Structure of polymerizable lipid bilayers VII: lateral organization of diacetylenic phosphatidylcholines with short proximal acyi chains

ELSEVIER

BiochimicaL et Biophysics Acta Biochimica et Biophysics Acta 1215 (1994) 237-244

Structure of polymerizable lipid bilayers VII: lateral organization of diacetylenic phosphatidylcholines with short proximal acyl chains David G. Rhodes a,* $I, S.W. Hui b, Y.H. Xu b, H.-S. Byun ‘, M. Singh ‘, R. Bittman ’ aPharmaceutics Division, College of Pharmacy, University of Texas at Austin, Austin, 7X 78712-1074, USA b Department of Biophysics, Roswell Park Cancer Institute, Buffalo, NY 14263, USA ’ Department of Chemistry and Biochemistry, CUNY Queens College, Flushing, h’Y 11367-1597, USA

Received 15 March 1994; revised 11 July 1994

Abstract As an extension of recent results (Rhodes, Xu and Bittman (1992) Biochim. Biophys. Acta 1128, 93; Hui, Xu and Bittman (1992) Langmuir 8, 2724) with a Cl8 diacetylenic phosphatidylcholine, bilayers of 1,2-bis(pentacosa-4,6-diynoyl)-sn-glycero-3-phosphocholine (C25) were investigated using X-ray diffraction on multibilayers and electron diffraction on Langmuir-Blodgett deposited bilayers. Monolayers of this lipid form solid (gel phase) domains at T > 14 mN/m. Electron diffraction data indicated that the chain spacing in these domains is 4.25 A and that the chains are tilted at angle of m 35” relative to the bilayer plane. Wide angle data frtm X-ray diffraction experiments indicated a similar spacing and chain tilt. Small angle data showed that the lamellar repeat was 70 A at high humidity and < 60 A at low humidity. The bilayer electron density profiles indicated a bilayer structure with no interdigitation. High angle reflections indicate thtt the principal acyl chain repeat is preserved as a function of hydration but some rearrangement occurs for other reflections. The - 10 A reflection corresponding to the headgroup spacing previously observed with ClS-diacetylenic phosphatidylcholine bilayers was not observed. The results are interpreted in terms of a packing model and possible limitations or constraints to the polymerization process. Keywords:

Diacetylenic lipid; Polymerizable lipid; X-ray diffraction; Electron diffraction; Langmuir-Blodgett film

1. Introduction Diacetylenic phosphocholines (DAPCs) often undergo polymerization when exposed to ultraviolet or more energetic radiation, provided that certain, largely steric, criteria are met. These molecules have been observed to form assemblies of unusual morphology. Several potential applications of these systems have been proposed, which would take advantage of the robust nature of the polymeric bilayers, the size and shape of the assemblies, or simply the unique phase behavior of these systems. Unfortunately, not all desirable properties are found concurrently. For

* Corresponding author. Fax: + 1 (512) 4717474. E-mail: [email protected]. * Most of D.G.R.‘s research was performed at the Biomolecular Structure Analysis Center at the University of Connecticut Health Center, Farmington, CT, USA. 0005-2760/94/$07.00 0 1994 Elsevier Science B.V. All rights reserved SsDIOOO5-2760(94)00136-7

example, tubules formed from DC,,,PC ’ do not polymerize well. Mixtures of DC,,,PC and dinonanoyl PC (DNPC) polymerize well, but do not form tubules. This laboratory and others have worked to determine (a) what criteria must be satisfied in a DAPC system in order to promote polymer formation, (b) what criteria must be satisfied (what forces are involved) to promote assembly of tubules or other morphologies, and (c> whether these conditions are compatible. The chemistry of simpler diacetylenes (RI-C = C-C = C-R,, where R, and R, are small organic substituents and often identical) has been well characterized [l]. For polymer to form, the monomers must be in a well-ordered solid, with the distance between adjacent diacetylenes and

z In the diacetylenic lipid nomenclature DC,,,PC, m and n refer to the number of CH, between the glycerol backbone and the diacetylene (proximal) and between the diacetylene and the terminal methyl group (distal), respectively.

238

D.G. Rhodes et al. /Biochimica

et Biophysics Acta 11715(1994) 237--244

the angle of the diacetylenes relative to the polymerization axis, constrained within relatively narrow limits. In addition, there is generally a small increase (8% in the case of 1,6-di(N-carbazolyl)-2,4-hexadiyne, DCH> or decrease (4% in the case of 6-(~-toluenesulfonyl-2,4-hexadiyne, PTS) in the lattice dimension in the direction of the polymer formation, so the packing of the monomers must allow for some movement [2]. Because of these conflicting, often con~adictory, criteria, it is not obvious which systems will polymerize most efficiently. Polymer formation is accompanied by a color change, usually to orange, red, or blue, with blue polymer (resulting from longer conjugation length) representing a high degree of polymerization and good order. Although this is a useful indicator, it cannot be used quantitatively because red color could arise from short polymer or disorder. DC,,,PC, like most DAPCs [3], does not polymerize completely in the pure form, regardless of the overall morphology, yielding only the red polymer form. In the presence of ‘spacer lipids’, DNPC or DMPC, the extent of polymerization is increased many fold, and the polymer is blue. As with all other diacetylenic lipids, polymer formation occurs only at temperatures below the main chain phase transition temperature, T,. If blue DAPC/spacer mixtures are heated to T > T, there is an irreversible thermochromic transition to red or orange [4]. It is unlikely that this represents breaking of covalent bonds, and probably represents decreased conjugation length by ‘twisting’ of the diacetylene chain. Glutamate-based diacetylenic lipids undergo a similar thermochromic (or solvochromic) transition, but the transition is fully reversible [5,6]. A recent publication described a C,, DAPC, DC2,itrPC, which exhibited rapid, complete polymerization and a bilayer thickness comparable to that of biomembrane lipids [7]. This lipid rapidly polymerized to N 85% completion in the blue form, and underwent a thermochromic transition similar to that of the DC,,,PC/DNPC mixtures. It was also reported that Langmuir compression isotherms of this lipid featured an overshoot ‘bump’ in the first compression, and that these films contained apparently solid lipid in feather-like domains which tended to twist in a counterclockwise direction. Based on structures and polymerization behavior of this lipid and similar DAPCs, the relatively poor polymerization of long chain DAPCs may arise from their inability to adapt to changes to the lattice which result from polymer formation 181. It has been suggested that DC&PC formed a two-dimensional lattice with a high degree of order in the headgroup, backbone, and proximal acyl chain region which was compatible with polymer formation. We propose that the relatively short proximal acyl chains allow for conformational change because they would not form extended solid domains as would longer acyl chains. It was also noted that the bilayer thickness was similar to that of biological membranes. Thus, receptors or other integral membrane proteins might not lose functionality upon in-

H

c L

OC(CH,)&S

C-CE32(CH2),7CH3

i OC(CH&CC=

C-

C=:

C(CHp),,CH3

+

OPO~H~CH~N(C~~)~ 0-

Dc2,17pc Scheme 1. DC2,, , PC structure

corporation into this system, and robust biosensors might be produced. In the present study, lipids were synthesized in which the distal acyl chains were 7 CH, segments longer. This derivative, DC 2,17PC (Scheme 11, would be expected to have a greater tendency toward gel-state distal acyl chains, and should have a somewhat different (hopefully improved) polymerization behavior. The rationale of constraining the diacetylene of DCZ,rOPC by increasing the length of the distal acyl chain to 18 carbon atoms was to address the issue of high crystallinity vs. ~exibility in obtaining an optimal balance for diacetylene polymerization. The data reported here support the model proposed previously [4,6,8] to explain the mechanism by which certain DAPCs polymerize well while others polymerize with difficulty or not at all. We suggest that systems which provide diacetylene alignment in the proximal acyl chain segment but allow for conformational ‘relaxation’ in the distal acyl chain segment should polymerize rapidly and completely.

2. Materials and methods

DC,,,,PC was synthesized by a modification of a procedure used to prepare DCr,,PC [8]. 4,6-Dodecapentadiynoic acid was obtained by oxidation of 4,6-dodecapentadiyn-l-01 with pyridinium dichromate. The diacylation of glycero-sn-phosphoryl choline-cadmium chloride complex with 4,6_dodecapentadiynoic acid was performed as described previously [8]. DC,,,,PC was stored in the dark, desiccated at -20°C. Solvents were HPLC grade, purchased from Aldrich. Water was from an ultrafiltration system (Bamstead RGW-5). Polymerization and TLC visualization utilized a 254-nm light (UVG-11, rated at 1.8 . 1015 photons/s per cm2 at 76 mm).

D.G. Rhodes et al./Biochimica

et Biophysics Acta 1215 (1994) 237-244

239

240 pg. Temperature scans were from 20°C to 100°C (at ZC/min) to allow for the expected higher T,. TLC was carried out using standard methods with Merck silica gel 60 (20 X 20 cm, 0.2 mm) plates. Chromatograms were developed with CHCl~/methanol/20% methylamine/H,O (60:36:10:0.3). To visualize the separated components, dried plates were sprayed with 1 mM TNS (in 50 mM Tris-HCl, pH 7.4) and examined under ultraviolet light (254 nm). Phosphate content of spots removed from the plates by scraping was determined as described previously [91. 2.3. ~i~ere~ti~l scanning calorimetry DSC experiments were performed with a DuPont DSC 2910 differential scanning calorimeter with a TA2000 PC-based analysis system as described previously [8]. All sample preparation protocols were identical except that 30 ~1 of an 8 mg/ml suspension of phospholipid was used for multilayer experiments so that the sample mass was

2.4. X-ray diffraction As described previously [6,8,10,111, these experiments used multibilayer stacks formed on aluminum foil substrates by a sedimentation/dehydration process. Samples were mounted on curved glass substrates to provide 8 sampling. Samples were then exposed to UV (254 nm> light at 4”C, typically for 15 min, to effect polymerization. Samples turned bright orange, indicative of short and/or disordered polymer, and did not further discolor following heating. The samples were kept in sealed brass canisters with thin aluminum windows at constant temperature and relative humidity, Cu-K, X-rays from a GX-18 rotating anode generator (Enraf-Nonius) were incident at grazing incidence, and scattered photons were detected with a Braun 1-D wire detector (Innovative Technologies) or film (Kodak DEF-5). Data were integrated by subtracting a

Fig. 1. X-ray diffraction patterns from DC,,,, PC. (a) With point focus optics, a sample mounted on a curved aluminum foil substrate (98% relative humidity, 15°C) reveals several low angle lamellar orders and three well de&red wide angle reflections. Because of the large mosaic spread, no indication of tilted chains can be deduced from the wide angle reflectionsg (b) With a sample geometry designed to minimize mosaic spread, the inner wide angle reflections are found to be strongest at 0’ (4.5 A) and 35” (3.9 A). Beam origin is indicated (0).

240

D.G. Rhodes et al. /Biochimica

et Biophysics Acta 1215 (1994) 237-244

multi-exponential background and summing counts between minima. Phases were determined using the swelling method [6,8,10-121, by plotting the square root of the normalized intensities as a function of s (= h/d = 2. sine/A). The continuous structure factor could then be inferred by interpolation. For some experiments, samples with reduced mosaic spread were required, and multilayers were deposited on glass coverslips. Excess glass was trimmed from around the multilayer using scissors, and the remaining shard ( N 2 mm X 3 mm) was glued to the end of a glass capillary. This sample was then aligned with the multilayer nearly parallel to the incident X-ray beam. Patterns were collected using Kodak DEF-5 film on a point focus beamline. 2.5. Thin films As described previously [7], lipids in chloroform solution were spread on a temperature-controlled Langmuir water as the subphase. The trough, with distilled pressure/area (r/A) isotherms were recorded using a Wilhelmy plate to measure r. For fluorescence mapping of monolayer domains, 2 mol% of rhodamine-phosphatidylethanolamine (rhodamine-PE) was mixed with the lipid before spreading. The domains were visualized by the selective partitioning of fluorescent dye to the fluid phase, so that the solid domains appeared dark against a bright fluid phase background. The fluorescence images were recorded with a fluorescence microscope-SIT video camera system. 2.4. Electron

intermediate rate. The trace of material migrating at intermediate rate may represent dimer or similarly small ‘polymer’, and the material at the origin (15%) is polymeric DC,, ,,PC (larger than dimer). 3.2. Calorimetry For both unpolymerized multibilayers and multibilayers irradiated with 254 nm UV light (as described above), DSC scans (heating) of DC 2,,7PC revealed a single endotherm at - 70°C (AH N 19 J/g) and cooling scans revealed a single exotherm at - 60°C (AH h 12 J/g). As expected, these values of T, are considerably higher than those of DC 2,,,,PC (35°C for the initial heating scan) multibilayers [8]. Because DC2,,,PC polymerized only to a limited extent, it is not surprising that DSC patterns of UV exposed samples did not differ significantly from those of unexposed samples. 3.3. X-ray difSraction Diffraction from multibilayers, prepared on aluminum foil substrates as reported in previous publications [8,10,11], typically exhibited a significant degree of mosaic spread. Fig. la was obtained from a point-focus beamline using a sample at 98% r.h. and 15°C. Three wide angle

diffraction

For electron diffraction experiments, single monolayers were deposited on Wilhelm coated grids by dipping the grids through the air/water interface as described previously [13]. A gold layer was pre-evaporated on the opposite side of the Wilhelm film for conductivity and calibration purposes. Low dose electron diffraction patterns were observed using a Hitachi H-600 microscope with a tilting stage, and recorded on a high sensitivity Kodak DEF X-ray film to reduce damage to the monolayer. Electron diffraction patterns were recorded at tilting angles, including 90”.

100

200

300

400

500

20 (channels) 1 t

b

3. Results 3.1. Polymerization While DC 2,10PC polymerizes within seconds, DC,,,,PC, under identical conditions, polymerizes to a limited extent (based on color change to bright orange), and only following prolonged (1.5 min) exposure to UV light. In TLC of the orange samples, 15% of the phospholipid (as phosphate) remained at the origin, 75% migrated at the same rate as monomeric DC 2,,7PC, and 10% migrated at an

Fig. 2. Diffraction data from DC2,,, PC. (a) Lamellar diffraction pattern obtained with a 1D electronic detector, (b) the resulting structure factor reconstructed by swelling analysis, as described in the text. Individual data points are shown for both + and - phases, but only one continuous structure factor is plotted.

241

LAG.Rhodes et al. /Bioc~i~~ca et B~op~ysicaActa 1215 (i994) 237-244

reflections arz obseyed at posit@s correspoading to spacings of 4.5 A, 3.9 A, and 3.6 A. The 4.5 A reflection is stronger at the equator, but the others are fairly uniform. These spacings did not vary significantly over a range of 98% r.h. to N 0%. The lamellar reflections are visible to h = 13, with the strongest reflections for h = 1-t. The lamellar spacing determined from this film is 66.4 A. Samples on glass coverslips with reduced mosaic spread were used in a stationary geometry, allowing the intrinsic mosaic spread to provide B sampling. ~thou~ quantitative sampling of 8 was not possible, as the details of the mosaic spread were not analyzed fully, these samples did yield good quality diffraction in both the lamellar and equatorial directions (Fig. lb). Under these conditions, the three wide angle, ring-like reflections observed with curved geometry samples were resolved into reoflections with limited arcs, These appeared with the 4.5 A reflection on the equator, the 3.9 A, reflection at 35” from the equatorial axis, and the 3.6 A. reflection at 35” from the equatorial axis. Lame&r data were collected using a one-d~e~ional electronic wire detector. A typical pattern is shown in Fig. 2a. In most data sets, 6-8 orders were of suffitient signalto-noise ratio to be integrated, and the 4.5 A reflection could sometimes be s,een at high angle. The lamellar repeat distance was N 70 A for ~511 hydrated samples ( 2 80% r.h.), and decreased to N 60 A at low humidity. This range of d provided sufficient sampling of the structure factor to determine phases by the swelling method. Fig. 2b shows the discrete structure factor data (& 1 with the estimated continuous reciprocal space function. These phases were used to determine the electron density profile structure (Fig. 3). The profile structure is generally typical of phospholipid bilayers, except that the headgroup peak exhibits a distinct broadening f f 18 A), which presumably corresponds to the increased electron density due to the diacetylene. The width of the bilayer is exceptionally small for a lipid with acyl chains as long as those in DCzr7PC (C,,). This observation requires that model st~~tmes fit to the profile have tilted and/or disordered acyl chains. A model structure with distal acyl chains tilted at 35” relative to the

tkan

MolecularArea (A”)

Fig. 4. Langmuir compression isotherm from DC&,PC. obtained at the pressures indicated, as described in the text.

Data were

bilayer normal can be fit to the electron density profile (not shown). 3.4. Monolayer characteristics At low temperatures (22°C and below), liquid phase DC&PC monolayers were compressible to about 50 AZ/molecule before any surface pressure became measurable (Fig. 4). The area per molecule increased with temperature up to about 45°C. At higher tem~ratures (> 45”C), two kinks or transition points became observable in isotherms of DC 2,17PCmonolayers (Fig. 41, in contrast to a single, well-defined liquid-expanded/liquid-~ndensed transition observed in isotherms of DC, roPC [7] or DCs,,PC [14] monolayers. Dark, thin, netilike domains (Fig. 5a) became observable at higher pressures (> 15 mN/m) for all temperatures, and at pressures higher than that of the first transition, if it appeared, indicating that the first transition was the onset of a mixed phase containing certain ordered domains. If such transitions existed in low temperature isotherms, they were too weak to be detected. The second transition, which was unique to DC,,,,PC, represented the onset of another form of ordered domains. Polygonal (mostly triangular) domains were visible at pressures higher than that of the second tr~sition (Fig. 5b). These polygonal, domains were smaller and had straight edges, instead of curved edges (as in domains of DC2,,,PC monolayer& but were otherwise similar in shape to the ‘feathered’ domains in monolayers of DC,,,PC [7]. The straight edges implied a strong intermolecular force in well ordered domains sufficient to overcome a tendency for inter-facial tension to minimize the boundary length (see Discussion). 3.5. Electron diffraction

v -30

-20

-10

(A;'

0

20

z Fig. 3. Electron density profile structure of DC2,1,PC.

30

Monolayers were transferred to Formvar supporting films to be studied by electron diffraction. Diffraction data were obtained from selected areas assigned by fluorescence microscopy. No sharp diffraction pattern was detected from the net-like domains observed at low pressure.

242

D.G. Rhodes et ai./Bj~~~ica

et Bioph~s~ca Acta 1215 (1994) 237-244

Therefore, the first transition was interpreted as the alignment of the diacetylenic groups at the C2 position while the remaining portion of the chains was still in the fluid state. Diffraction from polygonal domains showed two sharp spots at a spacing of 3.9 A (Fig. 6a). Therefore, the polygonal domains were likely to be domains where the entire acyl chain was in the solid state. Because only two reflections were observed when the monolayer plane was normal to the electron beam, either the supporting film was highly tilted or the solid state acyl chains were not aligned perpendicular to the monolayer plane [15]. The former case was not likely. In the latter case, the tilt angle was sufficiently large that the Ewald sphere did not intercept other lattice points (rods, for a single monolayer) on the same tilted reciprocal lattice plane of a two-dimensional crystal. Only when the normal to the monolayer plane was tilted to between 35” and 40” from the beam direction was a pseudo-hexagonal diffraction pattern observed (Fig. 6b). From the specimen tilt angle and pseudo-hexagonal spacings, the tilt angle of the

Fig. 6. Low dose electron diffraction patterns from DC,,,,PC monolayers deposited at 5OT with rr = 30 mN/m (Fig. 5b). Patterns were obtained at (a) normal incidence, with a pattern indicative of tilted chains, and (b) with the sample tilted at 35” relative to the electron beam, so as to optimize the hexagonal pattern. The rings at wide angle are from the gold layer evaporated on the other side of the coated grid.

chains was estimated to be about 30” to 35” from the bilayer normal, using the formula given by Tristram-Nagle et al. [16].

4. Discussion

Fig. 5. Fluorescence micrographs of DC,,,,PC monolayers deposited at 50” with pressures of (a) rr = 20 mN/m, (b) rr = 30 mN/m. The net like domains in (a) comprise only a small fraction ( - 20%) of the total area. The small solid phase domains in (b) become more numerous at higher pressures, but do not increase significantly in size. Although the effect is not as pronounced as in comparable domains of DC,,,PC monolayers 171, the solid domains are chiral. Bar in (a) is 20 pm.

In previous reports, it has been suggested that limited polymerization of DAPCs results from constrained acyl chains which cannot accommodate the conformational shift associated with polymer formation [4,6,8]. In DC,,,,PC [8], the diacetylene packing may be constrained only by the requirements for headgroup packing. The packing energy of proximal acyl chains (C2) is negligible, and that of distal acyl chains (Cll) is small enough that the packing energy can be overcome by polymer formation. In DC2,t7PC bilayers, however, the C,, distal acyl chain tends to remain as an ordered gel-state domain. X-ray diffraction data suggest weak order in the lamellar direction (large mosaic, small h,,,), but local in-plane ordering may still exist. The mosaic structure could arise from

D.G. Rhodes et al. / Biochimica et Biophysics Acta 1215 (1994) 237-244

misaligned, highly crystalline domains. Wide angle data were quite strong relative to the lamellar reflections, suggesting a high degree of order in the in-plane (chain packing) direction. Electron diffraction data also indicate significant inplane ordering. The diffraction spots were relatively well defined, and are consistent with a condensed, solid state, gel phase layer. Although it is difficult to prove that the packing in a deposited monolayer is equivalent to that in a bilayer in suspension, a multibilayer, or even a compressed monolayer, the similarity of the wide angle spacings suggests good correspondence. Equivalence is strongly supported in the present case, because of the clear demonstration of acyl chains tilted at 30-35” by electron diffraction with tilted samples, and the presence of off-axis reflections at the same angle and lattice constant in F-ray diffraction patterns. The acyl chain packing on a 3.9 A pseudo-hexagonal lattice represents very tight packing. The data reported here represent the first unambiguous demonstration that the off-axis wide angle reflections observed in X-ray diffraction patterns from DAPC multilayers may result from chain tilt rather than from higher order reflections. Data obtained by Lando and Sudiwala [14] suggested that there may be interbilayer registry in Langmuir-Blodgett deposited DAPC multibilayers. Most lipid multibilayers are stacks of parallel bilayers which may have some in-plane crystallinity, but are randomly oriented about the stacking axis. These yield diffraction patterns with lamellar (O,O,1) reflections corresponding to bilayer thickness along one axis and equatorial (h,O,O) reflections corresponding to acyl chain packing along a perpendicular axis. Whereas off-axis reflections could result from tilted chains in any system, a degree of three-dimensional (stacking) crystallinity could produce higher order (h,0,1 with h,l > 0) reflections. Fluorescence micrographs clearly show that the monolayers are not homogeneous. This does not mean that the phase distribution in bilayers or multibilayers is similarly heterogeneous. The steric and energetic constraints of a bilayer differ from those of a monolayer. In other lipid systems, dehydration has been shown to increase T,, so the partial dehydration in this case may increase the lipid order, promoting gel-phase formation in the multibilayer samples. The existence of the gel phase is demonstrated by compression isotherms, and its predominance by X-ray and electron diffraction data. It is important to note that for DC,,,,PC, which polymerizes readily, there is no indication of highly ordered pseudo hexagonal packing in similar electron diffraction experiments. The morphology of the solid domains suggest that the intermolecular interactions in these systems are quite strong. It is apparent for both DC,,,,PC and DC, 17PC that the solid domains are stable against the tendency to minimize the interface length by forming round domains (see [17], for example). Suppose that for both DC,,,,PC and D&PC one could initiate two-dimensional lattices of

243

equal spacing based on the headgroup, proximal acyl chain, and diacetylene. If the intermolecular interaction is strong, asymmetric domains can form. If the acyl chain packing is not exactly matched to this lattice, one might expect distortions to the lattice, perhaps forming the ‘feathered’ domains observed with D&PC. If the distal acyl chains are of sufficient length that the energy of distal chain packing becomes a significant (or dominant) factor in the packing energy, the distance over which the lattice can be propagated may be limited. That is, the lattice eventually becomes so distorted that addition of more molecules to the solid domain is no longer favored. With lower energy of distal chain interaction (in DC2,10PC), it is more difficult to reach the limiting conformation.

Conclusions

Stabilizing the structure of DC&PC by increasing the acyl chain length did not enhance, and in fact impeded polymerization due to additional constraints in the lipid packing involving the distal segment of the acyl chains. The bilayer structure of DC,,,,PC features strongly tilted acyl chains (35”), and this tilt was observed using X-ray diffraction and electron diffraction. For the first time, through the combination of the two techniques, the origin of off-axis reflections in small angle X-ray diffraction patterns from DAPC multibilayers has been identified. Although the lamellar ordering is limited, the in-plane packing appears to be highly ordered and relatively independent of hydration state over the range of 2 0% r.h. to N 98% r.h. Dark (gel phase) domains of D&PC are smaller than those of DC 2,1,,PC, perhaps due to constraints on the diacetylene packing resulting from distal acyl chain packing. Chains longer than 10 ‘CH, units are required to form highly crystalline gel phase packing distal to the diacetylene, as might be expected from phase transitions of short saturated lipid systems.

Acknowledgements

This study was supported by the National Science Foundation (CTS-8904938, D.G.R.), Glaxo Group Research, Ltd. (D.G.R.), and the National Institutes of Health (HL-16660 to RB and GM-28120 to S.H.).

References H. (1984) in Polydiacetylenes (Cantow, H.J., ed.), p. 1, Springer Verlag, Berlin. [21 Enkelmann, V. (1984) in Polydiacetylenes (Cantow, H.J., ed.), p. 91, Springer Verlag, Berlin.

111Blssler,

244

D.G. Rhodes et al./Biochimica

et Biophysics Acta 1215 (1994) 237-244

[3] Singh, A., Singh, B., Gaber, B., Price, R., Burke, T., Herendeen, B., Schoen, P., Schnur, J. and Yager, P. (1989) Surfact. Sol. 8,467-476. [4] Rhodes, D. and Singh, A. (1991) Chem. Phys. Lipids 59, 215-224. [5] Kuo, T. and O’Brien, D. (1991) Langmuir 7, 584-589. [6] Rhodes, D., Frankel, D., Kuo, T. and O’Brien, D. (1994) Langmuir 10, 267-275. [7] Hui, S.W., Yu, H., Xu, Z. and Bittman, R. (1992) Langmuir 8, 2724-2729. [8] Rhodes, D.G., Xu, Z. and Bittman, R. (1992) Biochim. Biophys. Acta 1128, 93-104. [9] Herbette, L.G., Vant Erve, Y. and Rhodes, D.G. (1989) J. Mol. Cell. Cardiol. 21, 187-201

[lo] Blechner, S., Morris, W., Schoen, P., Yager, P., Singh, A. and Rhodes, D. (1991) Chem. Phys. Lipids 1991, 58, 41. [ll] Rhodes, D., Blechner, S., Yager, P. and Schoen, P. (1988) Chem. Phys. Lipids 49, 39-47. [12] Franks, N. (1977) J. Mol. Biol. 100, 345-358. [13] Hui, S. and Yu, H. (1993) Biophys. J. 64, 150-156. [14] Lando, J. and Sudiwala, R. (1990) Chem. Mat. 2, 594-599. [15] Hui, SW. (1989) J. Electron Microsc. Tech. 11, 286-297 [16] Tristram-Nagle, S., Zhang, R., Suter, R.M., Worthington, CR., Sun, W-J. and Nagle, J.F. (1993) Biophys. J. 64, 1097-1099. [17] Grainger, D.W., Reichert, A., Ringsdorf, H. and Salesse, C. (1990) Biochim. Biophys. Acta 1023, 365-379.