Investigative Ophthalmology & Visual Science, Vol. 32, No. 8, July 1991Copyright © Association for Research in Vision and Ophthalmology

Structural Characterization of Human Lens MembraneLipid By Infrared SpectroscopyDouglas Borchman,* M. Cecilia Yappert,t and Paul Herrell*

Fourier transform infrared spectroscopy was used to measure the structural changes in lipid hydrocar-bon chain and interface regions of membranes that were obtained from the cortex and the nucleus of thenormal human lens. Temperature-dependent studies, ie, phase transitions, were performed to evaluatethe differences in the thermodynamic properties of the nuclear and cortical lipids. The structure of thefatty acyl chain region showed a higher degree of order for the nuclear lipid membranes as comparedwith the cortical ones. At physiologic temperature, the acyl chains of lipid from the cortical region of a51-yr-old lens showed a degree of disorder of 63 ± 0.6% compared with 23 ± 1% for the nuclear region.The gel-to-liquid crystalline-phase transition temperatures were 27.2 ± 0.3 and 39.2 ± 1 °C for thecortical and nuclear lipids, respectively. From the phase transition data, the enthalpy (strength oflipid—lipid interactions), entropy (randomness of the bilayer), and cooperativity (influence of adjacentlipid molecules) were calculated to be 2.6,1.8, and 2 times greater, respectively, for the nuclear lipidtransition compared with the cortical lipid transition. These differences show stronger lipid interac-tions and higher order in the nuclear membranes as compared with those in the cortex. Energeticdifferences between the cortical and nuclear membranes may arise from differences in the level ofhydration or in the packing at the interface region. This last possibility is supported by changes in thecontour of the carbonyl band near 1743 cm"1. Invest Ophthalmol Vis Sci 32:2404-2416,1991

The structure of lipids can affect membrane proper-ties and functions such as membrane permeabilityand kinetics of enzymatic processes.1"4 The structureof lipid membranes has been studied indirectly withthe use of lipid order parameters that were obtainedwith electron spin resonance5 and fluorescence polar-ization probes.6'7 Both approaches present limita-tions,89 particularly in lipid matrices that containlarge amounts of cholesterol.10 Vibrational spectros-copy and x-ray crystallography allow the direct eluci-dation of lipid structure. With the use of infrared orRaman scattering vibrational spectroscopy, the nativesystem in the absence of any perturbing probe mole-cule can be investigated. Although Raman scatteringand infrared spectroscopy are proven techniques forthe characterization of lipid structure and lipid-pro-tein interactions, neither technique has been used tostudy lens lipid structure.

Alterations in lens lipid composition have been re-

From the *Department of Ophthalmology and Visual Sciences,Kentucky Lions Eye Research Institute and department of Chem-istry, University of Louisville, Louisville, Kentucky.

This work was supported by USPHS research grants EY06916,EY06917 and EY07975, The Kentucky Lions Eye Foundation,and an unrestricted grant from Research to Prevent Blindness, Inc.

Submitted for publication: August 10, 1990; accepted March 15,1991.

Reprint requests: Douglas Borchman, PhD, Department of Oph-thalmology, University of Louisville, KY 40292.

ported with lens region, age, and cataractogene-s j s 6,10-27 T n e s e changes in composition could lead toalterations in lipid structure that may account for thechanges in membrane transport28"33 and membranepermeability.34"36

In this study, infrared spectroscopy was used tostudy the structure of multilammellar lipid disper-sions from lipids that were extracted from the corticaland nuclear membranes. The lipids were preparedfrom a pair of normal lenses that were obtained froma 51-yr-old man.

Materials and Methods

All chemicals were obtained from Sigma ChemicalCompany (St. Louis, MO), except American Chemi-cal Society (ACS) grade methanol, and high perfor-mance liquid chromatography (HPLC) grade chloro-form, which were obtained from Fisher ScientificCompany (Cincinnati, OH).

Preparation of Lipid Vesicle Dispersions for InfraredMeasurement

A pair of clear lenses was extracted 4 hr post mor-tem from the globes of a 51-yr-old man. Nuclear andcortical lipid membranes were prepared, and lipidswere extracted from these membranes in CHC13.

14 Ap-proximately 400 txL of this lipid extract in CHC13 waslayered on a 13-mm diameter AgCl window in a glove

No. 8 HUMAN LENS LIPID IR SPECTROSCOPY / Dorchmon er ol 2405

box with an atmosphere of argon. CHC13 was evapo-rated by a stream of argon. The dry lipid was lyophi-lized on the AgCl window for 12 hr to remove residualCHC13 and water. Complete removal of the CHC13

was evident from the infrared spectrum of the driedlipid. Due to CHC1, no band at 3020 cm"1

3 could bedetected in our preparations. The lipid was dispersedin 10 ML of 100-mM KC1, 5 mM 4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid HEPES buffer, pH7.4 was added to the AgCl window. Control studieswith bovine brain sphingomyelin showed thatHEPES does not alter the thermodynamic parametersof the phase transition. Sphingomyelin is the majorphospholipid of the lens.12 A 0.10-mm Teflon spacerand a second AgCl window were added to completethe sample holder. The sample holder was then placedin a Fenwal variable temperature cell obtained fromWilmadglass Company Inc., Buena, NJ. We con-trolled the temperature to ± 0.1 °C during spectro-scopic measurement. The sample cell was equili-brated at 70 °C for 1 hr to ensure complete hydrationof the lipid. Our samples contained eight times morewater than was necessary to assure complete hydra-tion.37 The sample was cooled from 70°C to -2.5°C;then heated from -2.5°C to 70°C. Infrared spectrawere measured every 2°C. The sample was equili-brated for 5 min before spectral acquisition that re-quired 2 min. Temperature was changed at a rate of0.1 °C/min to avoid hysteresis that is caused by supercooling. Hysteresis, in this case, is the lagging behindof values, such as fluidity, when a sample is heated orcooled too quickly. Hysteresis, due to phase separa-tion caused by sphingomyelin-cholesterol interac-tions that last for days,38 is unavoidable, and fortu-nately, is inconsequential in our system. Lipid disper-sion spectra were measured with 2.0 cm"1 spectralresolution by a Nicolet SX-176 FTIR spectrometer(Cincinnati, OH). Each spectrum averaged 100 scans.Absorbance (or transmittance) values were displayedand digitally stored for further mathematic manipula-tion (background subtraction). Before the frequencyof the CH2 stretching bands was measured, the base-line was corrected for the contribution of the waterO-H stretching band that was centered near 3200cm"1. The baseline value for the CH2 stretching bandswas taken at 3100 cm"1 and 2700 cm"1.

Statistical Evaluation of the Lipid Phase Transition

Parameters, such as transition temperature and rela-tive cooperativity, are often difficult to measure byvisual inspection of the transition curves of the lipidphase. Furthermore, it is difficult to assess the preci-sion and accuracy with which those parameters aremeasured by visual inspection. To circumvent theseproblems, a nonlinear regression analysis program,

Frequency = P, + (P / [ 1 + (P3/Temp)P4]

TEMPERATURE (°K)

Fig. 1. Relationship between parameters P,, P2, P3, and P4 andthe shape of a lipid phase transition curve. A nonlinear regressionanalysis program, statistical analysis software, was used to evaluatethe parameters from experimental phase transition data pairs (vi-brational frequency in wave numbers, temperature in °K).

Statistical Analysis Software (SAS), was used to evalu-ate the parameter P,, P2, P3, and P4 from experimentalphase transition data pairs (vibrational frequency inwavenumbers, temperature in °K). The function usedin the fitting program is:

Wavenumber = P, + [P2/(l + (P3/T)P4] (1)

The relationship among the parameters P,, P2, P3,and P4 (denned below) and the shape of the phasetransition curve is shown in Figure 1. P, is the mini-mum wavenumber for the phase transition, and inour study, it is the most ordered state of the transition.P2, the magnitude of the phase transition, is the netchange in wavenumbers and represents the change inthe number of trans-to-gauche rotamers (Figs. 1 and2). P3 is the transition temperature and shows the tem-perature at which half of the population of lipid mole-cules have undergone the phase transition. P4 is therelative cooperativity and describes the degree bywhich one lipid affects the state of order of adjacentlipids.39 The symmetric stretching band frequencywas measured with a precision of ± 0.06 cm"1. Thisvalue was determined from the standard deviation ofthe wavenumber values on the flat portion of thephase transition curve. A phase transition of 0.12cm"1 or more would be statistically significant, P< 0.01, assuming that six data points were measuredbefore and after the phase transition.

The relative amount of gauche rotamers at a giventemperature is indicative of the level of disorder of alipid bilayer.41*42 For a completely ordered system (0%gauche rotamers), such as a dispersion of dipalmitoyl-phosphatidylcholine at -20°C, the frequency for theCH2 symmetric stretch was reported to be 2849.0

2406 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / July 1991 Vol. 32

ROTOMERS

:"H HUCCH,

C >

TRANS GAUCHE

Fig. 2. Molecules differ by rotation around single bonds and aredefined as rotamers. Rotational isomers, conformational isomers,and conformers are synonomous with rotamers when the carbonatoms are in the same plane and when the methyl groups form a180° dihedral angle when viewed along the bond axis. This confor-mation is termed the anti conformation by organic chemists,40 butnot by lipid spectroscopists. When methyl groups are in any otherconformation, they are referred to as gauche rotamers.

cm '. When the same model system was completelydisordered (100% rotamers) by solubilizing the lipidin CHC13, the CH2 band frequency was measured at2854.5 cm"1. Based on the considerations above, thepercentage of gauche rotamers was estimated usingthe following equation:

% gauche rotamers

(CH2 symmetry stretch frequency)- 2849.0 cm"1

(2854.5 - 2849.0)cm"(2)

A similar expression can be used for the estimationof percentage gauche rotamers that use the frequencythat is measured for the CH2 asymmetric stretch. Fora completely ordered and disordered model system,the band frequency was reported to be 2917.0 cm"1

and 2927.0 cm"1, respectively.

Results

Infrared spectroscopy was used in this investigationto characterize the fatty acyl chain and the interfaceregions of lens membranes. In this study, the CH2

infrared stretching bands were used to quantify transand gauche acyl chain isomerization (Fig. 2, descrip-tion of rotamers). When lipids are in an ordered state(gel phase), the fatty acyl carbons are in an all transconformation; van der Waals interactions and bilayerpacking are maximized (Fig. 3). In the disordered liq-uid-crystalline phase, a number of C-C gauche ro-tamers are introduced. Van der Waals interactions

and bilayer packing are diminished. Among many in-frared spectral changes, the frequency of the CH2

stretching bands near 2850 and 2920 cm"1 becomeshigher as the number of gauche rotamers increases.The rotamers increase the disorder of the lipid mem-brane.41'42

After the temperature is increased, many lipid bi-layer membranes undergo a change from a gel to aliquid-crystalline phase. In this study, gel to liquid-crystalline phase transitions were carefully monitoredwith the use of infrared spectroscopy. From the phasetransition curves, transition temperatures (tempera-ture at which half of the population of lipid moleculeshave undergone the transition) and relative lipid co-operativity (degree to which one lipid affects the stateof order of adjacent lipids) were determined.

Figure 4 shows a typical spectrum of the CH2

stretching region of a lipid vesicle dispersion; the dis-persed lipids were extracted from membranes thatwere prepared from the nucleus of a human lens.Three bands are discernible near 2850, 2920, and2980 cm"1, and correspond to the CH2 symmetricstretch, asymmetric stretch, and CH3 symmetric ter-minal methyl stretching band, respectively. A doublebond =C-H stretching band near 3035 cm"1 was notdetected and shows that lens nuclear lipids are satu-rated.

The CH2 symmetric and asymmetric stretchingband frequencies are shown to be dependent on tem-perature. The higher vibrational frequencies seen athigher temperatures show more fatty acyl chaingauche rotamers and, thus, a more disordered mem-brane. A phase transition from the gel to the liquid-crystalline phase is noticeable in curves plotted withthe symmetric and asymmetric stretching band fre-quency (Figs. 5 and 6). We believe that this transitionis a single broad transition rather than a collectivetransition due to several components for two reasons.First, we measured the phase transition for pure bo-vine brain sphingomyelin and found it to be broadand greater than 35°C from beginning to end (P4

= 40), unlike the transition for pure phosphatidylcho-line membranes, which shows a sharp phase transi-tion. Sphingomyelin is the major phospholipid in lensmembranes.12 Second, cholesterol, which is found inhigh concentration in lens membranes,14 would be ex-pected to broaden the phase transition.

The nonlinear regression program described in Ma-terials and Methods was used in the calculations toimprove the precision in the determination of thephase transition temperature, the magnitude, relativecooperativity, and minimum vibrational frequency ofthe transition. The values obtained for each parame-ter are shown in Table 1, and the statistical fit is plot-ted in Figure 7. The symmetric CH2 stretching bandat 2850 cm"1 provides a better estimate for the phase

No. 8 HUMAN LENS LIPID IR SPECTROSCOPY / Dorchmon er ol 2407

Headgroup

Interface

Acyl-Chain §

head head head

ORDERED DISORDERED

Fig. 3. Schematic diagram of the acyl chain region of bilayer membrane lipids in the ordered (gel) and disordered (liquid crystalline) phases.

INFRARED SPECTRUM OF NUCLEAR LIPID

3100 3000 2900 2800

WAVENUMBER (CM1)

2700

Fig. 4. Infrared spectrum of the CH2 stretching region for a lipiddispersion from the nucleus of a human lens. No smoothing wasused. The spectrum was corrected for the broad water background.(See Materials and Methods)

transition parameters than the 2920 cm ' antisym-metric CH2 stretching band because the 2850 cm"1

band is farthest removed from the broad H-O-H sym-metric stretching water band centered at 3200 cm"1

and from adjacent CH2 stretching bands that can af-fect the accuracy and precision with which the fre-quency at maximum peak height is calculated. Theparameters calculated with the 2920 cm"1 band wereincluded to confirm the trends of those calculatedwith the more accurate 2850 cm"1 band.

Minimum Wavenumber P, and Magnitude of PhaseTransition P2

The magnitude of the gel to liquid-crystalline phasetransition temperature (P2) was measured for the cor-tical and nuclear lipid vesicle dispersions (Table 1).Phase transition magnitudes of 3.9 and 4.2 cm"1 weremeasured for cortical and nuclear lipid, respectively,with the 2850 cm"1 CH2 symmetric stretching band.Calculations based on equation 2 indicated that thelipid progresses from an almost completely orderedstate (<15% gauche rotamers) to an almost com-pletely disordered state (100% gauche rotamer). The

2408 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / July 1991 Vol. 32

NUCLEAR LIPID PHASE TRANSITIONS

TRIAL 1 TRIAL 2

1

Eo^ ^

JMB

ER

l

LJ

2855

2854

2853

2852

2851

2850

2849

2848

I

Eo—'

aJGO

2854

2853

2852

2851

2850

2849

9R4fl

/

f

-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

TRIAL 3 TRIAL 4

Eu

toorUJ

m

LJ

IEo

«lfc •

JMB

ER

i

LJ

2855

2854

2853

2852

2851

2850

2849

10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

2848-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

Fig. 5. Change in the frequency of the CH2 symmetric stretching band with temperature for a nuclear lipid dispersion prepared from a5 l-yr-old lens. (O - O) cooling curve, (A - A) heating curve. Higher wave number values indicate a more disordered membrane.

magnitudes of the transitions were also measured forthe CH2 asymmetric band and were 5.5 and 7.7 cm"1

for the cortex and nucleus, respectively. These magni-tudes were also indicative of an almost maximumchange in the order of the lipid system, in accordancewith the results above. Because all of the lipids wereinvolved in the phase transition, our sample was com-pletely hydrated.

Values of 2849.15 ±0.11 cm"1 (n = 8) and 2849.8± 0.08 cm"1 (n = 8) for the minimum wavenumber,Pl5 of the nuclear and cortex transitions, respectively,also confirmed that at temperatures lower than theonset of the phase transition, the lipids are found in analmost completely ordered state.

Phase Transition Temperature P3

Slight hysteresis was seen in the nuclear phase tran-sition curves as shown by different heating and cool-ing phase transition temperatures, 42.6 ± 0.8°C and39.6 ± 1.6°C, respectively (Table 1). A similar trendwas seen in the phase transition curve that was ob-tained with the CH2 asymmetric stretching band.However, the temperature P3 measured from the CH2

asymmetric stretch band, was 4.4°C lower than thatmeasured from the CH2 symmetric stretch band. Thisdifference was attributed to the lower accuracy of thefrequency measurements of the CH2 asymmetricstretching band.

No. 8 HUMAN LENS LIPID IR SPECTROSCOPY / Borchmon er ol 2409

NUCLEAR LIPID PHASE TRANSITIONS

TRIAL 1 T R | A L 2

EuocUJen

2928

2926

2924

2922

2920

2918

2916-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

TRIAL 32928

2926

2924

2922

2920

2918

2916-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

2928

2926

2924

2922

2920

2918

2916-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

TRIAL 42928

2926

2924

2922

2920

2918

2916-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

Fig. 6. Change in the frequency of the CH2 asymmetric stretching band with temperature for a nuclear lipid dispersion prepared from a51-yr-old lens. (O - O) cooling curve, (A - A) heating curve. Higher wave number values indicate a more disordered membrane.

The cortical lipid phase transition temperature(Fig. 8) evaluated with the 2850 cm"1 band was about12°C lower than that of the nuclear lipids (Table 1).Hysteresis was less pronounced in the phase transi-

Table 1. Lipid phase transition parameters

tion curves for the cortical lipid. Phase transition tem-peratures of 29.9°C and 28.9°C were measured fromthe heating and cooling curves, respectively (Table 1).The temperatures that were calculated with the less

Parameter

P,, minimum of transition (cm"1)

P2, magnitude of transition (cm"1)

P3, phase transition temperature (°C)Increasing temperatureP3, phase transition temperature (°C)Decreasing temperature

2850 Band

2849.8 ± 0.08(3)

3.9 ± 0.2(3)

29.9(1)

28.9(1)

Cortex

2920 Band

2919.43 ±0.51(3)

5.5 ±0.1(2)

29.9(1)18.9(1)

Nucleus

2850 Band

2849.15 ±0.11(8)

4.2 ±0.2(8)

42.6 ±0.8(4)

39.6 ± 1.6(4)

2920 Band

2917.9 ±0.16(8)

7.7 ± 0.2(8)

38.1 ± 1.0(4)

35.2 ± 1.2(4)

2410 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / July 1991 Vol. 32

STATISTICAL FITA) B)

2855

2854 •

Ifo

JUM

BE

I/A

VE

h

•5-

2853

2852

2851

2850

2849

2848

-

1//

-

2928

2916-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

-10 0 10 20 30 40 50 60 70 8C

TEMPERATURE (°C)

Fig. 7. Computed phase transition curves from experimental data in Figures 4 and 5. Heating ( ), cooling ( ) study for symmetric (A)and asymmetric (B) stretching bands.

accurate 2920 cm"1 band were 29.9° and 18.9°C forthe heating and cooling cycles, respectively.

Relative Cooperativity P4

Relative cooperativity of the phase transition wasdetermined by evaluating parameter P4 (Materialsand Methods). A high cooperativity shows that whenone lipid melts, the adjacent lipids melt more easily.The cooperativity that was measured for the nuclearlipid transitions was 1.7 times higher than that mea-sured for the cortical transitions, (Fig. 9). The sameratio was obtained with the symmetric and asymmet-ric bands.

Acyl Chain Order at 37°C

The percentage of gauche rotamers (disorder)showed that the fluidity of the cortex was about 63%of the maximum fluidity and 2-3 times larger thanthe fluidity of the nuclear lipid (Fig. 10). This percent-age indicated that at physiologic temperature, the cor-tical lipids are more disordered than the nuclear lipidsat the fatty acyl chain level.

The van't Hoff Enthalpy

The amount of heat that was required for gel toliquid-crystalline phase transitions was twice as large

for the nuclear lipid vesicle dispersion as comparedwith that for the cortical dispersion (Table 2). To esti-mate the van't Hoff enthalpy, Arrhenius plots, ie, In(fraction gauche) vs 1/T were constructed (Fig. 11)and the slopes evaluated. The fraction of gauche ro-tamers at 1 °C below and above the phase transitiontemperature, P3, was estimated by interpolation ofequation 2. The correlation coefficient for 11 datapoints was 0.919. The entropy (randomness) of the gelto liquid-crystalline transition was about twice aslarge for the nuclear compared with the cortical lipidvesicle dispersion (Table 2). The entropy values werecalculated by dividing the enthalpy values in Table 2by the phase transition temperature (AS = H/P3 atequilibrium conditions, AG = 0):

Acyl Chain Packing

A doublet at 1473 cm"1 and 1462 cm"1 was seen forthe CH2 scissoring band when lens lipids were in thegel phase (Fig. 12). A doublet is indicative of an acylchain orthorhombic or monoclinic crystal lattice.41 Inthe liquid-crystalline phase, the doublet becomes asinglet at 1467 cm"1 and is indicative of a hexagonalcrystal lattice.42 Sharp narrow bands at less than 1462cm"1 were found in all spectra at the same location.Their origin is unknown.

No. 8 HUMAN LENS LIPID IR SPECTROSCOPY / Borchman er ol 2411

A)

oa:UJ

m

ZUJ

2855

2854 -

2853 •

2852 -

2851

2850 -

2849

CORTICAL LIPID PHASE TRANSITION

B)2926

2916-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

-10 0 10 20 30 40 50 60 70 80

TEMPERATURE (°C)

Fig. 8. Change in the frequency of the cortical lipid CH2 symmetric (A), and asymmetric (B), CH2 stretching band for cooling (O - O) andheating (A - A) studies.

Interface Region Structure

The carbonyl band near 1740 cm"1 (Fig. 13-15) isuseful for the structural determination of the bilayerinterface region. The frequency of this band was con-stant at 1742 cm"1 for cortical lipids in the gel or liq-

RELATIVE COOPERATIVITY OFOF LIPID PHASE TRANSITION

uid-crystalline phase (Fig. 13). A carbonyl band fre-quency of 1742 cm"1 for the lens lipid membranesshowed an acyl chain C,-C2 bond dihedral angle thatarranges the acyl chains in an energetically favoredtilted position.43

The carbonyl band width was measured to be 19.7-21.6 cm"1 for the lens lipids. For comparison, the car-bonyl bandwidth for dipalmitoyl-phosphatidylcho-line is 31-34 cm"1.44 The narrow band width for lens

80

CHLJLJ

Q_OOOLd

LJ

70 [-

60

50

40

30

20

10

0

•

I h

he

atin

a I

I ^

coo

ling

ILIPID ORDER PARAMETER

CORTEX NUCLEUS

oo

ro

!<rrLJ

ooQ:11 1LLJ

Xo

o

80

70

60

50

40

30

20

10

0

%c

oQ)

SI

1

c"ooo

iFLUID

RIGID

Fig. 9. Relative cooperativity estimated from lipid phase transi-tion. Curves for cortical (n = 3) and nuclear lipids (n = 8). Estimateswere made using the symmetric and asymmetric CH2 stretchingbands.

CORTEX NUCLEUS

Fig. 10. Lipid order parameters were determined from the fre-quency of the symmetric CH2 stretching band. (See Materials andMethods). For nuclear lipids, n = 4 and for cortical lipids, n = 2.

2412 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / July 1991 Vol. 32

.00301/T (°K)

.0033

Fig. 11. A typical Arrhenius plot where K is the percentagegauche rotamer and T is given in °K. Circles indicate a heatingstudy; triangles indicate a cooling study.

lipids showed restricted freedom of motion in the in-terface region. Cholesterol, found in excess in thelens,14 has been shown to lower the carbonyl band-width.44

Discussion

With Fourier transform infrared spectroscopy, wefound lipid acyl chain and interface structure in mem-branes that were obtained from the cortical and nu-clear regions of a normal human lens to be consider-ably different. Temperature-dependent studies, ie,phase transitions, were performed to evaluate differ-ences in the thermodynamic properties of the nuclearand cortical lipids. A relevant aspect in the study oflipids is the determination of the order of the system.Previously reported differences in lens membranelipid structures had been inferred from lipid order pa-rameters that were obtained with electron spin reso-nance5 and fluorescence polarization6-7 probes. Dueto the high molar ratio of cholesterol to phospholipid(>2:1) in lens lipid membranes, fluorescence polariza-tion probes may not be successfully applied to deter-mine lipid order. Van Bleitterseurjk et al8 believe thatat cholesterol/lipid molar ratios that exceed 2:1, cho-lesterol clusters are formed and that probes can nolonger sense the lipid order because they may be ex-cluded from these clusters. Another possibility is thatthe probe may be drastically perturbed by an adjacentcholesterol molecule, and consequently, it could nolonger sense the order of other adjacent lipids.9 Toavoid these problems and to obtain structural infor-

mation directly, x-ray, infrared, and Raman scatter-ing spectroscopy can be used.

Our infrared results show that the structure of thefatty acyl chain region shows a higher degree of orderfor the nuclear lipid membranes as compared with thecortical ones. This result is evidenced by the highertransition temperatures (Table 1), cooperativity (Fig.9), and values for the enthalpy and entropy (Table 2)that are measured from the phase transition curves.At physiologic temperature, the fluidity of the cortexis maintained at about 60% of maximum fluidity (Fig.10). Ca-ATPase has been shown to be sensitive tolipid acyl chain fluidity. The cortical lipid environ-ment meets the fluidity requirement for the functionof the Ca-ATPase pump. Ca-ATPase has been shownto be sensitive to lipid acyl chain fluidity. No Ca-AT-Pase activity is detected in the nuclear region,33 per-haps because of the higher lipid acyl chain order of thenuclear membranes.

Figures 5,6, and 8 show a single-phase transition inwhich cortical and nuclear lens lipid membranes un-dergo a complete change from a very ordered liquid-crystalline phase to a highly disordered, more fluid,gel state. Multiple-phase transitions, such as thoseseen in the lipids of the sarcoplasmic reticulum, showlateral-phase separation45 that is understood as theformation of clusters of fluid lipids within an orderedmatrix. Membranes that show a single transition,such as lens membranes, do not show lateral-phaseseparation and are therefore less permeable thanthose that are phase separated.39'46"48

As shown by the values of P, and P2, the acyl chainsof the cortical and nuclear lipids undergo the samechange in the fraction of gauche rotamers during aphase transition, ie, <15% ordered state to 100% dis-ordered state. Although the initial and final levels oforder in the acyl chains are identical for cortical andnuclear lipids, the transition temperatures, P3, andchanges in enthalpy and entropy show that more en-ergy is required to fluidize the nuclear membrane, ie,increase the fraction of gauche rotamers, as comparedwith the cortical membrane. These energetic differ-ences between the cortical and nuclear membranes

Table 2. Lipid phase transition thermodynamicparameters

Parameter

AH Van Hoff(Kcal/mol)

AS (cal/mol°K)

2850Band

6.9 ± 0(2)

23 ± 2(2)

Cortex

2920Band

.7 4.9 ± 0.8(2)

16 ± 3(2)

Nucleus

2850Band

18±2(7)

57 ±6(7)

2920Band

12 ± 1(7)

38 ± 4(7)

No. 8 HUMAN LENS LIPID IR SPECTROSCOPY / Dorchmon er ol 2413

Fig. 12. The CH2 scissoring band fornuclear lipid dispersions in the gelphase ( ), and in the liquid crystal-line phase ( ). The lower curve isthe difference spectrum.

LUoz<CDCCOCOGO

KM 473

1500 1480 1460 1440 1420 1400 1380

WAVENUMBERS (cm1)

may arise from differences in the level of hydration orin the packing at the interface region. This possibilityis supported by changes in the contour of the carbonylband near 1743 cm"1 as discussed in the precedingparagraph. The small changes in the interface regionand possibly stronger interactions in the head groupregion, not well characterized at this time, may alsoaccount for the higher cooperativity values, P4, calcu-lated for the nuclear membranes.

No change in the shape of carbonyl band with tem-perature exists for the cortical lipid (Fig. 13). This find-ing indicates that the energetically favorable interfaceregion packing remains constant, whereas the acylchain region of the bilayer undergoes isomerization

and a change between orthorhombic or monoclinicand hexagonal packing.

The nuclear lipid carbonyl band in the liquid-crys-talline phase was identical to the cortical lipid band inthe gel and liquid-crystalline phase. Only the profileof the nuclear lipid carbonyl band in the gel phase wasdifferent (Figs. 14 and 15). This band showed a profilewith reduced intensity at 1745 cm"1 and indicatedthat the nuclear lipid membrane underwent a changein interface region packing during a gel to liquid-crys-talline phase transition. The exclusion of water mightresult. The difference in the interface region hydra-tion could account for the larger enthalpy and en-tropy in the nuclear-lipid phase transition (Table 2).

2414 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / July 1991 Vol. 32

UJ

Oz<CQCCOCOCQ

1690 1710 1730 1750 1770

WAVENUMBERS (cm1)Fig. 13. Carbonyl band spectra for cortical lipid dispersions in the

gel phase ( ), and in the liquid crystalline phase ( ).

The high degree of order in lens membranes com-pared with membranes from other tissues, generallyfluid at physiologic temperatures , mus t result from

composit ional differences. The most striking compo-sitional difference is the cholesterol content , which isabout four t imes greater in the lens than in other tis-sues. Within the lens, the nuclear contains about one-third more cholesterol than the cortex. In modelm e m b r a n e systems, an increase in the concentrat ionof cholesterol lowers the transit ion temperature , co-operativity, and changes in enthalpy and entropy.4 9

Unless cholesterol interacts differently with unknownlipids in the lens,50 the energetic differences cannot beattr ibuted to differences in cholesterol content . Re-gional changes in the extent of oxidative damage,5 1 '5 2

head group composit ion, acyl chain length, and acylchain saturation5 3 could account for differences in thephase transition parameters that are measured for thenuclear and cortical lipids.

There is no evidence that acyl-chain length and /o rsaturation or head group composi t ion are dramat i -cally different in the h u m a n cortex or nuclear. Thelack of a = C stretching band at 3010 cm" 1 shows tha tlipids from both of these regions are saturated andcontain less than an average of two double bonds perlipid molecule. Oxidative damage may not explainthese differences because oxidative products t end tofluidize ordered model membranes5 1 ' 5 4 and the nu-clear m e m b r a n e shows a higher degree of oxidative

LUOz<CQCCOCOCD

Fig. 14. Carbonyl band spectra for nu-clear lipid dispersions in the gel phase( ), and in the liquid crystalline phase( ). The lower curve is the differencespectrum.

1690 1710 1730 1750

WAVENUMBERS (cm1)

1770

No. 8 HUMAN LENS LIPID IR SPECTROSCOPY / Dorchmon er ol 2415

Fig. 15. Carbonyl band spectra for anuclear lipid dispersion ( ), and acortical lipid dispersion ( ) in the gelphase. The lower curve is the differencespectrum.

1690 1710 1730 1750

WAVENUMBERS (cm1)

1770

damage.52'5556 The correlation between differences inmembrane composition and structure is still unclear.

We showed the applicability of FTIR spectroscopyto measure regional structural differences in lens lipidmembranes. Studies are currently being performed todetermine regional differences in lipid structure withage and cataractogenesis.

Key words: infrared spectroscopy, lipids, membrane struc-ture, crystalline lens, human

References

1. Deuticke B and Haest WM: Lipid modulation of transport pro-teins in vertebrate cell membranes. Ann Rev Physiol 49:221,1987.

2. Strittmatter W, Hirata F, and Axelrod J: Biochem Biophys ResCommun 88:147, 1979.

3. Yeagle PL: Lipid regulation of cell membrane structure andfunction. FASEB J 3:1833, 1989.

4. Anner BM: Interaction of (Na+ + K+) -ATPase with artificialmembranes. I. Formation and structure of (Na+ + K.+) -ATP-ase-liposomes. Biochim Biophys Acta 822:319, 1985.

5. Puskin JS and Wiese MB: A spin label study of human lensmembranes. Exp Eye Res 35:251, 1982.

6. Gooden M, Takemoto L, and Rintoul D: Evidence for reducedlipid order in plasma membranes from cataractous humanlenses. Curr Eye Res 2:367, 1983.

7. Li LK, Spector A, Logan U, and Schacter D: Dynamics of lipidfluorophores in bovine lens fiber membranes. Exp Eye Res34:145, 1982.

8. Van Blitterswijk WJ, Van Der Meer BW, and Hilkmann H:Quantitative Contributions of Cholesterol and the IndividualClasses of Phospholipids and their Degree of Fatty Acyl (Un)Saturation to Membrane Fluidity Measured by FluorescencePolarization. Biochemistry 26:1746, 1987.

9. Hoffmann W, Pink DA, Restall C, and Chapman D: Intrinsicmolecules in fluid phospholipids bilayers. Eur J Biochem114:585, 1981.

10. Feldman GL: Lipids of the Human Lens. In Biochemistry ofthe Eye, Dardenne MU and Nordmann J, editors. Symp Tutz-ing Castle Basel/New York: Karger, 1966, pp. 348-357.

11. Andrews J and Leonard-Martin T: Total lipid and membranelipid analysis of normal animal and human lenses. Invest Oph-thalmol Vis Sci 21:39, 1981.

12. Broekhuyse RM, Kuhlmann ED, and Jap PHK: Lens mem-branes IX. Some characteristics of fiber membranes in relationto ageing and cataract formation. Ophthalmic Res 11:423,1979.

13. Broekhuyse RM: Phospholipids in tissues of the eye III. Com-position and metabolism of phospholipids in human lens inrelation to age and cataract formation. Biochim Biophys Acta187:354, 1969.

14. Borchman D, Delamere NA, McCauley LA, and Paterson CA:Studies on the distribution of cholesterol, phospholipid andprotein in the human and bovine lens. Lens and Eye ToxicityResearch 6:703, 1989.

2416 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / July 1991 Vol. 32

15. Bakx T, Ringens PJ, van den Oetelaar PJM, Muskiet FAJ, Jan-sen G, and de Jong JMBV: Accumulation of cholesterol sul-phate in a human cataractous lens. Lens Res 4:277, 1987.

16. Broekhuyse RM, Roelfzema H, Breimer ME, and KarlssonKA: Lipids in tissues of the eye X. Molecular species of sphin-gomyelins from different parts of calf lens in relation to differ-entiation and aging. Exp Eye Res 19:477, 1974.

17. Feldman GL, Culp TW, Feldman LS, Grantham CK, andJonsson HT Jr: Phospholipids of the bovine, rabbit and humanlens. Invest Ophthalmol Vis Sci 3:194, 1964.

18. Roelfzema H, Broekhuyse RM, and Veerkamp JH: Lipids intissues of the eye XII. Phospholipid and sphingomyelin con-centrations in bovine lenses in relation to differentiation andageing. Exp Eye Res 23:409, 1976.

19. Feldman GL and Feldman LS: New concepts of human lenticu-lar lipids and their possible role in cataracts. Invest OphthalmolVis Sci 4:162, 1965.

20. Zelenka PS: Lens lipids. Curr Eye Res 3:1337, 1984.21. Varma SD and Reddy VN: Phospholipid composition of

aqueous humor, plasma and lens in normal and alloxan dia-betic rabbits. Exp Eye Res 13:120, 1972.

22. Rosenfeld L and Abraham S: Changes in lipid distribution inthe human lens with the development of cataract. Exp Eye Res33:641, 1981.

23. Cenedella RJ: Regional distribution of sterol and fatty acidSynthesis in the ocular lens. Exp Eye Res 38:95, 1984.

24. Tao RVP and Cotlier E: Ceramides of human normal and cata-ractous lens. Biochim Biophys Acta 409:329, 1975.

25. Cotlier E, Obara Y, and Toftness B: Cholesterol and phospho-lipids in protein fractions of human lens and senile cataract.Biochim Biophys Acta, 530:267, 1978.

26. Li L-K, So L, and Spector A: Membrane cholesterol and phos-pholipid in consecutive concentric sections of human lenses.Lipid Res 26:600, 1985.

27. Zigman S, Paxhia T, Marinetta G, and Girsch S: Lipids ofhuman lens fiber cell membranes. Curr Eye Res 3:887, 1984.

28. Gupta JD and Harley GD: Decreased adenosine triphosphateactivity in human senile cataractous lenses. Exp Eye Res20:207, 1975.

29. Nordmann J and Klethi J: L'activite Na-K-ATPase dans lecrystallin normal viellissant et dans. Arch Ophthalmol (Paris)36:523, 1976.

30. Auricchio A, Rinaldi E, Savastano S, Albini L, Curto A, andLaudolfo V: The Na-K-ATPase in relation to the Na, K andtaurine levels in senile cataract. Metabolic Pediat Syst Ophthal-mol 4:15, 1980.

31. Friedburg D: Enzyme activity patterns in clear human lensesand in different types of senile cataract. The Human Lens inRelation to Cataract. Ciba Found Symp 19:117, 1973.

32. Paterson CA, Delamere NA, Mauhotler L, and Culzon JV: Na,K-ATPase in stimulated eye bank and cryo extracted rabbitlenses, and human eye bank lenses and cataracts. Invest Oph-thalmol Vis Sci 24:1534, 1983.

33. Borchman D, Paterson CA, and Delamere NA: Ca2+-ATPaseactivity in the human lens. Curr Eye Res 8:1049, 1989.

34. Gandolfi SA, Tomba CM, and Maraini G: 86-Rb Efflux innormal and cataractous human lenses. Curr Eye Res 4:753,1985.

35. Pasino M and Maraini G: Cation pump activity and mem-brane permeability in human senile cataractous lenses. ExpEye Res 34:887, 1982.

36. Duncan G and Jacob TJ: Human cataract formation CibaFound Symp 106:132, 1984.

37. Shipley GG, Avecilla LS, and Small J: Phase behavior andstructure of aqueous dispersions of sphingomyelin. J Lipid Res15:124, 1974.

38. Yeagle PL: Cholesterol and the cell membrane. BiochimBiophys Acta 822:267, 1985.

39. Marsh D, Watts A, and Knowles PF: Evidence for phaseboundry lipid. Permeability of TEMPO-choline into dimyris-toylphosphatidylcholine vesicles at the phase transition. Bio-chemistry 15:3570, 1976.

40. Stereochemistry. Allinger NL, Cava MP, de Jongh DC, John-son CR, Label NA, and Stevens CL, editors. In Organic Chem-istry, Worth Publishers Inc., 1972, pp. 105-126.

41. Cameron DG, Casal HL, Gudgin EF, and Mantsch HH: Thegel phase of dipalmitoyl phosphatidylcholine and infrared char-acterization of the acyl chain packing. Biochim Biophys Acta596:463, 1980.

42. Mantsch HH, Cameron DG, Umemura J, and Casal JL: Four-ier transform infrared spectroscopy of aqueous systems: appli-cations to the study of biological membranes. J Mol Struct60:263, 1980.

43. Bicknell-Brown E, Brown, and Person WB: Configuration-de-pendent Raman bands of phospholipid surfaces. J Am ChemSoc 102:5486, 1980.

44. Umemura J, Cumeron DG, and Mantsch HH: A Fourier trans-form infrared spectroscopic study of the molecular interactionof cholesterol with 1,2 dipalmitoyl-sn-glycero-3-phosphocho-line. Biochim Biophys Acta 602:32, 1980.

45. Borchman D: Infrared, Raman and compositional studies ofskeletal muscle sarcoplasmic reticulum. Doctoral Thesis,Wayne State University, Detroit, MI, 1986.

46. Blok MC, van der Newt-Kok EMC, van Deenen LLM, and deGier J: The effect of chain length and lipid phase transitions onthe selective permeability properties of liposomes. BiophysActa 406:187, 1975.

47. Papahadjopoulous D, Jacobson K, Nir S, and Isac T: Fluores-cence polarization and permeability measurements concerningthe effect of temperature and cholesterol. Biochim BiophysActa 311:330, 1973.

48. Noordam PC, Killian A, oude Elferink RFM, and de Gier J:Comparative study on the properties of saturated phosphatidy-lethanolamine and phosphatidylcholine bilayers: barrier char-acteristics and susceptibility to phospholipase A2 degradation.Chem Phys Lipids 31:191,1982.

49. Jain MK: Role of cholesterol in biomembranes and relatedsystems. In Current Topics in Membrane and Transport 9,Kleinyeller A, editor. New York, Academic Press, 1977, pp.1-57.

50. Merchant TE, Lass JH, Meneses P, Greiner JV, and Glonek T:Human crystalline lens phospholipid analysis with age. InvestOphthalmol Vis Sci (Suppl)30:95, 1989.

51. Lamba OP, Yappert MC, Lou MF, and Borchman D: Compar-ative infrared studies of structural features in cataractous lensmembranes and in model membrane deranged by oxidativedamage. Invest Ophthalmol Vis Sci (Suppl)31:163, 1990.

52. Borchman D, Yappert MC, Rubini RQ, and Paterson CA: Dis-tribution of phospholipid-malondialdehyde-adduct in the hu-man lens. Curr Eye Res 8:939, 1989.

53. Oldfield E and Chapman D: Dynamics of lipids in membranes.Heterogeneity and role of cholesterol. Fed Eur Biochem SocLett 23:285, 1972.

54. Lamba OP, Lai S, Yappert MC, Lou MF, and Borchman D:Spectroscopic detection of lipid in a sphingomyelin model sys-tem. Biochim Biophys Acta (in press), 1990.

55. Bhuyan KC, Bhuyan DK, and Podos SM: Lipid peroxidationin cataract of the human. Life Sci 38:1463, 1986.

56. Bhuyan KC, Master RWP, Coles RS, and Bhuyan DK: Molecu-lar mechanisms of cataractogenesis IV. Evidence of phospho-lipid malondialdehyde adduct in human senile cataract. MechAgeing Dev 34:289, 1986.