eluted with the petroleum ether-ether-acetone mixture with considerable tailing. Those fractions showing only one spot on TLC (conditions same as above, Rf 0.80-0.87) were pooled. The factions showing more than one species on TLC were accidently discarded. The product yield (0.079, 7.14x10'2 mmol, 1%) therefore, represents a lower limit. IR and NMR spectra of the product are shown in Figure 8. 34 mm '“ 1 l l 1 l l l l I 3600 3200 2800 2400 2000 1800 1600 1400 1200 1000 WAVENUMBER cm" FIGURE 8 NMR hr -—n~..._.z~A~—.A——— - e, L4\.~LJL 8 £67 £63539 £9118» 100$ NMR and IR of 1,2-diphytanoy1-3-B-D-[2,3,4,6-tetraacetoglucosyl]- sn;glycerol. PART II Physicochemical Characterization of l,2-Diphytanoyl-§n;glycero-3-phosphocholine in Hydrated Systems In 1972 Redwood and coworkers [1] demonstrated that "black" lipid membranes (BLM) could be formed from decane solutions of 1,2-diphytanoy1-§gfglycero-3-phosphoch01ine (D¢PC). These BLM displayed characteristically high d.c. resistance, a typical current voltage response, and good mechanical stability. In recent years there has been much criticism of the use of BLM as a model for biological membranes. The principal objection has been that BLM formed by conventional methods {suspending a hydrocarbon solution of a lipid across a circular hole in a hydrophobic support immersed in an aqueous medium and allowing the droplet to thin to a film of molecular dimensions} contained significant amounts of solvent [15, 16, 17]. Therefore, it was suggested that those properties of lipids which were thought to be observable in this system {i.e., electrical resistance, capaci- tance, interfacial tension, and to some extent, ion permeation and selectivity} reflected more the properties of the dissolved solvents than the lipids themselves [16]. Moreover, the types of physical properties studied in this fashion did not lend them- selves to microscopic interpretation. 35 36 The author, having prepared a Master's thesis on the subject of BLM, was convinced that the then new technique of sonicating multilamellar aqueous suspensions of lipids to produce single bilayer "vesicles" [13] would provide a more useful and less problematic model membrane system. Therefore, the author proposed to make a study of DQPC in vesicular suspension. Specifically, the author wished to establish the rate of kink diffusion in the hydrocarbon region of D¢PC bilayers by using the hydrocarbon protons as a probe to this motion. Subsequently the permeation rate of a neutral molecule through the lipid bilayer was to be measured. A study was then to have been undertaken of this data to establish whether a correlation existed between the rate of kink diffusion and the permeation rate. However, before proceeding with the studies mentioned above, it was necessary to characterize the behavior of this synthetic lipid in aqueous suspension. To this end experiments were per- formed on unsonicated suspensions of DOPC to determine (i) the phase formed by this lipid when fully-hydrated, (ii) the gel to liquid-crystalline phase transition temperature, (iii) the properties of lipid mixtures containing DOPC, and (iv) the order parameter for the hydrocarbon chain. Then sonicated suspensions of D¢PC were examined to ascertain (i) the presence of single bilayer vesicles, (ii) the average size of the vesicles formed by soni— cation, (iii) the heterogeneity of the vesicle sample, (iv) the partial specific volume of D¢PC in vesicles, (v) the integrity of the vesicles toward ion permeation, (vi) the approximate rates 37 of kink propagation and chain reorientation in the hydrocarbon portion of the bilayer, and (vii) the ability to detect lateral phase separation in vesicles prepared from lipid mixtures containing DOPC. The results of these studies are presented and discussed below. Characterization of Unsonicated Aqueous Suspensions of DOPC. Establishment of phase formed by fully hydrated DOPC. The number of phases observed in lipid-water systems, as a function of composition and temperature, is quite large. However, without known exception, fully hydrated lipids exist only in the L0 or LB' lamellar phase where the "infinite" bimolecular leaflets of the lipids are well separated by a water region [19]. The individual lipid molecules in each half of the bilayer form a hexagonal array, and each half of the bilayer appears to be relatively independent of the other [20]. The author considered it important to establish that fully hydrated D¢PC assumed the expected lamellar phase. Therefore, samples of hydrated'DQPC and DPPC were applied to microscope slides and covered with cover slips to minimize dehydration of the samples. These samples were then examined under cross-polarized light with and without a calcite compensator using a polarizing microscope. Both samples of hydrated lipid were similar in appearance under cross-polarized light. The field was filled with squiggly bright and dark bands which alternated between turquoise and fuchsia when the calcite compensator was in place. The bright bands were 38 exchanged with the dark when the line being observed was exactly alligned with the cross-polarized field. This observation means that the sample displays an extinction angle of 0°. A number of "crystalline" forms possess an extinction angle of 00 under cross- polarized light, among them, infinite planer arrays of hexagonal symmetry. Since DPPC was known from other data to exist in the lamellar phase [21], and since the characteristics of the two samples under cross-polarized light were, for all extents and purposes, identical, this observation was considered to confirm that hydrated D¢PC existed in the lamellar phase. Determination ofggel to liquid-crystalline phase transition temperature for 00PC. The LB'lamellar phase undergoes an order- disorder transition which is generally known as the gel to liquid- crystalline phase transition. This transition is thought to occur when the thermal energy available to a molecule in the gel state of the LB'phase {in which all the hydrocarbon chains appear to be in the all-trans conformation} is sufficient to cause rotations about carbon-carbon bonds in the hydrocarbon chain which lead to the introduction of Kinks [22]. This gel to liquid- crystalline (L8' to L0) transformation occurs at a characteristic tem- perature for every lipid. The temperature at which this order-disorder transition takes place is a function primarily of the length of the hydrocarbon chain, but the charge in the region of the polar head group can also affect the onset temperature for the transition to a lesser extent [23]. 39 Since it had been established that hydrated unsonicated D¢PC existed in the lamellar phase, experiments were undertaken to determine the gel to liquid crystalline phase transition for DOPC. Samples of the hydrated lipid were scanned from -120° to +120°C using a Differential Thermal Analyzer. With the exception of the ice-water transition at 0°C no phase change was detected. Since it was possible that the gel to liquid-crystalline phase transition for D¢PC had been masked by the water-ice transition, the author performed the scanning measurements using (i) 2H20, (ii) 1:1(w/v) sucrose-water mixture and (iii) 1:1(w/w) ethylene glycol-water mixture as hydrating media for 00PC. The water-ice transition temperature was shifted and, in the case of the ethylene glycol- water mixture, the transition was shallow and took place over a wider temperature range. However, no phase transition was detected for the lipid. Examination of properties of lipid mixtures containinggmfig, When two lipids which carry the same polar head group but acyl substituents of differing chain length are mixed in a 1:1 molar ratio, certain changes in the shape and position of the heat absorption curves for the gel to liquid-crystalline phase transition are observed. In the case where the chains differ in length by two carbons (and no unsaturated site, nor branching is present) the lipids will form an "ideal" solution. The heat absorption curve for the gel to liquid-crystalline phase transition of the mixture (or solution) will not differ greatly in shape from that of either pure component, but only one transition curvelwill be observed in which 40 the position of the transition maximum will have been shifted to a temperature which is exactly intermediate to the transition temperatures of the pure components [24]. However, if straight chain saturated lipids carrying acyl substituents which differ in chain length by four or more carbons are mixed, the properties of the resultant lipid solutions indicate that the mixing is'hon- ideal. The shape and position of heat absorption curve for the phase transition of the lower "melting" lipid is relatively un- affected. However, the shape of the heat absorption curve for the higher "melting" lipid is greatly broadened while the position of the temperature maximum is shifted toward that of the lower "melting" lipid. Since 00PC and dipalmitoyl phosphatidylcholine (DPPC) carry acyl substituents of identical chain length (the four additional carbons of phytanic acid are methyl branches), the author had reason to expect that these two lipids might form an "ideal" solution. If such was the case, then the temperature shift in the transition maximum for the mixture would yield information about the undertermined phase transition temperature for DOPC. The effect of adding increasing amounts of 00PC to DPPC on the phase transition curve of the latter is shown in Figure 9. It is seen that the addition of 0090 to DPPC causes a broadening of the temperature region over which the order-disorder transition of DPPC takes place and a diminuition of the intensity of the curve. The onset temperature for the transition in DPPC is shifted to lower temperatures while the temperature at which the transition Mr.“ \“F Q?— 7 V— _ f V 9 V I J 20 °C 40 FIGURE 9 Differential thermograms of mixtures of diphytanoyl phosphatidyl- choline in dipalmitoyl phosphatidylcholine. Diphytanoyl phosphatidyl- choline: (a) 1 mole %; (b) 4 mole %; (c) 7 mole %; (d) 11 mole %; (e) 18 mole %; (f) 22 mole %; (g) 46 mole %. The sample consisted of 5 mg lipid in 5 ul of water. 42 terminates is only slightly affected. The temperature which is associated with the transition maximum is affected to an intermediate degree. At a concentration of DOPC in DPPC which approaches a 1:1 molar ratio, the temperature range over which the DPPC transition takes place has become so extended that the transition curve appears to almost disappear into the baseline. Apparently,the presence of DOPC in the DPPC matrix disrupts intermolecular order in DPPC and diminishes the size of the cooperative unit involved in the phase transition. {The breadth of the heat absorption curve for the gel to liquid-crystalline phase transition is generally inter- pretedto reflect the total number of molecules which simultaneously undergo the transition. This is called the cooperative unit. The larger the cooperative unit the narrower the heat absorption curve [25].} This experiment yielded no information about a gel to liquid-crystalline phase transition in DOPL. Therefore, the author concluded that DOPC possessed no detectable gel to liquid- crystalline phase transition. The lack of a gel to liquid-crystalline phase transition in DOPC becomes more comprehensible when data from experiments per- formed on the lipids extracted from the extreme halophiles are considered. The lipids extracted from the cell membrane of the extreme halophiles comprise a system similar to DOPC in that the polar fraction of these lipids consists of ether analogues of glyceride-derived phospholipids in which the hydrocarbon portion of the lipid is exclusively dihydrophytol. Low angle x-ray dif- fraction studies have been performed on hydrated suspensions of 43 these lipids [26]. In this paper the observation was made that in the radial densitometer tracings the 14 A band was weak indi- cating that the terminal methyl groups were not well localized at the center of the bilayer. In addition, the spacing of the wide angle diffuse ring was 4.9 R which was greater than the 4.5 to 4.6 R spacing observed for most natural source lipids. This observation provided evidence that the intermolecular chain spacing was greater than that encountered in other lipids. Moreover, the lack of sharp rings at wide angles suggested that the packing of the lipid molecules in the bilayer was non-crystalline. In another study the fluorescent probe, perylene, was dissolved in these halobacterial lipids [27]. The fluorescence yield of perylene in this system was interpreted as indicating that over the temperature range studied (15°C to 75°C) the halobacterial lipids exhibited a greater fluidity than that found for most other lipids with this probe. Furthermore, no gel to liquid-crystalline phase transition has been observed in the halobacterial lipids [28]. These evidences of disorder on a molecular scale in the hydrocarbon region of lipid bilayers prepared from halobacterial lipids can be considered in the light of a thermodynamic study by Patterson and coworkers [29]. These authors found that the short-range ordering forces between hydrocarbon molecules which leads to what Botherel and coworkers call "correlation of molecular orientations" [30] were highly sensitive to molecular shape. For pure n-Cn liquids these intermolecular forces increased with 44 increasing n. For heavily branched hydrocarbon compounds (e.g. 2,2,4,6,6-pentamethy1heptane and others) these short-range ordering forces were non-existent. The enthalpy change which is associated with the gel to liquid-crystalline phase transition involves, in part, the over- coming of the correlation of molecular orientations which exist in the gel state. If, therefore, the presence of methyl branches at regular intervals along the phytanic acyl chain has the effect of partially or wholly destroying the correlation of molecular orientations (as the x-ray data suggest), this could provide an explanation of the lack of a gel to liquid-crystalline phase transition in DOPC. Patterson and coworkers [29] also noted that when n-Cn hydrocarbons were mixed with methylated hydrocarbons, the short range order in the former was destroyed. This observation provides insight into the effect of mixing D¢Pc with DPPC on the gel to liquid phase transition of the latter. The data which were presented in Figure 9 suggest the partial phase diagram shown in Figure 10. If one chooses a sample compo- sition of 10 mole % of D¢PC in DPPC {Samples of this composition will be used in NMR experiments.}, at a temperature above termination temperature of the gel to liquid-crystalline phase transition for DPPC, the sample will containa homogeneous liquid crystalline phase. If the sample temperature is lowered to a temperature below the termination temperature for the gel to liquid-crystalline phase transition for DPPC, the sample will contain a liquid-crystalline 45 50 ] l U 01 l TEMPERATURE °c 8 1 25 - REGION 8 2K)-- 5;- 5 IS '- REGION C '0 I I I I 0 20 ‘40 60 80 100 MOLE °/o DIPALMITOYL PHOSPHATIDYLCHOLINE FIGURE 10 Partial phase diagram of mixtures of diphytanoyl phosphatidylcholine in dipalmitoyl phosphatidylcholine. Region A consists of a homo- geneous liquid-crystalline phase 10 mole % 00% and DPPC. Region 8 consists of a liquid-crystalline phase enriched with respect to diphytanoyl phosphatidylcholine in equilibrium with a gel phase enriched with respect to dipalmitoyl phosphatidylcholine. Region C consists of nearly pure gel phase dipalmitoyl phosphatidylcholine. 46 phase in equilibrium with a small amount of almost pure DPPC gel phase. As the temperature of the sample is further lowered, phase separation continues,and the liquid-crystalline phase becomes en— riched with D¢PC over and above 10 mole %. At some unknown temp- erature, presumably, phase separation is complete,and the sample contains a DPPC gel phase contaminated with DOPC in equilibrium with a liquid-crystalline DDPC phase contaminated with DPPC. This information will become important to the discussion of the detection of lateral phase separation by NMR. Determination of the order parameter for the hydrocarbon chain of DQPC. The state of motion of the hydrocarbon portion of lipid molecules in bilayer membranes can be ascertained by a number of physical methods {electron-spin resonance (ESR), 1H and 2H-nuclear magnetic resonance (NMR), and Raman spectroscopy}. A quantitative measure of the molecular order is given by the order parameter [31]. In motionally restricted systems such as lipid bilayers, it has been customary to describe the extent of order with respect to an effective laboratory fixed symmetry axis, a, usually refered to as the director (see Figure 11). The director is the axially symmetric axis for the slowest molecular motion which is still fast enough to give averaging over the time period of the measurement. In the model developed for the interpretation of the order parameter in terms of the motions of the lipid hydrocarbon chain [31], the order parameter is a measure of the distribution of orientations of an internuclear vector, 7, with respect to the director. 47 Director, 6 Instantaneous chain orientation , E Molecular interaction vector, 7 FIGURE” Illustration of vectors_and angles relevant to lipid hydrocarbon chain motional model. Htlis the direction of the applied magnetic field; d is the director which is normal to the lipid bilayer interface; the angles are described in the text. Although in the figure all the vectors appear co-planar, generally the relationship between the various vectors is three dimensional. {This figure is a partial reproduction of a figure from the Ph.D. thesis of Dr. N. 0. Petersen and is used here with the permission of the author.} 48 On the NMR timescale effective axial symmetry exists about the bilayer normal [3]]. However, there is evidence that molecules in these systems reorient about an average director normal to the bilayer [32, 33]. If this is the case, then it becomes necessary to identify the effective symmetry axis for motion in the system with the average chain orientation with respect to the director. Then the interpretation of the order parameter must take into account chain reorientation with respect to the bilayer normal in addition to the intramolecular motion of the vector F'with respect to the chain axis. {This motion is axially symmetric with respect to the instantaneous chain orientation.} In this model,which was developed by Chan and coworkers, [31, 34, 35] the distribution of B, the angle between the molecular vector F'and the director (see Figure 11), is governed by two types of motions (1) chain reorientation with respect to the bilayer normal and (2) intramolecular rotations in the chain. From Figure 11 it is apparent that chain reorientation results in changes in the angle a, whereas chain rotations result in changes in the angle 7. {These two motions are considered to be independent since the timescales of the motions are expected to be very dif- -ll -12 ferent (intramolecular chain rotations, 10 to 10 sec; '7 sec.) [31].} For proton NMR the molecular chain reorientation, 10 vector F'is taken to be the geminal methylene interproton vector which undergoes changes in the angle 7 with respect to the in- stantaneous chain axis due to chain rotations. For proton NMR, chain reorientation is considered to result from conformational 49 changes either in the glycerol backbone or about the ester linkages of the lipid. However, because of the steric restrictions of the neighboring chains, this tilting of the chain will be expected to be locally cooperative and to cover all angles within a certain range, i.e. Ad. According to the mathematical developement of Petersen and Chan [31], the order parameter associated with changes in 8, 58’ is a product of the chain order parameter, So, which depends only on the magnitude of the chain deflection, Ad, and the intramolecular order parameter, SY’ which depends on the population fraction of a given conformer and the angle which the geminal methylene interproton vector, F, makes with respect to the instantaneous chain direction, i.e. 2 Where Sa = 1/2 (cos 00 + cosAa), and SY = -3/8Pt-1/8 (Pt = probability of a given chain segment being in the trans conformation) [31]. 58 can be estimated experimentally in the following manner. A multilammellar suspension of.a lipid is subjected to an off-resonance Rf pulse of sufficient power to excite all proton spin transitions in the molecule. The free induction decay (FID) is recorded as the spins relax to their equilibrium populations. Since a multi- lamellar lipid suspension constitutes a powder sample, one is at the slow motion limit of the T2 curve where direct dipolar coupling is the most important contribution to the spin-spin relaxation. 50 Therefore, the expected T2 should be in the psec time domain. Hence, if the intensity of the F10 at very short times is plotted with respect to time, the effective T2 (T2*) can be determined for the hydrocarbon methylene protons in the multilamellar sample. Using the methodology devised by Seiter and Chan [34], T2* can then be interpreted in terms of the order parameter, S Figure 12 B' shows a photograph of a typical FID taken for DDPC multilayers, and a graphic representation of the data taken from this F10. The values of T2* determined for the acyl methylene protons of multilamellar DDPC ranged from 70-77 usec. This range in T2* corresponded to a range in the methylene order parameter, 58’ of -0.18 i 0.03. This range fell well within the range of reported literature values for this order parameter in similar systems, -0.17 i 0.04 [31]. It is necessary to make some assumptions in order to interpret the experimentally determined order parameter in terms of Sc and SY for DOPC. In the case of straight chain saturated lipids Petersen and Chan [31] have performed the calculations necessary to estimate the probability of the chain segments being in a trans or gauche conformation. Pt was found to fall in the range of 0.8 to 0.9 which these authors interpreted as signifying that the population of kinks in a straight acyl chain of a lipid above the gel to liquid-crystalline phase transition temperature was on the average one per chain. However, these calculations have not been performed for D¢PC due to the complex nature of the differing types of kinks possible in the phytanic acyl chain. |(DO INTENSITY FID 51 (a) (11) FIGURE 12 lllll1 1 I I I l I EE___J 100 300 500 700 900 H00 TIME (psec) (a) A representative FID taken at 57.4 MHz at 20°C. Total time elapsed is 500 psec, with each point corresponding to 0.5 usec. (b) A plot of FID intensity !§_time. 52 Nevertheless,it is plausible to expect that the population of kinks in the phytanic acyl chain will be higher than that found in the straight acyl chain of comparable chain length, i.e. palmitic acid, due to the fact that the ground state energy differences between the all trans and kinked forms in D¢PC are less (1 Kcal/mole) than the difference in a straight chain (2 Kcal/ mole). Therefore, the author assumed a range for P for DDPC t of 0.5 < Pt < 0.8. Figure 13 shows graphically the relationship between P and Au [31]. From the graph it can be seen that the t range in Au, the chain excursion angle, for DOPC would thus be predicted to be 40 < Ad < 50°. According to Petersen and Chan [31] if the chain excursion, or reorientation with respect to the bilayer normal, is treated as a normalized distribution of angles, then Au which represents the maximum angle of excursion, can be expressed in terms of a most probable angle do, which for DOPC would be 190 < 00 < 28°. The corresponding analysis performed for the straight cahin lipid DPPC yielded Au 2 55° and a0 = 30° [31]. The above analysis permits one to state that the angle associated with chain reorientation in D¢PC should be somewhat smaller than that observed for DPPC. This conclusion is reasonable in view of the larger cross sectional area occupied by the phytanic acyl chain and the expected lack of cooperativity in chain motions due to the predicted absence of intermolecular short-range ordering forces between chains [29]. This point will be important to the discussion of the spin-lattice relaxation results. 53 I I I I I 0 IO 20 3O 4O 50 60 7O 80 90 Au (degrees) FIGURE 13 The range of values of Pt, the probability of a trans orientation, for given values of Am, the limit of the chain reorientation, which yield a proton order parameter S = - 0.17 i 0.04 {This figure is a partial reproduction of a figuFU from the Ph.D. thesis of Dr. N. 0. Petersen and is used here with the permission of the author.} 54 Characterization of Sonicated Suspensions of DOPC. Confirmation of presence of and determination of average size of DOPC bilayer vesicles. Aqueous dispersions of phospholipids contain large planar fragments of the lipids in lamellar phase dispersed in excess water. If the multilamellar sample is diluted further and agitated mechanically, it forms liposomes, which are roughly spherically concentric bimolecular leaflets with water trapped in between. 0n irradiating a liposomal suspension with ultrasonic sound (sonication) for prolonged periods of time (i.e. > 15 min) the liposomes are broken down to produce vesicles. A vesicle is a small sphere of aqueous medium enclosed within a single bimolecular lipid wall. The limiting size of vesicles which can be produced by’bl hour of sonication appears to be m2203 [36]. Smaller vesicles are apparently not produced because of packing constraints upon the individual lipid molecules [36]. Since it had been established that DOPC produced the appropriate lamellar phase on hydration, multilamellar suspen- sions of DOPC were subjected to sonication and the resultant sample was examined via electron microscopy (EM). A representative electron micrograph of DOPC vesicles is shown in Figure 14. Many difficulties were experienced in preparation and interpretation of electron micrographs of DOPC vesicles. The author had chosen Kreb's Ringer solution [37] as the aqueous medium for DOPC vesicle suspensions. This selection had been made with an eye to the permeation experiments which the author wished to perform under conditions which approached physiological 55 Figure 14 A representative electron micrograph of D¢PC vesicles produced by continuous sonication in 2HzO-Ringer. 56 IOWA FIGURE 14 57 conditions. {As it evolved, this decision was pure serendipity.} It is customary, when transmission electron micrographs of lipid samples are to be taken, to mount the sample on carbon coated grids and to use negative staining to fix and render the lipid visible. The carbon coating for the grids is charged since it is generated by sputtering carbon on to a support such as mica or a monomolecular plastic film. Vesicles samples prepared from DPPC and egg yolk phosphatidy- choline in Ringer solution adhered to carbon grids (transferred from mica), but D¢PC vesicles would not. DOPC vesicles would adhere neither to freshly sputtered carbon films, carbon films which had been stored for some time, nor hydrated carbon films. This problem was suspected to have its origins in the presence of a net charge at the D¢PC vesicle surface. X-ray, neutron diffraction, and 3] P-NMR studies of the phosphocholine head group have been interpreted as indicating that this polar head group lies roughly parallel to the bilayer interface [38, 39, 40] and executes a rotation about the P-0 glycerol bond which, coupled with a wobble of the molecule as a whole, carries the phospho- choline group to all positions within a cone about the P-D glycerol bond [40]. Hence, it has been presumed that shielding of the phosphate by the quarternary ammonium in the phosphocholine morety is intermolecular rather than intramolecular. If the area per phosphatidylcholine molecule in a bilayer {a parameter which is generally conceded to be controlled by the area requirements of the fatty acid chains} becomes abnormally large, the intermolecular 58 shielding in the head group region would be expected to be inefficient, and the interface would appear to be charged. The author proposes that this is the case in DOPC. Monolayer studies of lipids which have been compressed to their limiting areas, called the close packing area,yielded values of this parameter for DPPC of 40-45 82 [41], for egg yolk phosphatidylcholine, 50 32 [42]. Since monolayer studies of fatty acids have given values such as 20.3 R2 for stearic acid {C18} [42] and 30.6 32 for oleic acid {C18’ 1 unsaturated bond} [42] for the zero pressure area, it is clear that the close packing area for phosphatidyl- cholines appears to be equal to the sums of the zero pressure areas for the fatty acid chains. A monolayer study of halobacterial lipids had yielded a value of 60 82 for the close packing area per lipid molecule in this system [27]. Since the bond to the glycerol "backbone" in the halobacterial lipids is an ether, not an ester linkage, and since the steric requirements of the ether ys_the ester linkage should permit more efficient packing of the dihydrophytol moiety in the halobacterial lipids than of the phytanic acyl moiety in DDPC, the value of 60 32 should be regarded as representing a lower limit of the close packing area for 00PC. In vesicles the area occupied per lipid molecule has been estimated for DPPC as 70-75 82 [35]. This value is 25-30 22 larger than the close packing area. If this difference in area per molecule is generalizable, then the area per molecule of DOPC in vesicles would be predicted to range from 85-90 32 (as a lower limit). Therefore, it is difficult to envision effective intermolecular 59 shielding of the phosphate moiety in DQPC since the area avilable per head group should be abnormally large in this lipid. It was found that carbon sputtered on to a plastic film of monomolecular thickness could be used as a support on the EM grids. This film could be more successfully hydrated than the carbon film thus reducing the charge. However, this film tended to retain Ringer solution which left salt deposits an evaporation. The salt crystals were very electron dense, and the resultant local heating in the electron beam caused the film to break down. Despite these experimental difficulties, satisfactory electron micrographs of D¢PC were obtained of samples prepared under all conditons of interest to the body of work which is yet to be discussed. These results are summarized in Table 1. However, the appearance of the electron micrographs of D¢PC vesicles was far from uniform. From the electron micrograph shown in Figure 14 it is apparent that the D¢PC vesicles are not homogeneous with respect to size. Moreover, some EM grids carried liposomal material in addition to vesicles. If electron micrographs were to be used to establish which species was typical of the sample, vesicles or liposomes, hundreds of grids would need to be examined in order to achieve a statistical sample. In view of the difficulty of preparing EM grids which could be expected to carry sample of D¢PC vesicles or liposomers, this was not considered to be a viable alternative. Determination of the heterogeneity of DOPC vesicle samples. The analytical ultracentrifuge has been extensively employed in the TABLE I. Average Size of Vesicles Produced by Continuous Sonication. 60 LIPID AQUEOUS MEDIA AVERAGE DIAMETER 0F VESICLES (R) D¢PC 5 mM La(Na3)3 280 - 360 20 mM La(Na3)3 300 - 400 100 mM La(Na3)3 200 - 250 +++ . .a La ”mock Ringer' 490 - 635 ZHZO-Ringer 220 - 330 DPPC ZHZO-Ringer 250 aLa+++"mock Ringer" consists of Ringer solution in which the divalent salts have been replaced with sufficient La(N0 give the appropriate ionic strength. 61 determination of the molecular weights of proteins. In addition it is an excellent non-destructive tool for examining vesicle suspensions to determine their homogeneity. The author wished to utilize this methodology to ascertain the amount of the multi- lamellar liposomal material which was artifact in the electron micro- graphs. {This was an important point because the liposomal material had been most consistently observed in vesicles samples which had been prepared by cycling sonication. Vesicle suspensions prepared in this manner had been used in the determination of the spin- lattice relaxation rates, and the presence of appreciable amounts of liposomal material could distort the values of these rates for the vesicles.} Dipalmitoyl phosphatidylcholine (DPPC) has been mentioned at intervals throughout the discussion. This lipid carries the same length chain as DDPC, and the polar head group is identical. Moreover, DPPC has been well characterized, and a large body of data exists describing this lipid. Therefore, the author will continue to make use of DPPC as a reference against which to compare and contrast the behavior of DOPC. Vesicle suspensions of DDPC and DPPC in H20 and 2H20 Ringer solution produced by cycling and continuous sonication were subjected to centrifugation in a Beckman Model E ultracentrifuge equipped with schieren optics. The results of two typical runs are shown in Figure 15. It is immediately apparent that the DOPC vesicle suspension is much more heterogeneous than the DPPC vesicle suspension. {The width of the sedimenting boundary which 62 Figure 15 Photographs of schlieren patterns from representative sedimentation velocity experiments: (a) D¢PC vesicles formed by continuous sonication in 2H.20-Ringer; (b) DPPC vesicles formed by continuous sonication in HZD-Ringer. Insert (a) Photograph of a schlieren pattern taken early in a sedimentation velocity experiment on DOPC vesicles formed by cycling sonication. Insert (b) Photograph of a schlieren pattern taken early in a sedimentation velocity ex- periment on DPPC vesicles formed by cycling sonication. Note the presence of a series of small "peaks" on the leading edge of the sedimentation boundary. Insert (c) A photograph of a schlieren pattern taken at a later time of the same DPPC vesicle sample. Note the appearance of a shoulder on the main vesicle peak. 63 .9 «home. 9 $50.". .a “hung. “‘1' Q“ . .u «home. any a: is translated into the breadth of the peak by the schlieren optics is related to the heterogeneity of the samples.} There is no evidence of liposomal material in either of the runs shown. {Liposomal material has a very different molecular weight from the vesicular material and would be expected to appear as a separate boundary, i.e. peak, at early times during the course of sedimen- tation.} However, these vesicle samples (Figure 15) were prepared by continuous sonication. The inserts in Figure 15 show early frames on an ultracentrifuge run on D¢PC and DPPC vesicle suspensions prepared by cycling sonication. The DPPC sample shows evidence of small amounts of liposomal material {the small peaks on the high molecular weight side of the sedimenting boundary}. In addition,the DPPC vesicle sample prepared in this manner is much more heterogeneous as evidenced by a shoulder which developes on the main peak (Figure 15, insert c). The D¢PC vesicle sample is more heterogeneous than that prepared by continuous sonication, but there is no clear-cut evidence for the presence of multi— lamellar liposomes. At long times for such D¢PC samples some material reaches the meniscus (2H20-Ringer) at a rate which is consistent with the presence either of large vesicles (> 750 A in diameter) or very small liposomes (500 A in diameter, 2 bilayer). But the percent of the mass of the vesicle sample tied up in such forms is < 10% and, at that level, is not expected to influence the values found for the spin-lattice relaxation rates. Determination of partial specific volume of DOPC in vesicles. When vesicle samples are centrifuged in media of varying density, 65 it is possible to obtain a value for the partial specific volume of the lipid. {The assumption must be made that the average size and size distribution of the vesicles as well as the hydration state of the lipid in the vesicles is unchanged in the different media. This should be a reasonable assumption in the case of 2 H20- and H20- Ringer solutions.} The mathematical treatment, based on the development given by Tanford [44], for the determination of the partial specific volume of the lipid is presented below. The total mass of a vesicle is given by the expression Mv = 4n[(Riz/Ai + Roz/AO)M/Nav + 1/3 Ri3 pSOIn]. Where Ri = inner radius of vesicle area occupied by one lipid molecule on inner radius outer radius of vesicle area/molecule an outer radius Ai Ro Ao M = molecular weight of lipid Nav = Avagadro 5 number 0501" = den51ty of aqueous medium Mv = mass of vesicle. The force on a sedimenting particle in a gravity field in a centrifuge is given by the expression _ 2 6(Rr) - Mm Rr' Where M = mass of the sedimenting particle w = angular velocity of the rotor Rr= the distance of the particle from the center of rotation. 66 Therefore, for a vesicle _ 2 2 GIRr) — 4"[(Ri lAi + Ro /Ao)M/Nav + 1/3 RRW501an The buoyant force on a sedimenting particle is expressed as _ 2 B(Rr) ' v psoln w Rr’ where v = volume of the sedimenting particle. {For a vesicle = 3 4/311Ro _ 3 3 3 ' 4/3W[Ri + (R0 ' R] )J-} Therefore, the buoyant force experienced by a sedimenting vesicle is given by _ 3 3 3 2 B(Rr) - 4/3"[R1 + (R0 --R1. )]pso.In w Rr' The net force experienced by a sedimenting particle in a centrifugal gravity field is then F(Rr) = w2 Rr(Mv - ), psoln which for a vesicle becomes = wZRrI4n[Ri2/A13 + R 02/A0 JM/Nav - 4/3flRi3p 3 4/3"[Ri + (R03 ' Ri 3:npsoln _ 2 - 2 - w Rr{4n[Ri lAi + Ro 2/A0]M/Nav - 4/3n(R03- R. 3)p soln soln}’ but the first term in the above expression is just the mass of the lipid shell of the vesicle, whereas, the expression 4/311(RO3 - Ri3) is the volume occupied by this lipid shell. The volume of the lipid shell can be expressed as 4/3n(R03 - R13) = V£(mass of the lipid shell) where 9% = partial specific volume of the lipid. When this expression is substituted into the equation for the net force experienced by a sedimenting vesicle in a centrifugal field, the equation becomes 67 F(Rr) = szr{4n[Ri2/Ai + Roz/AO)M/Nav] - (l - 92 psoln)} which resembles closely the form of this equation for proteins. The velocity of a sedimenting particle is given by the expression u = F/f. Where f = frictional coefficient. For the spherical vesicle f = 61mRo (n = viscosity of the suspending medium). Therefore, the velocity of a sedimenting vesicle is given by _ 2 2 2 - u - w Rr[4"(Ri /Ai + Ro /Ao)M/Nav]/6nnRO-(l - v£ p501”). The Svedberg equation for this system is 2 so = u/w Rr’ where 50 = sedimentation coefficient. Therefore, for a vesicle _ 2 2 - so - 4"(Ri lAi + Ro /Ao)M/Nav(l - v2 psoln)/6flnRo' Experimentally the sedimentation coefficient is obtained by plotting the logarithm of the distance of the sedimenting boundary from the center of rotation !§_the time the boundary has been undergoing sedimentation. From this graph a least squares slope is calculated and the apparent sedimentation rate is then determined from _ 2. l3 s - l/w (least squares slope) ‘10 (sec). The apparent sedimentation rates for D¢PC and DPPC vesicle suspensions in 2H20 and HZO-Ringer solution and egg yolk phosphatidylcholine in 2HZO-Ringer are listed in Table 2. It is interesting to note that the sedimentation rate of D¢PC vesicles was the same irrespective of whether the vesicles were formed by cycling or continuous 68 TABLE II. Apparent Sedimentation Rates of Bilayer Vesicles. LIPID AQUEOUS MEDIA APPARENT SEDIMENTATION RATE (Sec) o¢Pc 2H20-La+++mock -l16.7a Ringer ZHZO-Ringer -47.l4 i 1.20C HZO-Ringer -3.3i4a 2 . a DPPC HZO-Ringer -9.39 —l7.83e HZO-Ringer 13.34 i 2.49b EYPCd ZHZO-Ringer 42.54a afrom a single measurement; vesicles formed by continuous sonication. b . . . . average of two values; ve51cles formed by continuous sonication. C I 0 average of three measurements; veSicles formed by continuous and cycling sonication. degg yolk phosphatidylcholine. efrom a single measurement; vesicles formed by cycling sonication. 69 sonication. This, however, does not appear to be the case with DPPC. From the ratio of the apparent sedimentation rates of the D¢PC and DPPC vesicles in H20 and 2HZO-Ringer the partial specific volume for the two lipids can be calculated. The partial specific volume of D¢Pc in vesicles was found to be 1.013 ml/g at 25°C. The partial specific volume for DPPC in vesicles was calculated to be 0.950 ml/g at 25°C. The literature value of the partial specific volume of DPPC in vesicles which had been prepared in distilled, deionized water is 0.948 ml/g at 25°C [45]. The literature value and that determined from sedimentation rates for the partial specific volume of DPPC in vesicles at 25°C are very similar. The difference between the author's value and that quoted from the literature may reflect errors in density and viscosity or a failure to correct for non-Newtonian flow or both. The observation that D¢PC vesicles undergo flotation in both H20 and 2HZO-Ringer implies that the amount of lipid per unit volume in the lipid bilayer in D¢PC vesicles is less than that for other lipids which do not bear methyl groups on the hydrocarbon chain and, therefore, pack with greater efficiency. Examination of pmr spectral characteristics of D¢PC in various media, A series of NMR spectra were taken of D¢PC under various conditions for characterization purposes. The same series of spectra were also taken for DPPC for purposes of comparison with the D¢PC spectra. These two sets of NMR spectra are shown in Figure 16. 70 Figure l6d shows an NMR continuous wave (cw) frequency spectrum taken at 65°C and 220 MHz of a multilamellar suspension of D¢PC similar to the sample from which the order parameter was determined. The spectrum of the lipid consists of relatively sharp choline and hydrocarbon methyl resonances (Au'n350 Hz) superimposed on a broad methylene-methyne signal (Au $2000 Hz). {The equation which describes the relationship between line width, Au, and T2* is An = l/wT2*. The linewidths observed in the above cw spectrum thus corroborate the T2*'s obtained from the FID.} For comparison the corresponding spectrum of a DPPC multilamellar sample is shown in Figure l6h. It is noteworthy that the intensity of the resonances corresponding to the hydrocarbon chain terminal methyl and choline methyls are similar in DPPC multilayers, whereas the hydrocarbon chain methyl signal clearly dominates the DOPC multilamellar spectrum. This observation indicates that all or the majority of the methyl groups on the phytanic acyl chain are motionally free [34]. DOPC was further characterized with NMR spectra taken at 220 MHz for the lipid in an organic solvent and at lOO MHz for vesicular suspensions (see Figure l6a-c). For reference purposes spectra taken of DPPC under the same conditions are shown in Figure l6e-g. The frequency spectra of D¢PC and DPPC in deutero- chloroform together with their assignments are presented in Figure l6a and e. DQPC exhibits a series of overlapping doublets in the region of the methyl proton resonance which arise from the methyl branches on the diterpenoid hydrocarbon chain. This can 71 Figure l6 PMR spectra. (a) 220 MHz spectrum of diphytanoyl phosphatidyl- choline in CDClsat 20°C (sweep width 2500 Hz); (b) lOO MHz spectrum of small sonicated diphytanoyl phosphatidylcholine bilayer vesicles in 5 mM La(N03)3 at 55°C (swee width 1000 Hz); (c) lOO MHz spectrum of sonicated diphytanoyl phospRatidylcholine vesicles in 2HZO-Ringer at 55°C (sweep width 1000 Hz); (d) 220 MHz spectrum of diphytanoyl phosphatidylcholine multilayer suspension in 2Hzo-Ringer at 65°C (sweep width 20 KHz); (e) 220 MHz spectrum of dipalmitoyl phospha- tidylcholine in CDCla at 20°C (sweep width 2500 Hz); (f) lOO MHz spectrum of small sonicated dipalmitoyl phosphatidylcholine vesicles in 5 mM La(N03)3 at 55°C (sweep width lOOO Hz); (9) l00 MHz spectrum of sonicated dipalmitoyl phosphatidylcholine vesicles in 2Hzo-Ringer at 55°C (sweep width 1000 Hz); (h) 220 MHz spectrum of dipalmitoyl phosphatidylcholine multilayer suspension in 2HZO-Ringer at 65°C (sweep width 20 KHz). 72 ' CNS! TMS 0 II «343,3 N’(CHz){O-F"O"CH2 06 CHOLINE GYCEROL “*2 L x i O' O HOD 0. FIGURE lb 7 3 e W CH2 CH-O-ENWWW ' 11 -CH (CH3)3 N-(CHZWEE 2 BULK TMS -CH2'S cwouwe -N(CH3l3 CH2 ‘ C‘C‘Hz fl‘CHz CH . 43 [ HOD _ FIGURE lb 4E5. 74 be compared with the triplet which is observed for the terminal hydrocarbon methyl groups in DPPC. The chemical shift range covered by the methyl groups in D¢PC is necessarily larger than that found for the terminal methyl groups DPPC due to the range of chemical environments experienced by these methyls. This chemical shift range is approximately 0.ll ppm. The apparent linewidths observed for these hydrocarbon methyl signals in both lipids are governed partly by molecular aggregation into micelles and partly by long-range spin-spin coupling. The differences in chemical environments along the chain which produced the observed range of chemical shift values for the hydrocarbon methyl resonances in D()PC have a more pronounced effect on the spectral positions of the acyl methylene and methyne proton signals (see Figure 16a). The range of chemical shift values observed for the acyl methylene-methyne protons in DQPC extends fromitl.l4 ppm tO’b2.4 ppm (A61ul.26) in CDCl3. Due to complex spin-spin coupling and slight differences in chemical shifts,the homogeneous linewidth for the methylenes in diphytanoyl phosphatidylcholine cannot be ascertained. By comparison, all the hydrocarbon methylene protons in dipalmitoyl phosphatidylcholine with the exception of those a and B to the carbonyl have very nearly the same chemical shift (see Figure l6e). The resonances observed for the methyl protons associated with the choline head group are sharp in both DbPC and DPPC. The choline methyl resonance occurs at 6 = 3.35 ppm in D¢PC and 6 = 3.31 ppm in DPPC. The slight broadening of these resonances may be 75 attributable to micelles in the sample. The remaining signals in the spectra arise from the methylene and methyne groups associated with the choline and glycerol moieties. Fourier transform NMR spectra taken at l00 MHz at 55°C of sonicated suspensions of D¢PC and DPPC prepared in SmM La(N03)3 and 2HZO-Ringer are shown in Figure l6 b, c, f, g. In the 00PC vesicles the series of doublets observed for the acyl methyl protons in CDCl3 coalesce into one doublet at the temperature at which these spectra were taken. The doublet is a result of spin-spin coupling to the methyne proton (J = 5H2) as this splitting was found to be independent of the magnetic field strength. {At temp- eratures below 30°C the methyl signal broadens to the extent that the doublet is unresolved.} The linewidth for the methyl resonance is comparable to the chemical shift difference among various methyl signals observed in the spectrum of this lipid in C0013. In the DPPC vesicles the hydrocarbon methyl triplet seen in 00613 solution is broadened and poorly resolved. The sharp methylene and methyne proton resonances observed for 0¢PC in CDCl3 appear in sonicated suspensions as an envelope which spans the entire chemical shift range observed for these protons in CDCl3. No individual resonances are distinguishable. In fact, the lineshape for the methylene-methyne protons is distinctly asymmetric with an apparent line width (56-74 Hz at 100 MHz) largely determined by chemical shift dispersion. In contrast, for DPPC vesicles at temperatures exceeding its gel to 76 liquid-crystalline phase transition temperature, it is possible to distinguish the signal arising from the a methylene protons from that arising from the remaining hydrocarbon methylenes. Here the bulk of the methylene protons produce a sharp, reasonably symmetric, almost homogeneous line at 100 MHz (Au < 20 Hz). The choline methyl protons of both lipids give rise to sharp signals in vesicle suspensions (Au'v6 HZ at 100 MHz). The effect of sonication, thus, is to produce small spherical bilayer vesicles which give rise to very different proton NMR spectra from that observed for the multilamellar suspensions (compare Figure l6 a and d). The very broad lines observed for the hydrocarbon methylene protons in lamellar suspensions have their origins in the direct di- pole-dipole interactions between spins in the sample. The tumbling times of multilamellar fragments are much too slow (on the order of seconds) on the NMR time scale to cause any motional narrowing of the lines [45]. However, spatial averaging of the signals takes place due to the fact that the multilamellar fragments are disposed randomly in space. Spatial averaging would be expected to produce a power type spectrum from the multilamellar sample if the molecules in the lamellae were motionless [34]. The fact that a Gaussian line is observed indicates that there is molecular motion with respect to the bilayer normal [34]. This motion has been interpreted as arising from chain reorientations in the bilayer [34]. The motion on a molecular scale of the hydrocarbon chains in a phospholipid bilayer is both restricted and anisotropic. 77 The restriction of molecular motion in a phospholipid bilayer arises from the nature of the aggregate where the molecules are hexogonally packed in an "infinite" sheet held together by van der Haal's interactions. Therefore any "chain wagging" motion in one molecule is inevitably opposed to a greater or lesser extend due to the steric requirements of neighboring molecules. The anisotropy of motion derives from the fact that a single phospholipid molecule in solution {a purely theoretical situation, since phospholipids appear to possess an immeasurably low critical micelle concentration} would be slightly cigar shaped and, therefore, would possess two correllation times for motional narrowing of NMR resonances, one characteristic of molecular tumbling with respect to the long axis of the molecule, the other characteristic of molecular tumbling with respect to the short axis of the molecule. Confining the phospholipid to a lamella accentuates this anisotropy. The observed lines for the hydrocarbon protons in vesicles are reasonably sharp (Lorenzian) and appear to be homogeneous, i.e., completely motionally averaged. The question is, within the context of a restricted, anistropic system, what motions are responsible for the line narrowing? The tumbling time in vesicles is much shorter than that for multilamellar fragments, e.g. for a vesicle of 250 A to 350 R, «Jo's sec [45]. The timescale of this motion should be sufficiently fast to contribute to the motional narrowing of the vesicle proton signals in NMR. However, the linewidths of the hydrocarbon proton resonances have been repeatedly shown to be independent of the viscosity of the medium 78 in which the vesicles are produced [45]. This finding argues against vesicle tumbling as a controlling factor in the determ- ination of the hydrocarbon proton linewidths. Apparently as long as the vesicle tumbling is sufficiently rapid to provide an isotropic magnetic environment for the molecules in the vesicle, other factors control the observed hydrocarbon proton linewidth [31]. Another potential source of motional narrowing for the observed linewidths is lateral diffusion of the individual molecules on the vesicle surface. The lateral diffusion constant for lipidsin bilayer membranes has been found to be'm l0'8cm2 sec.1 [46]. However, the time required for a lipid molecule to undergo a translation through an angle of n/Z on the vesicle surface via lateral diffusion has been calculated to be on the order of l0'4 sec for a vesicle of 250 K - 350 X in diameter [45]. The timescale for this motion is expected to be too slow to contribute significantly to the motional narrowing of the hydrocarbon resonances [45]. Indeed, since lateral diffusion involves motion of the entire phospholipid molecule, the observation that certain resonances in the molecule are sharper than others (i.e. have a narrower linewidth) seems to preclude lateral dif- fusion as an important factor in motional narrowing [45]. The observation of differential broadening of resonances arising from certain protons in the phospholipid molecule implicates local and/or segmental motion as the mechanism for the observed motional narrowing of the hydrocarbon proton resonances in vesicles. Petersen and Chan [3l] compared proton and deuterium linewidths 79 in vesicles and, utilizing the predictions of their motional model for lipid bilayers, analyzed the effects of various vesicular and molecular motions upon predicted linewidths. They concluded that chain reorientation with respect to the bilayer normal was the dominant factor controlling the hydrocarbon proton linewidths in small bilayer vesicles (< 500 2 in diameter). They also presented evidence that within the context of their model this chain reorientation was more extensive in vesicles than in multilayers [31]. Chain reorientation with respect to the bilayer normal can be thought of as involving either an individual molecule and/or an ensemble of molecules (the cooperative unit). It is not possible at the present level of knowledge to distinguish the two processes. Petersen and Chan [31] point out that rotational diffusion might be invoked as a mechanism for chain reorientation in an individual molecule. Under this motional regime a larger chain excursion angle would have associated with it a longer timescale for the motion. If, however, the chain reorientation was principally cooperative, then the motion would appear to have associated with it a distribution of timescales with a fast and slow cut-off [31]. In this latter case the timescale of the chain reorientation would depend on the size of the cooperative unit. There is much evidence that the size of the cooperative unit is smaller in vesicles [47]. Under these circumstances, a larger chain excursion angle would appear to be associated with a faster timescale. 80 The development just presented for the identification of the motions of lipids vesicles which are responsible for the narrow lines observed in bilayer vesicles was based on data obtained from saturated straight chain lipids and lipids from natural sources. The same data is in many cases not available for DOPC, e.g. the rate of lateral diffusion. The vesicle tumbling rate 'fln~D¢PC vesicles will, of course, be unchanged. However, lateral diffusion rates in D¢PC may be somewhat greater than that observed for the straight chain lipids and lipids from natural sources because of the proposed lack of intermolecular ordering forces between phytanic acyl chains in D¢PC. {However, a rate increase in lateral diffusion of > 103 would be required to bring about line narrowing from this source. A rate increase of this magnitude seems very unlikely.} Chain reorientation is almost certainly not cooperative in D¢PC for the same reason mentioned above, and, therefore, the mechanism for chain reorientation in D¢PC in bilayers is probably rotation diffusion. This is in contrast to DPPC in bilayers where the chain reorientation is expected to be primarily cooperative. Hence the introduction of a surface of curvature in vesicles which is absent in multilamellar samples results in a larger chain excursion angle which would be expected to produce an apparent slowing of the chain reorientation in D¢PC in contrast to the apparent rate increase observed for DPPC [3l]. {The rotations of the chain associated with Kink diffusion in the hydrocarbon chain take place on a timescale which is too rapid to contribute significantly to the line narrowing and are, 81 therefore.neglected.} Notwithstanding the suggested differences in motion regimes between D¢PC and straight chain phospholipids in bilayer vesicles, the form of the equation which gives the dependence of the linewidth on the various motions is such that, for a chain excursion angle greater than «25°, the chain reorienta- tion is predicted to be the dominant motion responsible for line narrowing in vesicles [3]]. Since Aa in D¢PC multilayers was estimated to range from 40° to 50°, chain reorientation is almost certainly the dominant factor for line narrowing in D¢PC vesicles also. This brief analysis of D¢PC spectra as compared with DPPC spectra taken under various conditions is included to indicate that no unusual or singular feature is present in D¢PC NMR spectra. Some slight line broadening is encountered in the hydrocarbon methyl and methylene resonances in D¢PC vesicles over that found in DPPC which simply reflects an increase in the average size of the D¢PC vesicles. Otherwise the behavior of D¢PC as indicated by NMR is quite prosaic as presaged by the order parameter. Verification of the average size of DOPC vesicles. By means of NMR the choline methyl protons can be used as a probe into certain characteristics of the bilayer vesicle, amoung them: (i) the average size of the vesicles in a suspension and (ii) the integrity of the vesicles toward ion permeation. The technique usually involves the use of a.paramagnetic shift reagent to magnify the inherent asymmetry of environment experienced by the polar head groups located on the inner y§_the outer radius of the 82 vesicle. This asymmetry was first reported by Sheetz and Chan [45]. At 220 MHz in bilayer phosphatidylcholine vesicles averaging in diameter from 250-350 R a splitting of the choline methyl resonance was observed. The upfield less intense peak, which showed a separation in ppm from the hydrocarbon methyl resonance identical to that seen in multilamellar samples, was assigned to the choline methyl protons on the inner radius of the vesicle. The downfield more intense absorption which was slightly deshielded {0.04 ppm, the shielding of the phosphate in phosphatidylcholine by the quarternary ammonium of the choline moiety is intermolecular, and the larger are occupied by the polar head groups on the outer radius presumably causes the effectiveness of this shielding to diminish.} ‘was assigned to the choline methyl protons on the outer radius. It should be noted that although the average diameter of D¢PC vesicles in 2H20-Ringer solution ranged from 250-350 A, no splitting of the choline resonance was ever observed for this lipid in the absense of paramagnetic shift reagents. Arguments have been presented (vida supra) that the area available to the polar head group in D¢PC vesicles is abnormally large. If this were so, no asymmetry of environment would be expected to be experienced by the polar head group in the inner vs the outer radius of a DOPC vesicle, and no choline splitting would have been predicted in this system. Another prediction based on the argument that the polar head group is abnormally exposed in DOPC would be that the lipid should be sensitive to changes in the type and concetration of salts 83 present in the suspending medium. It has been seen that the average diameter of D¢PC vesicles is indeed sensitive to the ionic species in the suspending medium and its ionic strength. Another example of the salt sensitivity of this lipid was observed when the paramagnetic shift reagent, Eu(N03)3, was added to a vesicle suspension of D¢PC in 2H20-Ringer. The lipid precipitated im- mediately. The author attempted to add Eu(N03)3 to D¢PC bilayer vesicles in 2 HZO-Ringer as a solid and by l:l v/v dilution of the vesicle sample to bring the Eu+++ ion concentration to values ranging from 20 mM to 2.5 mM without changing the ionic strength, always with the same result, vesicle precipitation. The same manner of addition of Eu(N03)3 to vesicular suspensions in 5 mM, 20 mM, l00 mM La(N03)3 solution was attempted. Precipitation of the vesicles always resulted. A portion of the difficulty was discovered to be a result of changes in the pH of the solution which took place on mixing Eu+++ ion solutions with other salts. {Lanthanides hydrolize in aqueous solution [48].} If precipitation of the lanthanide as the hydroxide is to be avoided, the pH of the solution must be kept below a pH of 7.0-7.5 (depending on the lanthanide). However, it never became clear whether it was the lipid or the Eu+++ ion or both which responded to pH changes by precipitating. The author found that working at low pH 0»2.0) in 5 mM La(N03)3 solution, Eu+++ addition could be accomplished without precipitation of the vesicles if the vesicle suspension was maintained at a temperature of 55°C and not allowed to cool to room temperature. 84 The results of the experiment are illustrated in Figure l7 a and b. {The Eu(N03)3 must be added to the suspension of D¢PC vesicles in order to split the choline methyl resonance.} The choline methyl protons on the outer radius of the D¢PC vesicles were shifted upfield from those on the inner radius by the presence of the paramagnetic Eu+++ ion in the extravesicular medium. The areas under each choline methyl resonance were then determined, and the ratio of the outer choline methyl resonance area to that determined for the inner choline methyl resonance was calculated. A mathematical relationship exists between the ratio of the number of lipid molecules on the inner radius to the number on the outer radius and the radius of the vesicle. The choline outer/inner ratio should be an experimental measure of the ratio of the number of molecules on the outer !§_the inner radius. This experimental number can then be translated into a vesicle radius. The outer/ inner choline resonance ratio.for these D¢PC vesicles was found to range from 1.4 to 1.7 (7 measurements) which was consistent with the average diameter as obtained from electron micrographs, i.e. 290-360 R. The electron micrographs had indicated that the average vesicle diameter of D¢PC vesicles in 2 HZO-Ringer also ranged from 290-360 8. However, it was not possible to confirm this figure by'the use of Eu(N03)3 since under all experimental conditions the addition of Eum ion to Ringer caused vesicle precipitation. Another determination of the ratio of the outer/inner choline methyl resonance can be obtained by measuring the area of the a '3? FIGURE 17 *Jkk l 20.3 9 AH! A A l Time course of paramagnetically induced choline resonance splitting. The figure shows a series of spectra taken at 220 MHz of D¢PC vesicles in 5 mM La(N03)3 at 55°C. (a) standard; (b) immediately after addition of an equal volume of 5 mM Eu(N03)3; (c) 1 hr. 27 min. after addition; (d) 2 hr. 23 min.; (e) 23 hr. 4 min.; (f) 51 hr. 28 min.; (9) 74 hr. 47 min.. The paramagnetically induced separations of the inner and outer choline resonances are given on the figure. ' 86 choline methyl resonance in the absence and presence of the paramagnetic relaxation agent, MnClz, in the extravesicular solution. {The addition of Mn+++ ion to the extravesicular solution causes the outer choline methyl resonance to broaden to the extent that it appears to disappear into the baseline.} In this experiment the ratio of the internsities of the outer to the inner choline methyl signal was found to be «41.7. This result suggested that the vesicles were approximately 300 R in diameter which falls within the range of values determined from the electron micrographs. Subsequently, a sample of DPPC vesicles in 2HZO-Ringer solution was treated in the same manner. The outer/ inner choline methyl resonance ratio was determined to be m2.0 for these vesicles which is consistent with a vesicle diameter of ~250 X. This is the average diameter found for DPPC vesicles in 2HZO-Ringer solution in the electron micrographs. Investigation of integrity of D¢PC vesicles as measured by ion,permeation. When straight chain saturated phosphatidylcholines are sonicated at temperatures below the gel to liquid-crystalline phase transition temperature, the bilayer vesicles contain "crystalline defects" which permit the passage of ions, for example, Eu+++ [49]. These defects can be removed by heating these vesicles at temperatures which exceed the gel to liquid- crystalline phase transition temperature for the particular lipid for periods from 20 minutes to 1 hour, or avoided by sonicating the lipids at temperatures which exceed the gel to liquid-crystal- line phase transition temperature. Vesicles treated or produced 87 in this manner are said to be annealed [48]. The absence of the defects can be demonstrated by the addition of Eu+++ ion to the extravesicular solution. The separation of the choline resonances produced by the shift reagent is stable for long periods of time when no Eu+++ ion is permeating the lipid bilayer. However, if defects exist in the lipid bilayer, the initial separation of the choline resonances produced by the introduction of the extrave- sicular Eu+++ ion will decay rapidly with time as the inner choline methyl resonance is shifted toward the outer choline resonance by the Eu+++ ions which have permeated the bilayer. D¢PC possessed no detectible gel to liquid-crystalline phase transition. This observation provided no assurance that the vesicles formed from this lipid would not permit the passage of ions. Since the author proposed to perform permeation experiments utilizing vesicles prepared from DOPC, it was important to establish that D¢PC bilayer vesicles possessed no large scale defects which would render the vesicles freely permeable. Figure 17 presents the results of such a study. The veiscles were incubated at 55°C for 75 hours. During that time the choline methyl resonance separation changed by 4 Hz (0.02 ppm). This amount of change is the same as that which is observed in annealed DPPC vesicles and is regarded as arising from sample deterioration and errors in measurement of the peak separation rather than Eu+++ ion permeation. Therefore, the barrier characteristics of D¢PC bilayers toward ions were identical to that observed for DPPC bilayer vesicles prepared at temperatures.exceeding the gel to liquid-crystalline phase transition temperature. 88 In view of the effect which the addition of D¢PC to DPPC hydrated multilayers had on the gel to liquid-crystalline phase transition of the latter, it was considered of interest to de- termine whether the disruption of molecular order in DPPC apparent in the differential thermograms of the mixed lipids would result in the production of defects in bilayer vesicles generated from the mixed lipids which would allow passage of Eu+++ ions. There- fore, vesicles were produced from a mixture of 10 mole % DOPC in DPPC sonicated at temperatures greater than Tc for DPPC in 5 mM La(N03)3, and Eu+++ ion (2.5 mM final concentration) was added to the extravesicular solution. The separation of the choline methyl resonances was followed for 72 hours at 55°C with no significant change in separation observable over that time period. It was concluded that the introduction of D¢PC into the DPPC matrix caused no change in the barrier characteristics of the latter lipid toward ions. Interpretation of spin-lattice relaxation rates onQPC in bilayer vesicles in terms of kink diffusion and chain reorientation rates, The motional model for lipids presented earlier can be extended to aid in the interpretation of NMR spin-lattice relaxation measurements which have been undertaken to investigate the dynamic properties of lipids in bilayer model membrane systems [31]. Proton spin-lattice relaxation rates in lipid bilayer membranes have been found to decrease with increasing temperature and frequency of irradiation [50]. The spin-lattice relaxation rates of the hydrocarbon protons in D¢PC showed a similar temperature 89 and magnetic field dependence to that observed for other lipids. This combination of temperature and magnetic field dependence would occur if the relevant dipolar interactions were being modu- lated by at_1gg§t_two motions with different correlations times, 1" and “IS-L, such. that (4101“12 «1 and (no-ti)2 > 1. Then, in the presence of anisotropic motion the spin-lattice relaxation rate could be a approximated by 1n] 2 AT” + B/uiozT-L [50]. Under these conditions the first term dominates the observed temperature dependence, whereas the magnetic field dependence is contained in the second term. In is taken to be the correlation time associated with kink diffusion along the hydrocarbon chain, while tL is thought to be the correlation time connected with the reorientation of the hydrocarbon chain with respect to the bilayer normal. The behavior of this model for D¢PC is illustrated in Figure 18 a - d. For this model it is assumed that the observing frequency is intermediate to the rates for the two motions. The sum of the contributions to (l/Tl) from each correlation time with a particular activation energy represents the predicted experimentally observable (1/T1) curves (dotted lines, Figure 18). It should be noted that the correlation time with the highest activation energy dominates the temperature dependence of the observed (l/Tl) curve. 0n the other hand the frequency dependence observed experimentally should increase as the two correlation times approach one another. This point is illustrated by comparing 90 Figure 18 The predicted behavior of (1/T1) (dotted lines) at 100 and 220 MHz calculated according to the motional model where the observed spin- lattice relaxation rates arise from the summed contributions of two correlation times according to the equation l/T; = AI + B/wozt H J. .{For the sake of simplicity it has been assumed that A = 8.} For D¢PC: (a) Iu 10-3 sec, I = 5 x 10'8 sec, Ea for both motions .. -8 -l.5 Kcal/mole (b) III ="LID 8 sec, FL.= 5 x 10 sec, Ea for T“ = 2.0 Kcal/mol, Ea for T = 1.5 Kcal/mol; (c) T” = 1o'° sec, I = 5 x 10'8 sec, Ea for I - 2.5 Kcal/mole, Ea for I = 1.5 Kéal/mole; (d) I = I = 13-3 sec, Ea for I = 3.0 Kcél/mole Ea for Il-= 1.0 cal/mole. H For DPPC: (e) I = 10"8 sec, I = 10'10 sec, Ea for I = Ea I = 3.0 .1.1 II -11 _ -a II J. _ Kcal/mo e; (f) Ill = 10 sec, 74.- 10 sec, Ea for Ill - Ea for Iil= 3.0 Kcal/mole. 9l [UTIII - ' ' ' 'IOOMHZ E (d) l/T. __ 1/1'. _ (sec")‘ ......... 220" (sec‘) - 1 p I b I _J._ _L_ b l o I o 1 I 2 3 4 5 Io3/°K FIGURE 18 l/T, sec"- 92 ~— OOMH: “MHz 1 1 1 1 s 7 a 103/°1< OOMH: 220MHz / 1 '1' 1 1 1 1 4 3 a 9 IO /°K FIGURE 18 93 FigureslB e and f which show the theoretical curves for DPPC with those in Figure 18 a - d for D¢PC. The experimental results of the (l/T1) studies are summarized below. Methylene Spin-Lattice Relaxation Rates--The hydrocarbon chain methylene proton spin-lattice relaxation rates for D¢PC vesicles and for vesicles containing 10 mole % D¢PC in diperdeutero- palmitoyl phosphatidylcholine (DPPC-d62) matrix at various temper- atures and at two NMR frequencies (100 and 220 MHz) are shown in Figure 19 a and b. For comparison data obtained on DPPC under similar conditions is included [50]. In the pure D¢PC the spin- lattice relaxation rates (l/T1) for the methylene protons decrease with increasing temperature at both NMR frequencies. In the mixture at temperatures well below the gel to liquid-crystalline phase transition temperature (Tc) for the deuterated host lipid the spin-lattice relaxation rates resemble closely those found for the pure D¢PC at the same temperatures. This is especially evident at 220 MHz where the values are identical within experimental error. The relaxation rates in the mixture decrease with increasing temperature at both NMR frequencies until the To for the host lipid is reached at which point the plots of (1/T]) versus the inverse of the temperature show a discontinuity. At temperatures in excess of the phase transition temperature for the host lipid, at 100 MHz and to a lesser extent at 220 MHz, it was noted that the plots of ln[hz(m)-hz(I)] versus I(de1ay time) for the lipid mixture from which the (l/T])'s are determined, 94 Figure 19 Spin-lattice relaxation rates (1/11) of the hydrocarbon methylene protons in sonicated bilayer vesicles as a function of reciprocal temperature and at two nmr frequencies: (a) 100 MHz, (b) 200 MHz. G) diphytanoyl phosphatidylcholine; O 10 mole % diphytanoyl phosphatidylcholine in dipalmitoyl‘hosphatidylcholine-d62; E] dipalmitoyl phosphatidylcholine; 10% dipalmitoyl phosphatidyl- choline dispersed in 90% dipalmitoyl phosphatidylcholine-d62. 95 Om.m _ Oman, _ OVM _ 9 $50: omen OO.m _ OmN 3 md 0.. “gm non “we “we A-oomv .3 0.0_ 96 004..” _ 00AM _ OVM 1 0— $50.”. 0N.m _ A5 0.0 0.. 0.N on 3 oh .783 p: 0.0. 97 became non-linear. The residual protons on the acyl moiety of the deuterated host lipid were contributing to the observed peak height in the partially relaxed spectra, particularly for values around the "null point". To compensate for this the methylene peak areas from spectra taken at 100 MHz at identical temperatures for a sample of 10 mole % D¢PC in DPPC-d62 and for a sample con- taining an identical amount of DPPC-d62 alone were subtracted. From these data the corrected spin-lattice relaxation rates plotted in Figure 19a were obtained. This procedure proved to be unnecessary for the data taken at 220 MHz. At 100 MHz and at temperatures in excess of To for the host lipid the corrected spin-lattice relaxation rates for the hydrocarbon methylene protons of diphytanoyl phosphatidylcholine show no temperature dependence, while the spin-lattice relaxation rates for the hydrocarbon methylene protons at 220 MHz show only a very slight decrease with increasing temperature. The signal from the hydrocarbon methyne protons on the phytanic acid chain in the vesicle spectra is masked. Since it falls beneath the hydrocarbon methylene signal envelope, it is not possible to independently determine a spin-lattice relaxation rate for the methyne protons. In order to ensure that the Tl's determined for the hydrocarbon methylene protons were not influenced by the methyne relaxation, peak heights were sampled in the chemical shift region well removed from that characteristic of the hydrocarbon methyne protons. 98 Methyl Spin—Lattice Relaxation Rates--The spin-lattice relaxation rates for the hydrocarbon methyl protons in vesicle samples of pure pope and 10 mole % hope in the DPPC-d62 host lipid at various temperatures and two NMR frequencies (100 and 220 MHz) are summarized in Figure 20. It should be noted that all_the hydrocarbon methyl protons appear to undergo spin-lattice relaxation at the same rate. However, since the relaxation rate is rapid, the rate difference between two methyls must be greater than 20% of the measured l/T1 to permit observation of a relaxation rate difference exceeding the experimental error inherent in the measurement. The spin- lattice relaxation rates for the hydrocarbon methyl protons in the pure lipid are very similar to those found for the hydrocarbon methylene protons in this 1ipid. These relaxation rates decrease with increasing temperature at both NMR frequencies. In the mixture of D¢PC with DPPC-d62 at temperatures well below the phase transition temperature for the deuterated host lipid the longitudinal relaxation rates resemble closely those found for the pure lipid. At 220 MHz the rates for the pure lipid and the mixture are indistinguishable within experimental error in this temperature range. The relaxation rates for the mixture decrease with increasing temperature at both NMR frequencies until a temperature of roughly 50°C is reached. At this temperature the plots of (1/11) versus the inverse of the temperature show a dis- continuity at 100 MHz and 220 MHz. At higher temperatures there appears to be no temperature dependence at 100 MHz and only a very slight one at 220 MHz. 99 Figure 20 Spin-lattice relaxation rates (l/Tl) of the hydrocarbon methyl protons in sonicated bilayer vesicles as a function of reciprocal temperature and at two NMR frequencies (100 and 220 MHz). 0 diphytanoyl phosphatidylcholine; .' 10 mole % diphytanoyl phosphatidylcholine dispersed in dipalmitoyl phosphatidylcholine- d62. 100 Om.m _ 0N $50.... 00d 0mm 0mm 00.m 00.N _ _ _ _ _ \. | hi 0 I'll?! \3\ \ \IOIIIIO-I 0.0 0.N 0.m 0..» 0.0 . 78m. .3 0.0. 101 The correlation times for the motions which are thought to be contributing to the observed spin-lattice relaxation rates for D¢PC in bilayer vesicles can be estimated qualitatively by comparison of the temperature and frequency dependence of the (l/T1)'s in D¢PC and DPPC in bilayer vesicles. The timescales of chain re- orientation and Kink propagation which have been estimated for the 8 9 methylene protons in DPPC in bilayer vesicles are IloJO' to 10' "O to 10“] sec and T” ~10 sec [31]. The spin-lattice relaxation rates from which these correlation times were estimated are two to three times smaller than those measured for the hydrocarbon methylene protons in DOPC in bilayer vesicles (see Figure 19 a and b). This observation suggests that the rate of kink diffusion is slower in D¢PC than that detected for the melted straight chain phosphatidyl- choline of comparable chain length. In the discussion in the intro- duction, the barrier heights to chain rotations leading to the introduction of kinks in the phytanic acyl moiety in DOPC were predicted to be gt_1ga§t_4 Kcal/mole. Using the Arrhenius relation- ship, this activation energy for kink propagation yields a jump -10 -11 5 rate of 10 to 10 ec. This jump rate, by the formulation presented by Petersen and Chan [31], can be converted into a '8 to 10'9 sec. This correlation time for kink diffusion, T“ =10 rough approximation yields what appears to be a reasonable estimate of the observed I Considerably more frequency dependence was observed for the spin-lattice relaxation rates obtained from D¢PC in bilayer vesicles than those obtained from DPPC in the same system. This effect is 102 seen for this motional model when the time scales associated with chain reorientation and kink diffusion are approaching one another. The timescales of these motions in DPPC in bilayer vesicles have 2 to 103, been estimated to be separated in time by a factor of 10 but the extent of the frequency dependence observed for the spin- 1attice relaxation rates in DPPC suggests that the difference between the two correlation times is closer to 102 than to 103 (see Figure 18 e and f). Comparison of the amount of frequency dependence of the spin-lattice relaxtion rates predicted by the model for DPPC (see Figure 18 a - d) with the experimentally observed frequency dependence led the author to conclude that the correlation times associated with chain reorientation and Kink propagation in D¢PC were separated in time by approximately a factor of ten. Therefore, EL.for the hydrocarbon methylene protons of 7 sec to 10'8 sec. The D¢PC was estimated to range from 10' similarity of the order.parameters determined for the hydrocarbon methylene protons in D¢PC and DPPC multilayers indicated qualita- tively that the chain reorientation angle (Au) did not differ greatly in these two lipids in that system. It was predicted that Aa should be slightly smaller in the case of D¢PC (1<50° as opposed to ~55° for DPPC). If the mechanism for chain reorienta- tion is molecular in D¢PC but cooperative in DPPC, a similar chain excursion angle could have associated with it very different time- scales for the motion. It would be predicted that the correlation time for chain reorientation would be faster in D¢PC than in DPPC. Therefore, since IL for DPPC in multilamellar suspension has been 103 estimated to be 'le-7 sec [35], the author proposes that EL.f°r D¢PC in multilamellar suspensions is <10'7 sec. Sonication of the miltilamellar suspensions to produce bilayer vesicles has the effect in the case of DPPC of reducing the size of the cooperative unit which is thought to be associated with the chain reorientation in this lipid. According to the motional model of Petersen and Chan[31]. reduction of the size of the cooperative unit in vesicles is coupled with the appearance of a larger chain excursion angle and a faster apparent timescale of this motion (I for DPPC changes from 4:10"7 sec in multilayers to 10'8 sec to 10'9 sec in vesicles). Conversely,since chain reorientation is presumed to be molecular rather than cooperative in DOPC, sonication to produce bilayer vesicles should have the opposite effect on the correlation time associated with chain reorientation in DOPC. A larger chain excursion angle should produce a slower apparent timescale for this motion. However, the correlation time associated with chain reorientation in DOPC in bilayer vesicles was estimated above to be approximately 10'7 sec to 10'8 sec. Therefore, it appears that this slower motion in D¢PC is relatively insensitive to the geometry of the system. There is a notable lack of temperature dependence in both the linewidth and spin-lattice relaxation rates of the hydrocarbon methylene proton resonance in D¢PC bilayer vesicles when compared to the straight chain phosphatidylcholines. The temperature dependence of the methylene proton resonance linewidth is generally used to obtain estimates of the activation energy associated with 104 the slow chain reorientation with respect to the bilayer normal. 0n the other hand, the apparent temperature dependence of the spin-lattice relaxation rates {in systems where the rates for chain reorientation and kink propagation are very different} can be used to obtain an estimate of the energy of activiation associated with kink diffusion. However, in the situation where the rates of the two motions contributing to the observed spin- lattice relaxation rates are comparable, i.e. D¢PC, the temperature dependence of these rates becomes a complex function of the activation energy associated with both Ti.and III and not just T” alone. A plot of the apparent hydrocarbon methylene linewidth observed in D¢PC bilayer vesicles v§_the inverse of the absolute temperature yielded an apparent activation energy for EL.°f 1.5 to 2.0 Kcal/mole. This activation energy is much lower than that found by the same methodology for DPPC in bilayer vesicles, i.e. 8.0 - 9.0 Kcal/mole [51]. Perhaps this difference has its origins in the proposed cooperative mechanism in DPPC as opposed to the molecular mechanism for D¢PC for producing the chain reorientation motion in the two lipids. From literature values of barrier heights to internal rotation [51] the activation energy associated with kink propagation in D¢PC has been estimated to be at least 4 Kcal/mole. However, the apparent activation energy for this motion estimated from the temperature dependence of the spin-lattice relaxation rates was approximately 2 Kcal/mole. The 4 Kcal/mole activation energy for 105 kink propagation in D¢PC gave a timescale for the motion which appeared to be supported by the observation of a greater rate of spin-lattice relaxation for the hydrocarbon methylene protons in DOPC vs DPPC vesicles at comparable temperatures. However, examination of the behavior of the theoretical curves for D¢PC (Figure 18 a - d) makes it difficult to understand why the presence of a 4 Kcal/mole activation energy would not be readily visible in a strong temp- erature dependence of the spin-lattice relaxation rates. Unfortun- ately the amount of data on hand for D¢PC at this time is insufficient to resolve this paradox. The temperature and frequency dependence of the longitudinal relaxation rates of the hydrocarbon methyl protons in D¢PC bilayer vesicles closely resemble those of the hydrocarbon methylene protons. The motions which are expected to contribute to the relaxation of these protons are rapid rotation of the methyl "top" {the more rapid the rotation, the longer the T1} and restricted motions of the axis of the methyl "top" arising from conformational changes in the hydrocarbon chain and "wagging of the chain" [34]. The spin-lattice relaxation rates for these methyl protons are approxi- mately an order of magnitude faster than those of the terminal hydrocarbon methyl protons of straight chain lipids of comparable chain length. In straight-chain lecithin bilayer vesicles the (l/T])'s of the terminal methyl group are determined by the rate of internal rotation of the methyl “top", which is fast, and to a limited extent by the rate of motion of the "top" axis with respect to the chain axis [34]. From examination of space filling 106 models of D¢PC it appears that the spinning of the methyl groups along the phytanic acyl chain may be hindered by steric factors, especially where the methyl group is involved in a kink. In addition, the methyl groups on the phytanic acyl chain in D¢PC bilayer vesicles would be expected to undergo more motion with respect to the chain axis than in a straight chain lipid because they participate in the kink propagation motion in addition to chain reorientation. Combination of the above factors may account for the unusually rapid rate of spin-lattice relaxation observed for the hydrocarbon methyl protons in diphytanoyl phosphatidyl- choline vesicles. Detection of lateral phase separation in vesicles prepared from lipid mixtures containing DQPC. In this work the method of isotpic dilution was used to attempt to determine the inter- and intramolecular contributions to the spin-lattice relaxation rates of 00PC. The method consists of diluting a protonated lipid in a deuterated host lipid followed by a determination of the spin- lattice relaxation rates for the mixture of the lipids at various temperatures. If the protonated and deuterated host 1ipid are identical in structure with the exception of having replaced all hydrocarbon protons on the host 1ipid with deuterons, then isotope dilution can be used to determine the inter- and intramolecular contributions to the spin-lattice relaxation rates measured for the protonated lipid in its undiluted state [50]. If, on the other hand, the protonated and deuterated lipids are too structurally dissimilar to give "ideal" mixing this method could, in principle, 107 also be used to observe lateral phase separation, to investigate the nature of the phase diagram for the mixture of the two lipids, and to elucidate the nature of the lipid—lipid interactions in the mixture. The results obtained for a vesicle sample containing 10 mole % D¢Pc in DPPC-d62 are shown in Figure 19 a and b. The incomplete phase diagram presented in Figure 10 permits one to predict that when this vesicle sample is cooled to temperatures below the phase transition temperature for the deuterated host lipid (37°C),lateral phase separation should occur. At temperatures below the gel to liquid-crystalline phase transition for the deuterated host lipid, there should exist two phases in equilibrium in the vesicle bilayer, a gel phase which is enriched with respect to the higher melting lipid and a liquid-crystalline phase which is progressively enriched with respect to the D¢PC over and above the 10 mole % which was added to the deuterated host lipid as the temperature of the mixture is lowered. These predictions are borne out by our (l/T]) measurements. The plots of the temperature dependence of the spin-lattice relaxations rates of the hydrocarbon methylene protons of 10 mole % D¢PC in DPPC-d62 at 100 and 220 MHz shown in Figure 19 a and b reveal a discontinuity in both curves at approximately 37°C. This result suggests that lateral phase separation is taking place. This same conclusion is supported by the fourfold increase in the methylene proton signal linewidth which is observed for the D¢PC in DPPC-d62 on cooling from 60°C to 10°C (see Figure 21). This temperature range includes the gel 108 ICK) LINEWIDTH (Hz) ,0 1 1 1 1 1 1 .1 o 20 4o 60 T°C FIGURE 21 Observed PMR linewidths of the hydrocarbon chain methylene protons at 220 MHz. (0) small sonicated dipalmitoyl phosphatidylcholine vesicles (measurements below the thermal hase transition were obtained using 20 Khz sweep width); (El vesicles prepared by sonicating a mixture of 10 mole % diphytanoyl phosphatidylcholine in dipalmitoyl phosphatidylcholine-d62; and (.AS) sonicated diphy- tanoyl phosphatidylcholine vesicles as a function of temperature. 109 to liquid-crystalline phase transition temperature of the deuterated host lipid. The dependence of the hydrocarbon methylene proton signal linewidth on temperature is most marked in the region of Tc for DPPC-d62. This abrupt change in the hydrocarbon methylene proton linewidth is not observed over this temperature range for pure D¢PC (see Figure 21). For comparison the change in linewidth which has been observed for the dipalmitoyl phosphatidylcholine hydrocarbon methylene proton signal in vesicle suspensions over the same temperature range is also included in Figure 21. Here there is a fiftyfold increase in linewidth on cooling the vesicle sample through the DPPC phase transition [53]. These results taken together indicate that the temperature dependence of the methylene proton signal linewidth for DOPC in DPPC-d62 is a consequence of the gel to liquid-crystalline phase transition in the deuterated lipid. The resonances observed from the methylene protons of DCPC in DPPC-d62 at temperatrues above Tc for the host lipid are narrower than those observed in the pure lipid at comparable temperatures. This decrease in linewidth is due in part to changes in average size of vesicles in the two samples as established by electron microscopy. This represents the first reported instance where lateral phase separation has been demonstrated by the use of NMR techniques. Characterization of Permeability of DQPC Vesicles. Biological membranes and artificial model membranes have been demonstrated to have a relatively high permeability to water and neutral molecules [54, 55, 56, 57]. Explanations for the existence of this process of non-specific permeation, or passive diffusion, 110 have invoked such phenomena as channels and/or water-filled pores through the hydrocarbon region of the bilayer. These proposed mechanisms have not been supported by either experimental evidence or theoretical calculations [58,59]. A popular theory for passive diffusion regards the membrane as a homogeneous hydrocarbon phase into which the permeating species must dissolve [60]. This process is viewed as progressing via a series of successive jumps of the permeating species from one equilibrium position to another over various energy barriers. However, this theory is difficult to interpret on a molecular scale. It has recently been demonstrated that non-electrolyte diffusion in biological membranes resembles the diffusion encountered for similar species in soft polymers [57]. On the basis of this work and that of Peachhold and coworkers [61,62] which identified certain types of mobile structural defects in polymers (kinks), Trauble [4] proposed a detailed molecular mechanism for passive diffusion through lipid membranes. Trfiuble contended that the thermally- induced, mobile structural defects which had been detected in polymers should also be present in lipid bilayers at temperatures above the gel to liquid-crystalline phase transition temperature. These kinks were thought of as creating mobile pockets of "free volume" in the hydrocarbon portion of the bilayer which would act as carriers for the diffusing neutral molecule. At approximately the same time Demel and coworkers [63] found that the permeability of multilamellar liposomes toward glycol could be enhanced by the presence of unsaturated acyl substituents 111 on the lipids from which the liposomes were prepared. These authors invoked increased "fluidity" of the hydrocarbon region of the bilayer as the cause of the increased permeability. The term ”fluidity" used by the Dutch authors is probably ill-chosen and certainly ill-defined. A fluid or liquid has classically been viewed as being that condensed state which cannot sustain sheer forces. Fluidity on a molecular scale in bilayer lipid membranes probably involves contributions from all the motional modes available to the lipid in its liquid-crystalline state. Therefore, fluidity in model and natural membranes is probably best regarded as a weakening or loss of correlation of intermolecular interactions. This property is not suceptible to facile measurement, and hence is not a very experimentally useful concept. The author had intended to perform experiments with D¢PC vesicles which would serve to test Trauble's model for passive diffusion. A method devised by Lossen [64] in 1972 called sequential dialysis had been chosen for use in determining the permeability coefficients for the D¢PC vesicles. An apparatus, designed according to Lossen's description, was constructed, and experiments were commenced. The permeant was 3H-labelled glucose. Preliminary results indicated that the rate of escape of 3 H-labelled glucose at 25°C from D¢PC vesicles was slightly greater than three times that measured for DPPC vesicles. Unfortunately, in the author's hands the apparatus did not perform as expected. The probable source of the difficulty was diagnosed, but its correction would have involved considerable time and expense. Therefore, the author 112 and others involved with this aspect of the work concluded that these experiments should be abandoned. It is not clear that the permeability coefficients obtained for D¢PC vesicles would have provided any test for Trauble's model of passive diffusion through membranes. Kink propagation and population levels are insufficiently characterized for D¢PC at this juncture (in the author's opinion) to enable one to draw any conclusions as to the possible correlation between the kink diffusion rate, as estimated from the motional model which accounts for the observed hydrocarbon proton spin-lattice relaxation rates, and the permeability coefficient. It is the author's opinion that Trauble's model for passive diffusion of non-electrolytes through membranes can only be adequately tested, at this time, by measuring permeation rates for vesicles prepared from saturated straight chain lipids which have been well characterized physically and for which kink diffusion has been adequately modelled. In addition, it has been argued that permeation studies on bilayer vesicles do not provide data which can be considered biologically significant. This argument is based on the large differences in the surface of curvature encountered in bilayer vesicles (250 R to 350 R in diameter) as opposed to the average cell (0.5m to lu in diameter). This argument can, however, be adequately refuted by the observation that a very high surface of curvature is known to exist in such biologically important membranes as the neurotransmitter storage vesicles in the synaptic regions of neurons, the highly convoluted cristae of the mitochondrial 113 inner membrane, the edges of retinal rod outer segments and disks, and the brush borders of intestinal epithelial cells. In a recent 3‘? NMR study Chrzeszczyk and coworkers [36] set about by use of to elucidate the effect of packing constraints on molecular con- formations assumed by DPPC in highly curved systems. They inter- preted their data as indicating that although the bilayer thickness in small vesicles (220 X in diameter) was approximately the same as that found from x-ray data for multilamellar samples above the gel to liquid-crystalline phase transition temperature (~36-38 R), that the outer half of the bilayer appeared to be thicker (v20 2) than the inner half (N15 3). They hypothesized that the individual lipid molecules in each half of the bilayer occupy truncated prismatic cells. The area occupied by a polar head group on the outer radius was found to be N76 32 while the hydrocarbon chains of the same molecule were confined to an area of 51 32. On the other hand, a polar head group on the inner radius was restricted to an area which resembled that found for a molecule in a multi- lamellar sample, 68 22, whereas the base of this truncated pris- matic cell which contained the acyl chains measured ~94 32. The 20 3 thickness of the outer half of the bilayer would indicate that the acyl chains are approximately fully extended, but since these same chains are purported to be confined to an area of 51 82, this model would predict a very low population of kinks in the outer half of the bilayer. This is in stark contrast to the pre- dicted situation for the inner half of the bilayer where the hydrocarbon chains cannot be fully extended and, therefore, must 114 be concieved of as being bent (probably semi-permanently) and considerably kinked. This effect is, of course, only expected for highly curved surfaces. For instance, these unusual packing constraints should have practically disappeared in 500 X diameter vesicles. Notwithstanding, this model for molecular packing of lipid molecules in systems with high curvature has some very interesting biological and physical implications which should be investigated. EXPERIMENTAL II Materials. 1,2-diphytanoy1-§g-glycero-3-phosphocholine (diphytanoyl phosphatidylcholine) was synthesized followingaiprocedure set forth in a published synthesis [1] with slight modifications. l,2-dipalmitoyl-§n-glycero-3-phosphocholine (dipalmitoyl phosphatidyl- choline) was purchased from Calbiochem. 1,2-diperdeuteropalmitoy1- .gn-glycero-3-phosphocholine (dipalmitoyl phosphatidylcholine-d62) was prepared as described in a previous publication [50]. All lipids were purified by extensive silicic acid chromatography. Nonspecifically labelled glucose (50 mCi, 3H) was purchased from New England Nuclear Co. Salts utilized in making up Ringer solutions were reagent grade and were used without further purification. The water used in making up solutions and buffers was distilled and deionized. 2H20 was purchased from Stohler Iostope Co.. Sample Preparation. Samples for differential thermal analysis were prepared by dissolving weighed amounts of lipid (total sample size «6 mg) in a minimum of chloroform and transferring with a Hamilton syringe into capillary tubes. These tubes were evacuated at high vacuum overnight to remove solvent. 511] of deionized water was then 115 116 added to each sample and the capillary tube sealed. The sealed tubes were incubated for several hours at approximately 45°C. Unsonicated lecithin multilayers were prepared by addition of 0.4 ml 2HZO-Ringer solution to approximately 100 mg of lipid in an NMR tube. The NMR tube was then sealed and the sample agitated using a vortex mixer and allowed to stand overnight at 42°C. The samples were again agitated before spectra were taken. Vesicle suspensions of mixed diphytanoyl phosphatidylcholine and diperdeuteropalmitoyl phosphatidylcholine were prepared by dissolving weighed amounts of lipids (total lipid m 20 mg) in a minimum of chloroform in glass centrifuge tubes. The samples were evaporated nearly to dryness by placing the tubes in a sand bath at approximately 60°C and directing a stream of dry nitrogen into the tube. The tubes were then evacuated at high vacuum overnight. HZO- or 2HZO-Ringer (1 ml) was added to each tube, and the sample was either sonicated continuously at high power in an ice bath for 15 minutes or for the same period of time using an alternating 30 second on 30 second off cycle with a MSE 150 N ultrasonic disintegrator. Vesicle suspensions of pure lipid were prepared in the same fashion except that the chloroform step was omitted. ' Vesicle suspensions for the permeation experiments were prepared as described above with the exception that the lipid concentration was 50 mg/ml, and the HZO-Ringer solution contained 3 20 14Ci/ml of H-glucose. 117 For electron microscopy measurements vesicle suspensions prepared above were diluted to 1-5 mg/ml in lipid. A drop of this solution was then applied to a ZOO-mesh copper grid coated with parlodian onto which a thin film of carbon had been evaporated. After approxi- mately 30 seconds the excess liquid was blotted off, and a drop of 2% w/v phosphotugstic acid (pH 7.4) was applied to stain the sample. After approximately 40 sec, the excess staining solution was blotted off, and the grid allowed to dry before observation. Instrumentation and Methods. The differential thermal analyses were performed on the Du Pont Model 900 Differential Thermal Analyzer. Operation at sub-ambient temperatures was achieved by cooling the heating cell with a stream of dry nitrogen which had been precooled by passing through a liquid nitrogen bath. A rate of heating of 7°C/min was used in the recording of the thermograms. Hydrated multilamellar samples of diphytanoyl phosphatidylcholine were examined under cross-polarized light using a Zeiss Standard NL polarizing microscope. The proton NMR free induction decay of a multilamellar sample was recorded at 14.1 kG (MHz) courtesy of Dr. R. N. Vaughan using pulse NMR equipment described elsewhere [65]. 100 MHz proton NMR measurements were made using a Varian XL-100 system, and measurements at 220 MHz were performed on a Varian HR-220 spectrometer. Both spectrometers are equipped with a Fourier transform accessory and interfaced with a Varian 620i 16K computer. Probe temperature in both instruments was regulated 118 to i 1°C by a Varian 4540 variable temperature unit. The probe temperature was determined using either a methanol or an ethylene glycol standard sample together with temperature calibration charts which were prepared using a copper-constantan thermocouple. T1 measurements were made using the (I-I-n/2) inversion recovery sequence. Normally 36 transients were collected for vesicle suspensions of pure diphytanoyl phosphatidylcholine. One hundred transients were collected for diphytanoyl phosphatidyl- choline diluted in the diperdeuteropalmitoyl phosphatidylcholine host matrix. The spin-lattice relaxation values (l/T1) were calculated from the slope of a plot of In [h(w) - h(I)] versus I, where h(I) is the peak height of the partially relaxed resonance line. The EM grids prepared as described above were observed on a Phillips 201 C electron microscope operating at 60 kV at magnifi- cation level 10 (actual magnification on camera, 13, 900). Vesicle suspensions (1 mg/ml) prepared as described above were subjected to ultracentrifugation in a Beckman-Spinco Model E analytical ultracentrifuge equipped with schlieren, interference, and scanning UV optical systems. (Only the schlieren optics were used.) The experiments were performed at 25°C in an An-H rotor using double-sector cells with a 12 mm light path. The temperature was regulated by the RTIC unit. The velocity of the run was con- trolled via the Electronic Speed Control unit. The results of each run were recorded on Kodak Metallographic glass plates. 119 Vesicle samples prepared for permeability coefficient determination were placed inside the dialysis bag of the sequential dialysis apparatus [64], and dialysis was started. The temperature of the apparatus was regulated within i0.l°C using a Haupt circulating water bath. The dialysis was carried out in steps of 30 minutes (3 to 4) followed by steps of 5 minutes (10) to remove the trace molecules from the suspending medium and to determine the time constant for the dialysis bag. This procedure was followed by 10 to 15 dialysis steps of 60 minutes duration each. The dialysates were sampled individually, and 1.5 ml from each was pipetted into scintillation vials. These samples were mixed with 15 ml of Aquasol II (New England Nuclear) and, after allowing for dark adaptation of 10 minutes, were counted twice for l min. on a Beckmann Model LS 350 Scintillation Counter. A logarithmic plot of the counting rates v§_time gave a straight line. The least squares slope of this line was determined and the data analyzed according to the method outlined by Lossen [64]. PART III Reconstitution of Cytochrome g_0xidase in l,2-Diphytanoyl-§n-glycero-3-phosphocholine At this time relatively little is known about the nature of lipid-protein interactions. However, it is clear that many membrane- bound proteins are dependent on the presence of lipids for main- tenance of enzymatic activity. This dependence expresses itself most clearly in the fact that of those membrane-associated proteins which have been studied, i.e. the [Ca++ and Mg++] ATPase of the sarcoplasmic reticulum [662, rhodopsin [67], cytochrome g_oxidase [68], and bacteriorhodopsin [69], all have been found to contain residual lipid in the "delipified" form. These residual lipid molecules cannot be removed without irreversibly denaturing the enzyme. In addition, there is considerable evidence that membrane- bound proteins "immobilize" the lipid in their vicinity in the membrane. The fraction of immobilized lipid in the membrane has been primarily estimated via ESR spin-lable studies and found to be approximately 20-25% of the lipid in the membrane [70, 71]. '{These studies have been performed on natural [70, 71] and recon- stituted membranes [67].} Some membrane-bound enzymes have been shown to require the presence of a particular phospholipid head group for activity, for example,the galactosyl transferase of S, 120 121 tryhimurium requires phosphatidylethanolamine,and mitochondrial D-B-hydroxybutyrate dehydrogenase requires phosphatidylcholine [72, 73]. Cytochrome §_oxidase which is located in the inner mitochondrial membrane is the terminal enzyme in the oxidative phosphorylation chain. It catalyzes the four electron reduction of oxygen to two molecules of water. Enzymatic activity of delipidated cytochrome g_oxidase can be restored to a greater or lesser extent by adding lipid to the delipidated enzyme in the presence of detergent and dialyzing the mixture to produce bilayer lipid membranes in which the protein is imbedded [74]. Following this procedure, delipidated cytochrome g_oxidase was reconstituted in various synthetic phosphatidylcholines (in- cluding DOPC) and egg-yolk phosphatidylcholine. The level of activity of each preparation was assayed. The results of this study which was performed by Dr. Valevie Hu with the author's assistance with those experiments involving DOPC are summarized in Table 3. The activity of cytochrome g_oxidase in DOPC is less than that observed for the protein in dioleoyl phosphatidylcholine (DOPC), greater than that observed in dimyristoyl phosphatidylcholine (DMPC), and comparable to that observed for the protein in egg-yolk phos- phatidylcholine. However, neither DQPC nor DMPC are representative of naturally occurring lipids. The same is true of DPPC. The activity of cytochrome g_oxidase in DPPC membranes is considerably lower than the average activity in DMPC. This drop in activity 122 TABLE III. Activity of Cytochrome §_Oxidase Reconstituted with ““““" Various Lipids [Lipid/Protein = 20/1 w/w];a: b PHOSPHOLIPID ACTIVITY AT 30°Cc’ d Egg Yolk Phosphatidylcholine 5.6 Dioleoyl Phosphatidylcholine 8.0 Dimyristoyl Phosphatidylcholine 1.8e Diphytanoyl Phosphatidylcholine 4.0 Dipalmitoyl Phosphatidylcholine 37°C) to ensure that the lipid bilayer contains no crystalline defects.} The author views the comparable activities achieved with cytochrome g_oxidase reconstituted in DOPC and egg-yolk phosphatidylcholine as an encouraging sign that DOPC and other phospholipids containing the phytanic acyl group might prove to be versatile lipids for the investigation of lipid-protein interactions. A lipid containing the phytanic acyl moiety can be obtained pure and handled experi- mentally without the precautions which must be observed with natural source lipids, yet, as in the instance just cited, it appears that it may be possible to achieve reconstitution of enzymatic activity which is similar to that produced by the presence of a naturally occurring lipid with the same polar head group. EXPERIMENTAL III The following experiments are taken in part from the Ph.D. thesis of Dr. Valerie Hu with the author's permission. They are included here for ready reference. Materials. l,2-diphytanoy1-§n-glycero-3-phosphocholine was prepared as described previously. l,2-dipalmitoyl-§g;glycero-3-phosphocholine, l,2-dimyristoyl-§g-glycero-3-phosphocholine, l,2-dioleoyl-sn- glycero-3-phosphocholine were purchased from Calbiochem and used without further purification. Egg-yolk phosphatidylcholine was purified essentially by the method of Lea and coworkers [76] followed by extensive preparative layer chromatography and stored at 0°C under N2 atmosphere. The delipidated cyctochrome g_oxidase was supplied by Dr. Tsoo E. King [68]. The sodium deoxycholate, Tris were purchased from Sigma. The cytochrome c was purchased from Calbiochem. Sample Preparation. The lipid-deoxycholate mixtures for reconstitution were prepared either by adding the lipid (20 mg/ml) to 2% sodium deoxycholate in 10 mM Tris (pH 7.59) and sonicating for 1 hour in a bath sonicator or by taking up the lipid (20 mg/ml) in buffer 124 125 (10 mM Tris, pH 7.59) and sonicating for l/2 hour on low power in an ice bath followed by addition of 2% sodium deoxycholate in 10 mM Tris (pH 7.59). Delipidated cyctochrome c_oxidase (MN 'b 160,000) was added to the lipid-cholate mixtures at a concentration of 20:1 w/w lipid to protein. The protein-1ipid-cholate mixtures (total vol. m 0.5 ml) were incubated for l to 3 hours on ice. The mixtures were then dialyzed at 8°C against 10 mM Tris (pH 7.59) according to the following scheme. Dialysis # Dilution Factor Time (hrs), 1 1:600 7 2 1:800 13 3 1:400 5 4 1:400 6 5 1:800 13 Instrumentation and Methods. The dialyzed samples were assayed at 30°C (temperature regulated to : o.1°C via a Lauda K-2/R circulating bath) on a Biological Oxygen Monitor (Yellow Springs Instrument Co.) equipped with a Model 53 oxygen electrode. Oxygen uptake was recorded on a Bausch and Lamb x-y recorder. Assay conditions were as follows: to 1.3 ml of 50 mM phosphate buffer (pH 7.4) was added 100 ul of the dialyzed reconstituted cyctochrome g_oxidase (M 2.3 mM of hemea) and 20 ul of cytochrome c (v 17.7 mmol). The electrode 126 was inserted taking care to remove all air bubbles, and the reaction was initiated by the addition of 70 ul of 0.6 M ascorbate 4 mM in EDTA. The controls consisted of assays conducted on delipidated cytochrome g_oxidase under identical conditions. SUMMARY AND CONCLUSIONS The introduction of an acyclic diterpenoid hydrocarbon chain into a phosphatidylcholine engenders major modifications to the lipid's behavior in comparison with straight chain phosphatidyl- cholines in aqueous systems. These modifications undoubtedly arise from the steric effects of the methyl groups at the branch points along the hydrocarbon chain. There is a loss of the gel to liquid-crystalline phase transition which is characteristic of the straight chain synthetic phosphatidylcholines and phosphatidy- cholines extracted from biological sources. This loss is probably attributable to the elimination of intermolecular ordering forces between the phytanic acyl methylene groups on neighboring chains brought on by the introduction of the methyl branch points along the chain. The finding that the partial specific volume of diphytanoyl phosphatidylcholine in bilayer vesicles exceeds 1.0 ml/g further corroborates the prediction that the presence of methyl branches on the hydrocarbon chain is expressed in a loss of intermolecular packing efficiency. Despite the apparent disorder on a molecular scale in the hydrocarbon region of the diphytanoyl phosphatidylcholine bilayer the order parameter found for this lipid falls within the range of values obtained from other liquid-crystalline systems. 127 128 In the branched lipid, the rate of the fast motion which is contributing to spin-lattice relaxation of the hydrocarbon protons in 8 9 sec") than that 1‘ sec", [50]). bilayer vesicles is somewhat slower (10 to 10 found for dipalmitoyl phosphatidylcholine (1010 to 10 According to the motional model for lipids in hydrated systems proposed by Chan and coworkers [31, 34, 35] this fast motion is normally attributed to Kink diffusion in the hydrocarbon chain. Howeven,the near total lack of temperature dependence for this fast motion in diphytanoyl phosphatidylcholine makes this motional interpretation dubious for diphytanoyl phosphatidylcholine. The rate of chain reorientation in diphytanoyl phosphatidylcholine in bilayer vesicles appears to be comparable to that estimated 7 sec-1). for dipalmitoyl phosphatidylcholine in multilayers (i.e.'v 10 The addition of diphytanoyl phosphatidylcholine to dipalmitoyl phosphatidylcholine in hydrated multilamellar samples greatly reduces the observed intensity of the gel to liquid-crystalline phase transition of the latter lipid. This broadening effect probably arises from loss of cooperativity in the phase transition brought on by the disruptive effect of the methylated chain of diphytanoyl phosphatidylcholine on the ”correlation of molecular orientations" in dipalmitoyl phosphatidylcholine. Lateral phase separation in mixtures of diphytanoyl phosphatidylcholine in diperdeuteropalmitoyl phosphatidylcholine was detected via changes in the linewidth of the hydrocarbon methylene proton resonance of diphytanoyl phosphatidylcholine and by dis- continuities in the spin-lattice relaxation rates of these same protons over the temperature range which includes the gel to liquid-crystalline phase transition temperature for the deuterated 1ipid. The lack of intermolecular ordering between phytanic acyl moieties does not have any apparent effect on the permeability of diphytanoyl phosphatidylcholine vesicles to ions. Indeed, addition of diphytanoyl phosphatidylcholine to dipalmitoyl phosphatidyl- choline did not affect the permeability of this lipid toward paramagnetic ions used to monitor their vesicular integrity. This indicates that diphytanoyl phosphatidylcholine does form a bilayer which is impermeable to ions. {Preliminary experiments using glucose as a permeant indicate that the rate of escape of neutral molecules from diphytanoyl phosphatidylcholine vesicles is approximately three times that measured for annealed dipalmitoyl phosphatidylcholine vesicles at temperatures below the gel to liquid-crystalline phase transition for this lipid.} The stability of vesicles formed from diphytanoyl phosphatidylcholine with respect to fusion and multilayer formation, as well as the size of vesicles formed on sonication, is extremely sensitive to the ionic strength and the ionic species present in the vesicle suspension media. This effect probably reflects the unusually large area available to the polar head group in this lipid which diminishes or eliminates the normal intermolecular shielding experienced in the polar head group region of the phosphatidylcholine bilayer. 130 The information obtained concerning the effect of chain branching on bilayer structure and dynamical properties brought about by inclusion of the phytanic acyl moiety may be useful in understanding the role of phytanic acid in Refsum's syndrome. He would predict, as did Cushley and Forest [74] that the presence of phytanic acid in nerve membranes would cause large scale disruptions of membrane order. However, in order to understand more fully the effect of the presence of phytanic acid in the nerve membrane on the con- duction of nerve impulses, some comprehension of its effect on the permeability of the membrane would be necessary. This study of the motional and dynamic properties of diphytanoyl phosphatidycholine in hydrated bilayer systems may also prove useful in understanding the physiological function of the branched chain lipids found in the extreme halobacteria. One member of this genus of bacteria, Halobacterium halobium, produces the "purple membrane." Current interest in reconstitution studies on bacteriorhodopsin [69] as well as the utility of the reconstituted halobacterial membrane as a model system for examination of the validity of Mitchell's chemiosmotic hypothesis point up the value of detailed physicochemical data on a lipid which is structurally similar to the halobacterial lipids. Indeed, should it be found that the presence of the branched chain diether analogues of glycerol lipids are necessary to the proper functioning of this "purple membrane," then information which permits comprehension of the probable reasons for the maintenance of function by the branched chain 1ipid analogues would be useful. 131 The results of the cytochrome g_oxidase reconstitution experiments utilizing diphytanoyl phosphatidylcholine, though preliminary, indicate that diphytanoyl phosphatidylcholine, which is a saturated lipid, mimics lipids from natural sources in restoring enzymatic activity. Thus the substitution of the phytanic acyl group for saturated straight chain acyl groups in synthetic lipids could conceivably provide a very useful and easily manipulatable series of synthetic lipids for the study of protein-lipid interactions. BIBLIOGRAPHY —-l 0 014-500 10. 11. 12. 13. 14. 15. 16. 17. BIBLIOGRAPHY H. R. Redwood, F. R. Pfeiffer, J. A. Heisbach, and T. E. Thompson, Biochim. Biophys. Acta, 233, l (1971). Dr. Valerie Hu, personal communication. Dr. Nils O. Petersen, unpublished results. H. Trfiuble, J. Membr. Biol., 4, 193 (1971). M. Kates, L. S. Yen oyan, and P. S. Sastry, Biochim. Biophys. Acta, 98, 252 (1965]. M. Kates, B. Palameta, C. N. Joo, D. J. Kushner, and N. E. Gibbons, Biochem., g, 4092 (1966). D. Oesterhelt, "Membrane Proteins in Transport and Phosphorylation", Elsevier, Amsterdam, 1974, p. 79. E. Racker and N. Stoeckenius, J. Biol. Chem., 249, 662 (1974). S. C. Kushwaha, E. L. Pugh, J. K. G. Kramer, and M. Kates, Biochim. Biophys. Acta, 260, 492 (1972). H. C. Brown, E. J. Mead, and B. C. Subba Rao, J. Amer. Chem. Soc., 11, 6209 (1955). Dr. F. H. Dahlquist, personal communication. 0. M. Brown, Adv. in Org. Chem., 3, 75 (1963). J. C. Dittmer and R. L. Lester, J. Lipid Res., 5, 126 (1964). N. Kornblum, N. J. Jones, and G. J. Anderson, J. Amer. Chem. §92,, 81, 4113 (1959). R. Berry, 0. Frfilich, P. Lauger, M. Montal, Biochem. Biophys. Acta, 324, 323 (1975). R. Fettiplace, D. M. Andrews, and D. A. Haydon, J. Membr. Biol., ,5, 277 (1971 ). S. H. White, Biophys J.,‘L§, 95 (1975). 132 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 133 C. H. Huang, Biochem., 8, 344 (1969). V. Luzzati in "Biological Membranes," D. Chapman, Ed., Academic Press, London, 1968, p. 81. R. G. Kornberg and H. M. McConnell, Biochem., L9’ 1111 (1971). D. Chapman, R. M. Williams, and B. D. Ladbrooke, Chem. Phys. Lipids, 1, 445 (1967). W. Pechhold, Kolloid Z., 228, 1 (1968). H. Trfiuble, M. Teubner, P. Woolley and H. Eibl, Biophys. Chem., 4, 319 (1976). Dr. Nils 0. Petersen, unpublished results. H.-J. Hinz and J. M. Sturteuant, J. Biol. Chem., 2&1, 3697 (1972). J. S. Chen, P. 6. Barton, 0. Brown and M. Kates, Biochim. Biophys. Acta, 352, 202 (1974). A. E. Blaurock, J. M01. Biol., 2%, 139 (1975). J. K. Lanyi, Biochim. Biophys. Acta, 326, 245 (1974). V. T. Lam, P. Picker, D. Patterson and P. Tancrede, J. Chem. Soc. Faraday Trans. 11., 70, 1465 (1974). P. Bothorel, J. Colloid Sci., 27, 529 (1968). N. O. Petersen and S. 1. Chan, Biochem., 16, 2657 (1977). . A. Hemminga and H. J. C. Berendsen, J. Magn. Res., 8, 133 (1972). . H. A. Seiter and S. I. Chan, J. Amer. Chem. Soc., 25, 7541 (1973). M A. Seelig and J. Seelig, Biochem.,.Lé, 1585 (1975). C G. W. Feigenson and S. 1. Chan, J. Amer. Chem. Soc.,‘gg, 1312 (1974). > . Chrzeszczyk, A. Wishnia, and C. S. Springer, Jr., Biochim. Biophys. Acta, 470, 161 (1977). A. H. Krebs and K. Henselert, Z. phys. Chem., 210, 33 (1932). W. Lesslauer, J. E. Cain, and J. K. Blasie, Proc. Natnl. Acad. Sci. 0.5., 99, 1499 (1972). J. Seelig, H.-U. Gally, and R. Wohlgemuth, Biochim. Biophys. Acta, 467, 109 (1977). 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 134 S. J. Kohler and M. P. Klein, Biochem., 48’ 519 (1977). M. C. Phillips and D. Chapman, Biochim. Biophys. Acta, 163, 301 (1968). M. C. Phillips, B. D. Ladbrooke, and D. Chapman, Biochim. Biophys. Acta,,LgQ, 35 (1970). L. Salem, J. of Chem. Phys., 37, 2100 (1962). C. Tanford, "Physical Chemistry of Macromolecules," John Wiley and Sons, New York, N. Y., 1961. M. P. Sheetz and S. I. Chan, Biochem., 11, 4573 (1972). A. C. McLaughlin in "International Conference on Magnetic Resonance in Biological SystemsJ'7th, St. Jovite, Canada, 1976. P. W. M. Van Dijck, B. de Kruijff, P. A. M. M. Aarts, A. J. Verkleij and J. de Gier, Biochim. Biophys. Acta, 586, 183 (1978). Dr. C. A. Springer, Jr., personal communication. R. Lawaczeck, M. Kainosho, J.-L. Giradet, and S. I. Chan, Nature (London), 256, 584 (1975). P. A. Kroon, M. Kainosho, and S. I. Chan, Biochim. Biophys. Acta, 433, 282 (1976). Dr. P. A. Kroon, unpublished data. A. Abe, R. L. Jernigan and P. J. Flory, J. Amer. Chem. Soc., 88, 631 (1966). D. Lichtenberg, N. O. Petersen, J.-L. Girardet, M. Kainosho, P. A. Kroon, C. H. A. Seiter, G. W. Feigenson and S. I. Chan, Biochim. Biophys. Acta, 382, 10 (1975). R. C. Bean, W. C. Shepherd, and H. Chan, J. Gen. Physiol., 52, 495 (1968). . Finkelstein and A. Cass, J. Gen. Physiol., 52, 145 (1968). T. Hanai and D. A. Haydon, J. Theoret,gBiol., I], 370 (1966). W. R. Lieb and W. D. Stein, Nature (London), 224, 240 (1969). H H > . 0. Price and T. E. Thompson, J. Molec. Biol., 4], 443 (1969). . R. Hirsch, Currents in Mol. Biol., 1, 139 (1967). 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 135 B. J. Zwolinski, H. Eyring, and C. E. Reese, J. Phys. Coll. Chem., 53, 1426 (1949). S. Blasenbrey and W. Pechhold, Rheol. Acta, 6, 174 (1967). w. Pechhold, Kolloid 2., 3gp, 1 (1968). R. A. Demel, W. S. M. Geurts van Kessel, and L. L. M. van Deenen, Biochim. Biophys. Acta, 266, 26 (1972). 0. Lossen, Biochim. Biophys. Acta, 282, 31 (1972). R. W. Vaughn, 0. D. Elleman, L. M. Stacey, W.-K. Rhim, and J. W. Lee, Rev. Sci. Instrum., 43, 1356 (1972). G. 8. Warren, P. A. Toon, N. J. M. Birdsall, A. G. Lee and J. C. Metcalfe, Proc. Natl. Acad. Sci. U.S., Z], 622 (1974). K. Hong and W. L. Hubbell, Proc. Nat. Acad. Sci. U.S., 62, 2617 (1972). C. Yu, L. Yu, and 1500 E. King, J. Biol. Chem., 259, 1383 (1975). D. Wildenauer and H. G. Khorana, Biochim. Biophys. Acta,,vgg, 315 (1977). H. M. McConnell, K. L. Wright, and B. G. McFarland, Biochem. Biophys. Res. Commun., 41, 273 (1972). A. Stier and E. Sackmann, Biochim. Biophys. Acta, 311, 400 (1973). L. Rothfield and D. Romero, Bacteriol. Rev., 35, 14 (1971). N. C. Nielsen and S. Fleischer, J. Biol. Chem., 248, 2549 (1973). Dr. Valerie Hu, unpublished results. R. J. Cushleg and B. J. Forrest, Can. J. of Chem., 55, 220 (1977). C. H. Lea, D. N. Rhodes, and R. D. Stoll, Biochem. J., 60, 353 (1955). M. H. Silk and H. H. Hahn, Biochem. J., 56, 406 (1954). APPENDIX APPENDIX Unsuccessful Synthetic and Purification Steps Purification Attempts. 3,7(R), 11(R), l5-Tetramethy1hexdec-Z-ene-l-ol (phytol). The author's first attempts to purify the commercially available phytol (E.M. Merck, Darmstadt) were performed on high-pressure- 1iquid-chromatographic (HPLC) columns. The first absorbent used was Chlorasil. Phytol injected on a Chlorasil column and eluted with methylene chloride at 600 psi appeared as one peak followed by a broad shoulder. The author then injected the phytol on a larger silicic acid column which was eluted with chloroform at 100 psi. The elution pattern on this column was a large peak phytol followed by eight minor peaks. The author felt that though the silicic acid column would affect purification of the crude phytol, its use was impractical for the preparation of the quantities needed 0» 50 9, column loading maximum 35-50 mg). The author then attempted to purify the crude phytol via a large scale silica gel column (Baker, 60-200 mesh). Phytol (29.7 gm, 0.10 mol) was applied to 300 gm of silica gel and eluted with 1000 m1 of hexane-ether (1:1, v/v). Some highly colored material was removed but no purification of the phytol was observed (re- covered 24.1 gm, 0.08 mol) as checked via TLC (developed hexane- ether (1:1, v/v); revelatory agent, iodine). 136 137 3(R,S), 7(R), 11(R), lE-Tetramethylhexadecan-l-ol (dihydrophytol). Dihydrophytol (24.1 g, 0.08 mol) was applied to m 250 g of silica gel (Baker, 60-200 mesh) and the column was eluted with hexane-ether (1:1, v/v). The product was recovered (21.2 g, 0.07 mol) essentially unchanged as verified by TLC (conditions: those used above for phytol). The author attempted to remove the high molecular weight colored component (suspected to be composed of tetrapyrrole fragments) via acid wash. Dihydrophytol (21.2 g, 0.07 mol) was taken up in 250 ml of hexane. The hexane phase was washed with 5% hydrochloric acid (2x 250 ml), 2.5% sodium bicarbonate (2x 250 ml), and water (250 ml). The hexane phase was dried over anhydrous sodium sulfate. The solvent was removed on the rotary evaporator, and the recovery of the dihydrophytol was essentially quantitative. However, its chemical composition was unchanged as verified by TLC. 3 (R,S),7(R), 11(R),15-Tetramethylhexadecanoic acid (phytanic acid), The author attempted to purify phytanic acid with the use of column chromatography. The various procedures used are described below. Phytanic acid (0.33 g, 1.0 mmol) was applied to 10 gm of Sephadex LH-20, and the column was eluted with chloroform. Fractions of 3 ml each were checked via TLC (condition same as for phytol and "dihy- drophytol) for the presence of the acid. Fractions #13-15 were found to contain all the crude acid which had been unchanged in com- position by passage through the column. Phytanic acid (0.36 g, 1.2 mmol) was applied to 10 gm of activity III neutral alumina (Woelmn), and the column was eluted with 2% acetic acid in methanol. Fractions of 3 ml each were 138 collected, and the appearance of the acid was followed via TLC. Twelve fractions were collected, and the acid first appeared in fraction 5. Fractions 5-7 were found to be most contaminated with fractions 8-12 showing some purification as followed by IR and NMR. Recovered acid yield was not recorded. A reverse-phase column was prepared according to a published procedure [77]. The sample was applied to the column by adding 35 mg (1.2 mmol) of crude phytanic acid to 1 ml of paraffin oil. This mixture was dissolved in 15 m1 of ether, and 1.3 g of silated Celite was added. The solvent was removed under a heat lamp with stirring, and the resultant mull was heated overnight at 60° in a vacuum oven (3mm). The mull was taken up in 50% acetone in water saturated with paraffin oil and applied to the column. The column was eluted with acetone-water mixtures saturated with paraffin oil (SO-90% acetone). No acid was recovered. ' Methyl-3(R,S), 7(R), 11(R),15-Tetrmmethylhexadecanoate (Methyl phytanoate). Crude methyl phytanoate (160 mg, 0.5 mmol) was applied to a 6'x 1/4", 1% SE-30 column at 190°C and elution was achieved at a flow rate of 65 ml/min. of helium. The major peak, presumed to be the ester, was collected (~100 mg). The "ester" fraction was subjected to analysis on the GC-mass spectro- meter using a 6' x 1/4", 1% SP-2100 glass column (mass spec. ionizing potential 70 e.u.) and was found to contain the ester (m/e 326[M+], 101[100]) and three other components (m.e 197 [M+], 57 [100] possibly C14H30; m/e 312 [M+], 88[100] consistent with branched C19H37C02CH3 a - Me ester; m/e intensity to 544, region 139 200-350 very complex, 88[100] possibly tetrapyrrole degradation product). Crude methyl phytanoate (50 mg, 1.5 mmol) was applied to a 20 x 20 cm TLC plate of silica gel G which was developed in hexane- ether (1:1, v/v). The fractions were eluted with ether and the composition of each fraction was Checked via analytical TLC. The fraction which had the same Rf as the major component of the crude ester was then subjected to analysis via GC-mass spectrometry (conditions as above) and was found to contain methyl phytanoate (m/e same as above) and two of the three contaminants above (m/e 197[M+]; 312[M+]). Synthetic Attempts. B-D-Glucopyranosyl (1,2-diphytanoyl-§g-glycerol)-2,3,4,6- tetraacetate. l,2-diphytanoyl-§g-glycerol (6.8 g, 0.010 mol) was dissolved in 10-20 ml of dry chloroform. Silver oxide (2.55 g, 0.011 mol) and small amounts of activated anhydrous CuSO4 were added to this solution in the dark. Then (0.5 9) I2 was added, and the solution was stirred in the dark while acetobromoglucose (5.5 g, 0.012 mol) in 15 ml of dry chloroform was added dropwise over a period of m 1 hour. The reaction mixture was stirred in the dark for 24 hours. The reaction mixture was then filtered and the solid residue was washed with chloroform. The filtrate was concentrated on a rotary evaporator. The residue was applied to a Florisil column which was eluted with petroleum ether, a petroleum ether-ethyl ether mixture (4:1, v/v), petroleum 140 ether-ether-acetone mixtures (1:1:0.05; 4:1:1, v/v) and acetone. The principal product was judged to be 1,3 diphytanoyl-sn-glycerol on the basis of NMR and IR (virtually identical to the l,2-diphy- tanoyl-sn;glycerol) and behavior on TLC (in 65:24z4 (v/v) chloroform- methanol-water, Rf of l,2-diphytanoyl-§fl-glycerol, 0.35-0.42; product Rf, 0.80-0.85). Uncharacterized Solid. Sodium bicarbonate (509) was dissolved in 300 ml of water to which L-(-)-lysine (14.6 g, 100 mol) was added while the mixture was vigorously stirred. Then benzychloro- formate (50 m1,'v 200 mol) was added dropwise with stirring over a period of 1/2 hour. After the addition was complete, the reaction mixture was stirred for one hour. The reaction mixture was then extracted with 60 ml of ether to remove the excess benzylchloro- formate, and acidified with 3N hydrochloric acid. The reaction mixture, which had been a heavy slurry, solidified and became intractable. It was insoluble in every common solvent, except a slight solubility was observed in water-dimethyl-sulfoxide mixtures. It was assumed that this solid contained the desired product. Reduction of Uncharacterized Solid. The uncharacterized solid (7.0 g) from the previous reaction was added to a solution of l M diborane in tetrahydrofuran (Aldrich, 20 ml) at 0° under a stream of dry N2 gas. When the reaction mixture began to foam vigorously '{The uncharacterized solid had been dried for m 48 hours at 0.01 mm Hg.}, it was assumed that the carbonate was decomposing in the presence of the diborane, and this reaction sequence was abandoned.