Biomaterial99 Monoolein Review

Biomaterials 21 (2000) 223}234

The phase diagram of the monoolein/water system:metastability and equilibrium aspects

Hong Qiu, Martin Ca!rey*Department of Chemistry, The Ohio State University, 100 W. 18th Avenue, Columbus, OH 43210, USA

Received 14 September 1998; accepted 28 June 1999

Abstract

Interest in the liquid crystal structure, transport and membrane protein crystallizing properties of the monoolein/water system hasgrown in the recent past. Monoolein is also an important homolog in a series of monoacylglycerols used to decipher how lipidmolecular structure relates to liquid crystal phase behavior*information needed for rational design applications and for understand-ing the origin of membrane lipid diversity. To make intelligent use of the monoolein/water system, a reliable and detailedtemperature}composition phase diagram is needed. The phase diagram of Briggs et al. (J Phys II France 1996;6:723}51) wasconstructed for this purpose. However, we have established that the liquid crystal phases in the latter below ca. 203C are metastable.By implementing a sub-zero degree (3C) sample incubation prior to data collection in the heating direction, we can reset the systeminto the lamellar crystal phase which we assume represents equilibrium behavior. We have re-examined the low-temperature part ofthe phase diagram and characterized structurally the new &equilibrium' phases by static and time-resolved low- and wide-angle X-raydi!raction and by di!erential scanning calorimetry. A more complete phase diagram that incorporates the new equilibrium behaviorat low temperatures is reported. ( 1999 Elsevier Science Ltd. All rights reserved.

Keywords: Cubic phase; Lipid bilayer; Liquid crystal; Mesophase; Phase diagram; Undercooling

1. Introduction

1-Monoolein is a monoacylglycerol with 9-cis-oc-tadecenoic acid at the sn-1 position of glycerol. Themesophase propensities and structure of monoolein dis-persed in water are of interest in a number of areasranging from controlled uptake and release to cosmetic,food and pharmaceutical formulations. Most recently,the cubic phase of hydrated monoolein was reported toserve as an environment in which to grow crystals of themembrane protein, bacteriorhodopsin, for use in solvingits three-dimensional structure [2]. The tantalizing pros-pect exists that the monoolein, and related hydratedmonoacylglycerol systems, might be used in crystallizingother membrane proteins.

To make full use of the monoolein/water system insuch applications calls for a reliable and detailed temper-ature}composition (}C) phase diagram and for a struc-

*Corresponding author. Tel.: #1-614-292-8437; fax: #1-614-292-1532.

E-mail address: [email protected] (M. Ca!rey)

ture characterization of the phases formed. Over theyears, studies have been performed to realize these objec-tives and the literature abounds with such information.The most detailed phase characterization performed todate in the temperature range from 0 to 1103C wasreported from this lab [1]. We have since established,and report herein, that the liquid crystal phases existingbelow 203C described in that study represent metastablebehavior. Equilibrium properties of the system have beenassessed and quanti"ed in the current study by imple-menting an extended, low-temperature incubation ofsamples prepared at room temperature before perform-ing phase characterization in the heating direction.

Here, we describe the problems encountered in charac-terizing the equilibrium phase behavior of a system hav-ing phases with a pronounced tendency to undercool,and our methods for solving them. The equilibrium }Cphase diagram of the monoolein/water system is de-scribed in the temperature range from!15 to 1103C andin the water concentration range from dry to full hy-dration. In addition, structure characteristics of the as-sorted phases (lamellar crystal, lamellar liquid crystal,cubic (Pn3m, Ia3d), Fig. 1) found in the low-temperature

0142-9612/00/$ - see front matter ( 1999 Elsevier Science Ltd. All rights reserved.PII: S 0 1 4 2 - 9 6 1 2 ( 9 9 ) 0 0 1 2 6 - X

Fig. 1. Schematic representation of the various crystal, liquid crystaland #uid phases identi"ed in the temperature}composition phase dia-gram of monoolein.

region of the phase diagram are reported based on X-raydi!raction measurements. These are compared withthose of related hydrated monoacylglycerols. The objec-tive here is to extend our understanding of the relation-ship between lipid molecular structure and mesophaseproperties for use in rational design. The informationshould also prove useful in deciphering the membranelipid diversity enigma: why most biomembranes containsuch a bewildering array of lipid types.

2. Materials and methods

2.1. Materials

Monoolein was purchased from Nu Chek Prep Inc.(Elysian, MN). It had a purity of*99% as determined bythin layer chromatography [3] and was used withoutfurther puri"cation. Thin layer chromatography was alsoused to monitor monoolein purity change after the sam-ples had been used in X-ray and DSC measurements. Noobvious degradation of the lipid was detected. It is

assumed that the isomerization equilibrium between1-monoolein and 2-monoolein [4] is independent of tem-perature in the range studied and that the lipid can betreated as a single component. Water was puri"ed byusing a Milli-Q Water System (Millipore Corporation,Bedford, MA) consisting of a carbon "lter cartridge, twoion-exchange "lter cartridges and an organic removalcartridge.

2.2. Sample preparation for X-ray diwraction

2.2.1. Standard preparationDry solid monoolein (ca. 20 mg) was mechanically

mixed with appropriate amounts of water (ca. 1}50 mg)in a syringe-based mixing device as described [5] toachieve the desired sample composition. Mixing wascarried out at room temperature (ca. 213C) while prepar-ing the "xed hydration samples ranging in compositionfrom 0 to 50% (w/w) water. The homogeneously mixedsamples were transferred to 1 mm diameter quartzcapillaries (Charles Supper, Natick, MA), #ame-sealed and glued with 5-min epoxy (Hardman Inc.,Belleville, NJ), and were stored prior to data collection at43C for a period ranging from a few days up to4 weeks. The actual composition of the samples wasdetermined gravimetrically as described [6] with anaccuracy of better than 0.1% (w/w) water for low watercontent (ca. 0}30% (w/w)) samples and 0.5% (w/w) waterfor high water content (30% (w/w)) samples,using a microbalance (M3P-000V001, Sartorius Corp.,Edgewood, NY) [5].

2.2.2. 03C preparationBesides the standard room-temperature sample prep-

aration protocol described, samples were also preparedat 03C. A monoolein sample with 32% (w/w) water wasprepared by mechanically mixing 21 mg dry monooleinpowder and 10 mg milliQ water at 03C on ice. A slow rateof mechanical mixing as described [5] was used to min-imize the likelihood of inadvertent sample frictional heat-ing. The resultant mixture was placed in capillaries at03C. Some of the "lled capillaries were stored sub-sequently at 43C while others were held at room temper-ature (ca. 213C). X-ray di!raction was used to monitorthe phase state of these samples as a function of time at4 and 213C.

2.2.3. Passive swellingThe passive swelling technique of Briggs et al. [1] was

used to prepare fully hydrated samples from water-stressed samples without mechanical mixing and thus,the occurrence of possible inadvertent frictional heating.This approach was used to minimize the likelihood ofproducing an undercooled cubic phase as described [1].For this purpose, an 18% (w/w) water sample was pre-pared following the standard preparation method above

224 H. Qiu, M. Cawrey / Biomaterials 21 (2000) 223}234

and was subsequently stored at 43C for one week. Thecapillary was then carefully broken at its glued end, andan excess of milliQ water at 43C was placed in thecapillary in direct contact with the hydrated lipid. Thesample was subsequently equilibrated at 43C andmonitored visually and by X-ray di!raction at 43C.

2.3. X-ray source

Copper Ka

X-rays (1.542 As , nickel (0.025 mm-thick)"ltered) were produced using a two-beam port RigakuRU-300 rotating anode X-ray generator (18 kW, RigakuUSA Inc., Danvers, MA), operated at 40 kV and 200 mA,and a horizontally positioned 0.5 mm]10 mm "lament.The beam was focused horizontally and vertically bya pair of 6 cm]2 cm (length]width) nickel-coated glassmirrors (Charles Supper, Natick, MA) placed verticallyand horizontally ca. 20 cm downstream of the anode,giving a focused beam measuring 1.5 mm]1 mm at thedetector which was positioned ca. 25 cm from the down-stream mirror. This beam was used for both static andtime-resolved X-ray di!raction measurements.

2.4. X-ray diwraction data collection

Static X-ray di!raction patterns were collected in thetemperature range from !15 to 553C in increments of1}53C in the heating direction. Up to seven samples canbe accommodated at one time in a temperature-control-led beryllium, multiple sample holder. Seven di!ractionpatterns measuring 1 in]8 in (covering a real spacerange from ca. 2 to 160 As at a sample-to-detector dis-tance of 20}25 cm) were collected side-by-side behinda 1 in wide lead slit on an 8 in]10 in image plate (FujiHR-IIIN, Fuji Photo Film Co. Ltd., Japan).

Before making di!raction measurements, samples wereincubated at!103C or lower for at least 2 h to facilitatefull development of the L

#phase. This is what we refer to

as the standard &sub-zero degree' (Celsius) incubation.Subsequently, temperature was increased and sampleswere equilibrated at a particular measurement temper-ature for anywhere from 3 to 12 h before making the15 min}1.5 h exposure at a sample-to-detector distanceof ca. 25 cm.

Time-resolved X-ray di!raction measurements weremade following temperature-jump perturbations to thesample. The sample was initially incubated at 253C for atleast 30 min at which point temperature dropped to!103C within 5 min. Mesophase properties of thesample before, during and for 30 min after the temper-ature jump were monitored continuously in the streakmode with the image plate translating at a rate of1.5 mm/min while positioned approximately 5 mm be-hind a 3 mm wide lead slit at a 21.3 cm sample-to-detector distance (see [23] for details).

2.5. Image analysis

A phosphorimage scanner (Storm 840, Molecular Dy-namics Inc., Sunnyvale, CA) operating at a resolution of100 lm/pixel and a dynamic range of 105 was used toread images recorded on the image plates. The patternwas stored as a .gel "le and subsequently converted to .tif"le for viewing and analysis using Adobe Photoshop (v4.0, Adobe Systems Inc., San Jose, CA) and Matlab (v 5,The Mathworks Inc., Natick, MA). An image analysisprogram, written in Matlab, was used to locate the centerand to circularly average the di!raction patterns. Origin(v 5.0, The Microcal Software Inc., Northampton, MA)was used to "t the averaged intensity versus scatteringangle pro"les and to determine peak position. Sample-to-detector distance was calculated using silver behenate(sample provided by Dr. Blanton, Analytical TechnologyDivision, Eastman Kodak Company, Rochester, NY)and 58.4 As as the lattice parameter [7].

2.6. Calorimetry measurements

Di!erential scanning calorimetry (DSC) measurementswere made using a Hart DSC-II (CSC4100, CalorimetrySciences Corp., Provo, UT). Calorimetry was used toevaluate the e!ect di!erent sample thermal histories hadon phase behavior.

Appropriate amounts (tens of milligrams) of mono-olein and water were mechanically mixed at room tem-perature as stated in Section 2.2. Three samples withdi!erent concentrations were made, and a portion ofeach was placed in two DSC ampoules. In this way, twosets of three samples containing 14.5, 37.3 and 68.5%(w/w) water were prepared. Each set was held at 373C for10 min to induce the formation of the L

aor cubic phases.

Afterwards, the "rst set (Set 1) was incubated at!103Cfor 1 h while Set 2 was incubated at 03C for 6 h beforerecording the thermograms up to 603C at a heating rateof 153C/h.

3. Results

3.1. Temperature}composition phase diagram

In this study, great care has been taken to ensure thatequilibrium phase behavior prevails in the temper-ature}composition (}C) phase diagram reported for themonoolein/water system (Fig. 2). This includes incuba-tion of all samples at!133C for at least 2 h before datacollection performed in the heating direction and longequilibration times (3}12 h) at each measurement tem-perature. The phases identi"ed by X-ray di!raction andtheir location in temperature}composition space areshown in Fig. 2A. Fig. 2B shows the phase boundariesand coexistence regions as identi"ed by the di!raction

JBMT 1216

H. Qiu, M. Cawrey / Biomaterials 21 (2000) 223}234 225

Fig. 2. (A) Identity and location in temperature}composition space of each phase and coexisting phases in the monoolein/water system as determinedby X-ray di!raction in the heating direction in the!15}553C temperature and 0}50% (w/w) water composition ranges. The identity of the phases is asfollows: (v) ice, (n) L

#, (]) L

a, (L) cubic-Ia3d, (h) cubic-Pn3m and (#) FI. (B) Temperature}composition phase diagram of the monoolein/water

system based on an interpretation of the data in Fig. 2A, and the hydration data as in Fig. 3B (location of hydration boundary). The excess waterboundary of the L

#phase is at about 4% (w/w) water, a best estimate as described in the text, and is indicated by a dashed line. (C) Composite

temperature}composition phase diagram of the monoolein/water system that combines Fig. 2B of the current work covering the range of!15}553Cand Fig. 5B of Briggs et al. [1] in the region above 303C.

data. It was drawn to conform to both the experimentaldata and the Gibbs' phase rule.

The pure phases found in this system include theL#

phase, the La

phase, two inverted bicontinuous cubicphases belonging to space groups Ia3d and Pn3m, andthe #uid isotropic (FI) phase. The description of thephase diagram that follows is based on that presented inFig. 2B. The pure L

#phase exists only at low water

concentration below ca. 343C. However, the excess waterboundary for the L

#phase has not been determined with

accuracy. The latter is shown at ca. 4% (w/w) water(dashed straight line, Fig. 2B) based upon the followingobservations: (i) the lattice parameter of the L

#phase

does not change with hydration above 4% (w/w) (seeFig. 3B) and (ii) ice formation was seen in samples with4% (w/w) water but not in samples of lower hydration.At least two L

#polymorphs exist as identi"ed by their

disparate wide-angle di!raction patterns (Fig. 4). How-

ever, this work is not concerned with crystal polymor-phism and this issue was not explored in the currentstudy.

Fusion of dry monoolein crystals to the melt occursbetween 31 and 363C. The FI phase occupies the high-temperature part of the diagram at all hydration levelsand its low-temperature limit reaches a minimum of ca.263C in the vicinity of 4% (w/w) water. In the temper-ature range from 20 to 503C, the series of pure liquidcrystal phases that form with increasing hydration arethe L

a, the cubic-Ia3d and the cubic-Pn3m phases.

Phases that exist in equilibrium with excess water in the0}503C range include the L

#and the Pn3m phases.

Phase behavior involving L#

at temperatures below itsmelting point is relatively complicated (see Fig. 2B).Special care was taken to decipher the phase sequence inthis lower temperature region by making measurementsin increments as small as 0.53C and by implementing

226 H. Qiu, M. Cawrey / Biomaterials 21 (2000) 223}234

Fig. 3. Temperature- and composition-dependence of the structure parameters of the phases found in the monoolein/water system at the indicatedsample compositions in units of % (w/w) water and at selected temperatures. The structure parameter values reported are accurate to $0.1 As for all ofthe phases, with the exception of the FI phase (characterized by di!use scatter) and the cubic phases when coexisting with the L

#phase (where

scattering angle is small and di!raction is weak). The identity of the phases is as follows: (v) ice (the re#ection at 3.9 As is used in the plot), (n) L#, (])

La, (L) cubic-Ia3d, (h) cubic-Pn3m, and (#) FI.

long incubation times prior to data collection. The data(Fig. 2A) are consistent with L

#coexisting with ice below

03C, L#

plus water coexistence from 03C to ca. 83C,L#

plus cubic-Pn3m coexistence from 8 to 173C, andL#

plus cubic-Ia3d from 17 to 183C. The La

phase co-exists with L

#between 18 and 263C. Coexistence of

L#

and FI was detected at 0% (w/w) water at 22 and313C. In going from low to high temperature, the hy-dration range associated with each of these coexistence"elds decreases.

A detailed examination of the transition occurring inthe vicinity of 03C was not undertaken. Accordingly, theline drawn along the 03C isotherm in Fig. 2B and C couldrepresent a polytectic transition if it occurs at exactly03C. On the other hand, if it is isothermal below 03C,then a eutectic transition is implied. The poor solubility

of monoolein in water would place the latter justbelow 03C.

The current data refer to the phase behavior of themonoolein/water system in the !15}553C range andfrom the dry state to the condition of full hydration. Thephase diagram of Briggs et al. [1] covers much the samecomposition range, extends to considerably higher tem-peratures and includes metastable phase behavior in thelow temperature region. By way of producing a morecomplete representation of what we consider to be equi-librium phase behavior over the full temperature rangefrom !15 to 1103C, we have combined the two phasediagrams as shown in Fig. 2C. Thus, Fig. 2C combinesthe 303C data in Fig. 2B of Briggs et al. [1] and thedata in Fig. 2B of the current study. We consider thephase diagram presented in Fig. 2C to faithfully represent

JBMT 1216

H. Qiu, M. Cawrey / Biomaterials 21 (2000) 223}234 227

Fig. 4. X-Ray di!raction patterns of monoolein with 25% (w/w) water in the!15}313C range. The patterns were collected on an image plate at theAdvanced Photon Source (line 1-BM, Argonne National Lab) with an exposure time of 3 s. X-ray wavelength was 1.377 As and the sample-to-detectordistance was 373 mm. (a)!14.53C, coexistence of the L

#phase and ice; (b) 2.63C, coexistence of the L

#phase and water; (c) 15.13C, coexistence of the

L#

and Pn3m phases; (d) 17.43C, coexistence of the L#

and Ia3d phases; (e) 19.73C, coexistence of the L#

and Ia3d phases; (f ) 31.13C, pure Ia3d phase.Di!erent polymorphs of the L

#phase were observed (compare wide-angle pattern in (a), (b) and (c).

the equilibrium phase properties of the monoolein/watersystem in the temperature and composition range shown.

3.2. Structure parameter temperature- andcomposition-dependence

The temperature- and composition-dependence of thestructure parameter of the di!erent mesophases areshown in Fig. 3. The monoolein/water system exhibitstypical liquid crystal thermal and composition expansivi-ties to within the limits of measurement accuracy. Speci"-cally, the structure parameter of all mesophases decreaseswith increasing temperature (Fig. 3A), while the structureparameter of all pure phases increases with sample hy-dration (Fig. 3B).

The composition dependence of the lattice parameter(Fig. 3B) shows that the monoolein/water system exhibitsthermodynamic invariance with respect to composition.This is expected for a binary system where two phasescoexist at a "xed temperature (and pressure). Thus, thecomposition of the two coexisting phases remains con-stant while the relative amounts of the two phases changeas the overall composition is varied isothermally. Thismeans that in any two phase coexistence region, the

lattice parameter of each phase should remain constantand insensitive to overall sample composition. Thus, forexample, at 183C, the L

#phase coexists with Ia3d in the

concentration range from 18.4 to 39% (w/w) water(Fig. 3B). In this composition range, the lattice parameterof the L

#phase is constant at 49.3$0.3 As , while that of

Ia3d remains "xed at 173$2.2 As . The relatively largeerror associated with the Ia3d lattice parameter arisesfrom the fact that the phase does not scatter X-raysparticularly strongly, which exacerbates the problemwhen the cubic phase is not the only phase in the system.On top of this, the scattering angle is low which adds tothe uncertainty in peak scatter determination. The sameholds for the Pn3m phase in coexistence with theL#

phase.

3.3. Calorimetry measurements

DSC measurements were used to determine the e!ectof two di!erent pre-measurement incubation protocolson the phase properties of the monoolein/water system.For this part of the study, duplicate sets of samples with14.5, 38.5 and 87% (w/w) water were prepared in sealedcalorimetry ampoules. Following a 10 min incubation at

228 H. Qiu, M. Cawrey / Biomaterials 21 (2000) 223}234

Fig. 5. Di!erential scanning calorimetry thermograms of monoolein/water mixtures recorded in the heating direction at 153C/h. Each set consists ofthree samples: (a) 14.5% (w/w) water, (b) 38.5% (w/w) water, and (c). 87% (w/w) water. Set 1 was incubated at 03C for 6 h before performing the scan.Set 2 was incubated at!103C for 1 h before performing the scan. The y-axis scale in Set 1 is 10-times that in Set 2. What looks like a peak in the vicinityof 33C in Set 1 thermograms arises from sample-reference mismatch. The same e!ect is present in Set 2 but is less obvious because of the di!erent scalesettings. The phases encountered during the heating scans are noted and were identi"ed based on X-ray measurements.

373C to induce the existence of the La

and/or cubicphases, Set 1 samples were held at 03C for 6 h while thosein Set 2 were incubated at!103C for 1 h before record-ing the heating thermograms (Fig. 5). The protocol im-plemented with Set 1 samples mimics the conditions usedby Briggs et al. [1] where metastability was observed.The so-called sub-zero degree incubation conditions usedwith Set 2 were chosen to avoid undercooling.

The primary di!erence between the two sets of samplesin terms of calorimetric behavior is that a major en-dotherm is seen in Set 2 between 5 and 203C that isabsent in the Set 1 thermograms (compare Fig. 5(1) and(2)). Referring to the phase diagram shown in Fig. 2B, it isobvious that the endothermic feature in Set 2 arises froma series of transformations involving the L

#phase. The

nature of these changes depends on sample compositionas dictated by the data in Fig. 2B. In stark contrast, nosuch endothermic event occurs in Set 1 samples sugges-ting that the L

#phase has not formed under these condi-

tions and that the phases seen below ca. 203C areundercooled. This is precisely what was reported in theoriginal paper by Briggs et al. [1].

The heat change associated with the L#

phase trans-formation is highly endothermic regardless of the liquidcrystal phase to which it converts (Fig. 5(2)). The corre-sponding enthalpy changes (*H) are as follows: ca.20 cal/g of mixture (8 kcal/mol of lipid) for the L

#-to-L

a

transition at 14.5% (w/w) water; ca. 16 cal/g of mixture(9 kcal/mol of lipid) for the L

#-to-Ia3d transition at

38.5% (w/w) water; and ca. 4 cal/g of mixture (11 kcal/mole of lipid) for the L

#-to-Pn3m transition at 87% (w/w)

water.The thermograms also reveal weakly endothermic

transitions such as those associated with the La-to-Ia3d

phase change at 503C (scan rate 153C/h, *H"100 mcal/g of mixture (40 cal/mol lipid) Fig. 5(1a)and the Ia3d-to-Pn3m transition at 403C (scan rate153C/h, *H"10}20 mcal/g of mixture (5}10 cal/mol lipid)Fig. 5(1b)). Hyde et al. [15] estimated that the lattertransition should have an associated *H of &less thanabout 0.01 kJ/mol' (about 10 mcal/g of mixture) whenviewed in the context that the interconverting cubicphases incorporate in"nite periodic minimal surfaces andthat the phase transition involves a change in lipid bi-layer structure but no change in curvature or breakage ofthe bilayer itself. The experimental and estimated valuesare in remarkably good agreement. A similar value hasbeen measured for the Ia3d-to-Pn3m transition in hy-drated monovaccenin, a monoacylglycerol di!ering frommonoolein in that the double bond is at C11 (unpub-lished data from this lab).

Given our focus on metastability and undercooling,it is interesting to note that water in the hydratedmonoolein samples readily undercools, but it does so

JBMT 1216

H. Qiu, M. Cawrey / Biomaterials 21 (2000) 223}234 229

intermittently. The thermograms in Fig. 5(2) illustratethis point in that the ice melting endotherm at 03C is seenin just one of the three samples that had been incubatedfor 1 h at!103C. In this and related hydrated monoacyl-glycerol systems, we "nd that water will undercool read-ily to!173C.

4. Discussion

4.1. c

as the equilibrium phase at low temperature

In what follows, we present the evidence assembled todate in support of the claim that in the low-temperatureregion of the monoolein/water phase diagram, theL#

phase represents equilibrium behavior.

4.1.1. Sub-zero degree preincubationThe "rst piece of evidence is presented in Fig. 2 where

L#

exists pure or in coexistence with ice, water, ora variety of lipidic liquid crystal and liquid phases. Thephase diagram was constructed using samples that hadbeen subjected to a 2 h incubation at !133C prior todata collection in the heating direction. This so-calledsub-zero degree preincubation sets the system into theequilibrium L

#phase from which the other liquid and

liquid crystal phases emerge upon heating. Omitting thesub-zero degree preincubation with samples hydrated attemperatures *203C allows for undercooling and per-sistence of the liquid crystal phases down to 03C at leastas observed by Briggs et al. [1] and in the calorimetrypart of the current study.

4.1.2. Passive swellingThe idea we are following here is as follows. Fully

hydrated monoolein prepared at room temperaturespontaneously forms the cubic phase. When such sam-ples are cooled to 03C, for example, in preparation forphase studies to be performed in the heating directionbeginning at 03C, they are prone to undercool. Thus, thephase observed in the 0}203C range very likely does notrepresent equilibrium for the system. The question is howto avoid or to by-pass the cubic phase as an initial phasein preparing a sample that is fully hydrated at 03C and inits equilibrium state? Our approach here was to preparea hydrated monoolein sample at 18% (w/w) water whereit exists in the L

aphase at room temperature and to set it

into the equilibrium L#

phase following the protocoldescribed in Section 2.4. It was subsequently broughtfrom!13 to 43C and incubated in direct contact with anexcess of MilliQ water. The sample was monitored at 43Cby X-ray di!raction to follow any changes in phase stateand/or hydration with time. None was observed and thesample remained in the original L

#phase for a period of

at least 5 months. While not proof positive that the

L#

phase represents equilibrium, this result is consistentwith it being the equilibrium phase.

Particularly intriguing is the observation that a similarpassive swelling experiment performed at 43C by Briggset al. [1] produced the cubic-Pn3m phase from an under-cooled L

aphase. This requires the system to remain

undercooled while the La

phase imbibes water, trans-forms "rst to the Ia3d and then to the Pn3m phase. Thishighlights the enigmatic nature of metastability in thatphases behave structurally (see below) and undergophase transitions just like systems at equilibrium all thewhile being metastable. A similar observation has beenmade in the N-dodecanoyl-N-methylglucamine}watersystem [8].

4.1.3. Low-temperature sample preparation and incubationIn this experiment, a sample of monoolein at 32%

(w/w) was prepared by mechanically mixing the twocomponents, water and monoolein, on ice at 03C. TheL#

phase forms under these conditions as demonstratedby X-ray di!raction performed on the sample at 03C.Temperature was adjusted subsequently to 43C and thesamples were held at 43C for a period of at least5 months. By X-ray di!raction, the L

#phase persisted

throughout the entire incubation. This protocol avoidsall liquid crystal phases during sample preparation andthus, the possibility of the system becoming trapped ina metastable L

aor cubic phase.

4.1.4. CalorimetryThe thermograms in Fig. 5 demonstrate clearly the

existence of metastable phases in the monoolein/watersystem when sub-zero degree incubation conditions arenot implemented prior to data collection in the heatingdirection. Preincubation at!103C for 1 h is su$cient todestabilize the undercooled liquid crystal phases in favorof the equilibrium L

#phase.

The phase diagram in Fig. 2B includes several regionswhere coexistence with the L

#phase is observed and

where the lever rule should and does apply as presentedunder Section 3. At "rst pass, this might be taken asevidence that the L

#phase is behaving as an equilibrium

phase. However, the data in the paper by Briggs et al.([1], Fig. 7A, 0}303C) where metastable L

aand cubic-

Ia3d phases coexist also show this same behavior. Thus,adherence to the lever rule cannot be used as evidence forequilibrium.

We have implemented a sub-zero degree preincubationprotocol as a means for reliably setting monoolein/watersamples into the equilibrium L

#phase. We must now

address the possible e!ect that ice formation, which canaccompany such low-temperature treatment, has on thephase properties of the system. Thus, it is possible that iceformation triggers the production of the L

#phase as

a result of sample dehydration and that the L#

phase isreally an artifact of the sample preincubation protocol.

230 H. Qiu, M. Cawrey / Biomaterials 21 (2000) 223}234

Fig. 6. Phase changes in hydrated monoolein following a jump intemperature. Shown in (A) is an example of portion of a streak low- andwide-angle X-ray di!raction pattern collected before, during and afterthe temperature jump from 253C at t"5 min where the hydrated lipidis in Ia3d cubic mesophase, to!103C where coexistence of the L

#phase

and ice prevails. Sample composition was 32% (w/w) water. The imageplate was translated behind a 5 mm wide slit at a constant rate of1.5 mm/min. The identity of the di!erent phases is indicated at the topof the "gure. The d-spacing of the L

#(0 0 1) and Ia3d (2 1 1) re#ections

is indicated. The re#ections at 3.9, 3.7 and 3.5 As are from ice. Thehorizontal axis corresponds to elapsed time as indicated in (B). Thevertical axis represents scattering angle (2h) with 2h"0 at and increas-ing in either direction above and below the asterisk. The time course ofin-sample temperature shown in (B) was measured with a T-typethermocouple (0.41 mm in diameter) read by Physitemp BAT-12(Physitemp Instruments Inc., Clifton, NJ). In this temperature-jumpstudy, two L

#polymorphs (L

#1, L

#2) emerge during cooling. The

number of L#

polymorphs observed in such measurements is variable.In some cases, a single polymorph is seen. In others, as in this case, twopolymorphs are observed. What is consistently observed however isthat L

#formation precedes the water/ice transition.

Our time-resolved X-ray di!raction measurements dur-ing the course of a temperature jump from 25 to!103Cdemonstrate unequivocally that this is not the case(Fig. 6). What is seen is the emergence of the L

#phase, at

the expense of the liquid crystal phase, that precedes iceformation. The calorimetry data in Fig. 5(2) also supportthis claim. As noted, ice formation upon cooling to!103C is intermittent. Thus, in two of the three thermo-grams in Fig. 5(2), ice fails to form but the highly en-dothermic transition below 203C that is characteristic ofthe L

#phase, is present in all three.

4.2. Comparison of monoolein/water phase diagrams

An overview of the literature concerning monoolein/water miscibility suggests a three-stage development inour understanding of the system. The "rst stage is repre-sented by the work of Lutton in 1965 [9] where &consist-ency' and polarized light microscopy were used toidentify the major phases (L

a, cubic, inverted hexagonal

(HII), FI). The second stage involves further phase identi-

"cation and structure characterization by means of X-raydi!raction and NMR-based di!usion measurements[10}14]. It was in this period that the di!erent cubicphases were identi"ed. The third stage heralds in moredetailed studies of miscibility properties and an attemptto characterize true equilibrium behavior of the systemby a myriad of methodologies [1,15}19]. At temper-atures above ca. 303C, the reported phase diagrams are inreasonable agreement. However, signi"cant di!erencesare observed below this temperature. In particular, thediagram of Briggs et al. [1] shows liquid crystal phasesprevailing down to 03C (Fig. 7A) while that of Aomoriet al. [17] has liquid crystal phases stable to approxi-mately 83C and &crystal#water' in the region from 8 to!403C (Fig. 7B). The phase diagram reported in thecurrent study (Fig. 2B) is at variance with both of these.We now know that the disparity between the currentwork and that of Briggs et al. is due to inadvertentundercooling of the liquid crystal phases below 203Cencountered in the latter work. The cause of di!erenceswith the Aomori et al. phase diagram is a little moredi$cult to identify but could in part be attributed to thefact that the latter study was performed using relativelyimpure lipid; speci"cally, commercially available foodadditive containing 97.5% monoolein [17].

Our conclusion is that the phase diagram in Fig. 2Crepresents equilibrium over the full temperature andcomposition range shown and that phase boundary loca-tions are reliable to $23C and$2% (w/w) water. Thediagram reported by Briggs et al. [1] in the temperaturerange below 203C represents metastable behavior ofa system that is kinetically stable. Following the protocolfor sample preparation as described in that work will,with high probability, produce the corresponding meta-stable behavior. Indeed, this same metastability might beused to advantage in certain applications where thestructural and phase properties of the undercooled hy-drated monoolein satisfy a particular need that cannot bemet by another monoacylglyceride.

4.3. Correlation between molecular structureand phase behavior

In the long term, we wish to use phase behavior studiesof the monoacylglycerols to understand the relationshipbetween lipid molecular structure and mesophasepropensity and microscopic structure. For example, by

JBMT 1216

H. Qiu, M. Cawrey / Biomaterials 21 (2000) 223}234 231

Fig. 7. Temperature}composition phase diagrams of the monoolein/water system redrawn from (A) Briggs et al. [1] and (B) Aomori et al. [17].

comparing the members of a series of monoacylglycerolhomologs, we seek to establish the e!ect that fatty acylchain length and double bond position has on phasebehavior.

Such a comparison is particularly appropriate in thecase of monoolein and monovaccenin, two monoacyl-glycerols that di!er structurally only in the position ofthe cis double bond. In the case of monoolein, the site ofunsaturation is at C9. In monovaccenin, it is at C11.A comprehensive comparison of the two has been madein the temperature range above 303C [3]. Because theyare positional isomers with very similar e!ective lipidchain lengths, they exhibit the same phases in the sameorder with respect to temperature and hydration andhave a similar structure parameter temperature- andcomposition-dependence as expected. However, there areslight di!erences in phase boundaries that were at-tributed to the disparate molecular shapes arising fromthe di!erent location along the chain of the double bondkink. The current study that focuses on phase behavior ofthe monoolein/water system in the !15}553C temper-ature range shows that the two systems also exhibitqualitatively very similar phase behavior in this low-temperature region. There are notable di!erences how-ever in that the transition temperatures associated withthe L

#phase are lower and the hydration boundary shifts

to lower water contents in the case of the monooleinsystem. For example, the temperature where the L

#phase

"rst transforms to a liquid crystal phase is ca. 83C in thecase of monoolein and ca. 163C for monovaccenin. Fur-ther, the (L

##Pn3m)-to-(L

##Ia3d) transition occurs at

ca. 173C and ca. 203C in the case of monoolein andmonovaccenin, respectively. These di!erences will be in-terpreted in the context of the variation in molecularshapes of the two molecules as described quantitativelyby means of the so-called shape factor [20].

The shape factor or lipid packing parameter, c, isa simple and quantitative way of describing molecular

shape in a self-assembled aggregate such as the assortedlyotropic liquid crystal phases encountered in lipidic sys-tems. In the bicontinuous cubic phase, c can be calculatedbased on the in"nite periodic minimal surface model asfollows:

c"