Regional and Age-Dependent Differences in the
Phospholipid Composition of Human Lens Membranes
Douglas Borchman,^ W. CraigByrdwell* and M. Cecilia Yappert*
Purpose. The long-term purpose of this research was to establish the relationships between
composition, structure, and function that affect human lens membranes. The authors hypothesized that the functional differences of epithelial, cortical, and nuclear lens membranes are
related to compositional differences. Furthermore, age-dependent alterations in membrane
function and structure can also be related to variations in the phospholipid composition. To
explore these possibilities, the authors determined the phospholipid composition of epithelial,
cortical, and nuclear membranes from pools of human lenses of different ages.
Methods. Membranes were extracted from pools of clear human lenses of different ages
using a monophasic methanolic extraction that minimizes the interfacial fluff produced with
biphasic extractions. The phospholipid composition was determined by 31P NMR.
Results. Only minor differences were detected between cortical and nuclear fractions. All
phospholipids, except sphingomyelin, phosphatidylethanolamine, and the phospholipid with
a shift of 0.12 parts per million (ppm) in the 31P NMR spectrum, showed significant differences
in the epithelial fractions of all age groups compared to the fiber fractions; the percentage
of phosphatidylcholine was considerably higher than that in the cortical and nuclear membranes of the same age. Conversely, the percentage of phosphatidylglycerol and lysophosphatidylglycerol was significantly smaller in the epithelial membranes than in the fiber membranes.
The age-related changes in the composition of cortical and nuclear membranes were identical.
These membranes showed a steady increase with age in the percentage of sphingomyelin and
of an unidentified component with a shift of 1.2 ppm. The percentage of phosphatidylcholine
decreased with age in both epithelial and fiber membranes. The rate of decrease was greater
in the epithelial membranes than in the fiber membranes. Epithelial membranes contained
approximately five times more phosphatidylcholine than fiber membranes of corresponding
age.
Conclusion. Regardless of age, the composition of epithelial cell membranes was different than
that of cortical and nuclear membranes, which showed similar phospholipid content. This
suggests that significant compositional changes occur when epithelial cells become elongated
to form fiber cells. Invest Ophthalmol Vis Sci. 1994;35:3938-3942.
JL he understanding of lens membrane function in
terms of the composition and molecular structure of
its phospholipid components is one of the thrusts of
our research. Numerous reports have dealt with the
phospholipid composition of membranes isolated
from human lenses.1"5 Most of these results are based
on the extraction of the membrane by the Folch6
From the Departments of *Chemistry and f Ophthalmology and Visual Science,
Kentucky Lions Eye Research institute, University of Louisville, Louisville,
Kentucky.
Supported by Public Health Service (Bethesda, Maryland) research grant 07975 and
the Kentucky Lions Eye Foundation (Louisville, Kentucky), and by an unrestricted
grant from Research to Prevent Blindness, Inc., New York, New York.
Submitted for publication January 10, 1994; revised May 11, 1994; accepted May
17, 1994.
Proprietary interest category: N.
Reprint requests: Dr. M. Cecilia Yappert, Department of Chemistry, 2320 Brook
Street, University of Louisville, Louisville, KY 40292.
3938
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
method, followed by the separation and quantification
of the phospholipids by normal-phase (i.e., silica stationary phases), thin-layer chromatography. More recently, Glonek et al developed a powerful method for
the analysis of phospholipid membranes based on 31P
NMR.7"9 Because phospholipids contain one phosphorus atom per molecule, the 31P NMR spectral response is expected to be related to the molar amount
of phospholipids. Until fairly recently, the resonances
from individual phospholipids could not be resolved
because of broadening caused by the multiple microenvironments of the phosphorus atoms and the aggregation of the different phospholipids. In 1988,8 Meneses et al reported the use of an aqueous solution
of potassium or cesium ethylene diamine tetraacetate
(EDTA) dissolved in methanol. This reagent acted as
Investigative Ophthalmology & Visual Science, October 1994, Vol. 35, No. 11
Copyright © Association for Research in Vision and Ophthalmology
Regional Composition of Human Lens Membranes
a detergent and resulted in the narrowing of the bands
for the different phospholipids. They applied this reagent and high-resolution 31P NMR to establish agedependent alterations in the phospholipid composition of human lens membranes.9 In their study, however, the membranes were not separated by lenticular
region; thus, the observed trends represented overall
changes in lens membrane composition. In this report, we establish the compositional differences in the
membranes from epithelial, cortical, and nuclear
cells. Although the lipid compositions of chicken10
and rabbit11 epithelial lens membranes have been reported, the phospholipid composition of human lens
epithelia is not known.
In this work, we extracted the phospholipids with
a monophasic protocol, and the phospholipid quantification was carried out by the 31P NMR method developed by Glonek et al.7"9
MATERIALS AND METHODS
Clear human lenses were obtained within 8 hours of
death from the Kentucky Lions Eye Bank. The epithelium, cortex, and nucleus were dissected. The three
regions were pooled according to age as follows: pool
1, 0 to 15 years (n = 16); pool 2, 16 to 30 years (n =
28); pool 3, 31 to 45 years (n = 34); pool 4, 46 to 60
years (n = 42); pool 5, 61 to 75 years (n = 91); and
pool 6, 76 years and older (n = 44).
Phospholipid Extraction
All solutions were bubbled with argon, and the extraction was carried out in an atmosphere of argon where
possible. The lenses were homogenized, without the
addition of water, with a blade homogenizer. In the
initial extraction, 20 ml of methanol was added per
gram of homogenized lens material. The nonlipid
components were removed by centrifugation (3000
rpm for 15 minutes). The solvent was decanted and
evaporated. To eliminate the possibility of contamination from methanol-soluble, nonphospholipid components, chloroform was added in an amount equal to
the initial volume of methanol. The sample was centrifuged at 3000 rpm for 15 minutes. The supernatant
was decanted to separate the lipid from the nonlipid
components in the pellet. For the further removal of
polar components, a 0.74% KC1 aqueous solution
(40% of the chloroform volume) was added to the
supernatant. The aqueous phase was then discarded,
and the chloroform layer was reduced to a volume of
0.5 ml. The phospholipids were finally crystallized by
adding 20 ml of acetonitrile. To isolate the white precipitate containing the phospholipids, the sample was
placed in a centrifuge tube and spun at 3000 rpm for
10 minutes. The supernatant was carefully removed
by aspiration with a pipette connected to the house
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
3939
vacuum system. The phospholipid extract was then
spectrally analyzed.
Lipid Extraction Efficiency
Broekheuyse1 and Merchant et al9 used a chloroformmethanol monophasic procedure to minimize the
possible losses of polar phospholipids and to avoid the
difficulties created by the frequent formation of an
emulsion layer when biphasic extractions, such as
those based on the Folch method,6 are applied. We
chose to use methanol in the first extraction step because of the high solubility of phospholipids in this
solvent.1213 With the use of 31P NMR spectroscopy, we
determined that one extraction with methanol was
sufficient to remove 99.9% of the total lipid. Further
extraction of the remaining pellet with either methanol or chloroform extracted less than 0.1% of the total
lipid. The efficiency of the final step of the extraction,
that is, the addition of acetonitrile, was tested using
phospholipid standards and bovine lenses. We found
that neither the total amount nor the composition was
affected by this addition.
Spectral Studies
The phospholipid extract was dissolved in 400 fi\ of
deuterochloroform. An aliquot of 250 //I of the Meneses—Glonek reagent, prepared as in references 8 and
9 and using KOH as the counter ion source, was added
at least 15 minutes before spectral acquisition. The
mixture was then shaken, and the aqueous phase was
allowed to separate before data acquisition. A Bruker
500AMX NMR spectrometer (Billerica, MA), operating at 202.4 MHz, was used to acquire 31P NMR
spectral data. Other acquisition parameters were:
spectral width, 2032.5 Hz; resolution, 0.50 Hz; acquisition time, 1.0 seconds; pulse length, 10 //sec; dwell
time, 246 //sec; and number of scans, 1000. The data
treatment was performed on a personal computer using Bruker's WINNMR software. The spectra were
phase corrected, zero filled, base-line corrected, and
deconvolved. The percentage of each phospholipid
was evaluated by integrating the peak area corresponding to each phospholipid and then calculating
the ratio of each area to the sum of all the areas. Nine
components were quantified; the identities of seven
of them are known, and two await identification.
RESULTS
Regional Differences
No major differences were observed between the phospholipid composition of the cortical and nuclear
membranes, except phosphatidylglycerol, whose average percent composition and corresponding standard
deviation were determined to be 10.8% ± 1.5% (n =
12) and 13.9% ± 1.8% (n = 10) for the cortical and
3940
TABLE l.
Investigative Ophthalmology & Visual Science, October 1994, Vol. 35, No. 11
Regional Phospholipid Composition of Human Lens Membranes
Phospholipid
Component
XI
LPG
PG
UNK
PE plas
PE
SPH
LPC
PC
PA
PC plas
Percent in
Epithelial
Membranes*
Percent in
Fiber
Membranes*
Percent in
Lens
Membranes\
Significantly
Different
(P < 0.0005)1
1.9 (0.5)
1.40 (0.03)
2.6 (0.3)
53.6 (4.6)"
3-10 §
3.6 (0.2)
12.3 (0.5)
54.5 (1.2)"
II
8.4 (1.3)
7-13 §
1.0 (0.2)
4_1§
2.3 (0.3)
1.0 (0.1)
2.6 (1.3)
1.8 (0.8)
9.8 (1.0)
43.7 (6.8)
14.5 (4.9)
5-8 §
7_1 2 §
1.5 (0.9)
4_2§
2.6 (0.8)
1.1 (0.5)
yes
yes
yes
no
II
11.6 (3.6)
11.5 (1.4)
2.5 (0.5)
5-16 §
1.9 (0.3)
2.5 (0.2)
no
yes
yes
yes
yes
yes
XI = Unknown component with 8 = 1.2 ppm; LPG = lysophosphatidylglycerol; PG = phosphatidylglycerol; UNK = unknown
component with (5 = 0.12 ppm; PE plas = phosphatidylethanolamine plasmalogen; PE = phosphatidylethanolamine; SPH =
sphingomyelin; LPC = lyso-phosphatidylcholine; PC = phosphatidylcholine; PA = phosphatidic acid; PC plas = phosphatidylcholine
plasmalogen.
* The values in parenthesis represent the standard errors, n = 8, 12, and 10 for the epithelial, cortical, and nuclear membranes,
respectively, f Values averaged from the data reported in Ref. 9. % Statistical significance of the compositional difference between
epithelial and fiber membranes using die Student's Mest for unequal variances. § These values represent the limits of the age-dependent
trend. " The bands for UNK and PE plas were not resolved in our spectra.
nuclear membranes, respectively. This difference was
statistically significant (P < 0.0005). Because no other
statistically significant differences were observed between compositions of cortical and nuclear membranes, the values in Table 1 represent the average
phospholipid content of epithelial and fiber membranes. The latter values were averaged from the cortical and nuclear membrane compositions. Except for
sphingomyelin (SPH) and the unknown phospholipid
XI (spectral shift of 1.2 ppm) in the fiber membranes,
and phosphatidylcholine (PC) in both epithelial and
fiber membranes, all other values represent the average percentages obtained over all age groups, because
no significant changes were observed with age.
ponent, regardless of age and lenticular region, was
the unknown (UNK) phospholipid with a chemical
shift of 0.12 ppm. Although the identity of this species
is not known, Merchant et al9 found that this phospholipid could not be saponified and suggested that it is
a phosphorylsphingosine derivative or a plasmalogen.
Depending on the region and age, the order in the
abundance of the other phospholipids varied. Thus,
the discussion will be based first on the regional trends
Sphingomyelin
Age Dependence of Phospholipid Composition
Significant age-related trends were observed for SPH,
which increased with age, and for the unknown XI,
which decreased with age in both cortical and nuclear
membranes. Figure 1 shows the percentage of SPH
for cortical and nuclear membranes as a function of
age. The linear regression of the data and the 95%
confidence limit are included to demonstrate this linear increase. Figure 2 shows the opposite age trend
exhibited by the unidentified component XI. As with
SPH, XI exhibited statistically identical age dependence in the cortical and nuclear membranes.
As shown in Figure 3, the percentage of PC also
exhibited a steady decrease with increasing age, especially in the epithelial membranes.
0
20
40
60
60
100
Age (years)
DISCUSSION
As previously reported by Merchant et al9 and confirmed in Table 1, the most abundant membrane com-
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
FIGURE l. Age-related changes in SPH for cortical (O) and
nuclear ( • ) membranes. Solid line = linear regression; broken lines = 95% confidence limits.
Regional Composition of Human Lens Membranes
3941
Unknown XI
10
1
a
e
1
1
•
-
0
©
•
o
>
O
-
>
n
•
Vv
\N
\
•
\
1
20
I
•
40
60
1
60
100
Age (years)
FIGURE 2. Age-related changes in unknown component XI
(6 = 1.2 ppm) for cortical (O) and nuclear (•) membranes.
Solid line = linear regression; broken lines = 95% confidence limits.
in composition and second on the age-related compositional trends.
Regional Trends
No significant differences were found between the
phospholipid composition of the cortical and nuclear
Phosphatidylcholine
20
40
60
80 100
Age (years)
3. Age-related changes in PC for cortical (O), nuclear (•), and epithelial (•) membranes. Solid line = linear
regression; broken lines = 95% confidence limits.
FIGURE
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
membranes, except phosphatidylglycerol, which varied slightly (see Results). Because of their similarity,
the percent composition of the cortical and nuclear
membranes was averaged and presented in Table 1
as percent in fiber membranes. The fiber membrane
composition obtained was similar to that reported by
Merchant et al9 for membranes extracted from whole
lenses, with the exception of the UNK phospholipid,
which was more abundant by our determination in
fiber membranes. The difference in the reported values could be caused by the overlap of the spectral
band corresponding to PE plasmalogen (PE plas) with
that for the UNK phospholipid at 0.12 ppm. Our spectral data showed no resolution of these two bands.
Merchant et al also reported greater variability in both
the percentage of UNK and the percentage of PE plas,
which showed relative standard deviations of 16% and
40%, respectively, in their results for whole lens membranes. The relative standard deviation was 2.2% in
our data for percentage (UNK + PE plas) in fiber
membranes.
This is the first report quantifying epithelial membrane composition and age-related regional differences in human lens membrane composition. The
most dramatic differences were observed in the epithelial membranes as compared to the cortical and
nuclear membranes. As seen in Table 1, except for
UNK and phosphatidylethanolamine (PE), all other
phospholipid contents were significantly different (P
< 0.0005) in the epithelial membranes as compared
to the fiber (cortical and nuclear) membranes. The
levels of PC were significantly higher in the epithelial
membranes, particularly at younger ages. This trend
in PC has also been reported for chicken embryo lens
membranes.10 The possibility of contamination of epithelial membranes with cortical ones has been evaluated and found to be minor.*
The compositional differences observed between
epithelial and fiber membranes cannot be totally attributed to intracellular organelles that are only present in the epithelium. The UNK phospholipid, which
makes up approximately half of the epithelial membranes, has not been reported in intracellular organelles. The SPH content of nuclear, endoplasmic, and
mitochondrial membranes is generally reported to be
3%, 5%, and 0%, respectively.14 Thus, one would expect the percentage of SPH to be much lower in the
epithelial membranes compared to the fiber membranes if considerable amounts of intracellular organelle lipids were present in the epithelial extracts. Furthermore, if the fivefold increase in the content of PC
observed in epithelial versus fiber membranes were
due to the presence of intracellular organelles, which
* Unpublished data (1993) based on the crystallin content of
the tissue. Obtained by electrophoresis in collaboration with Dr. D.
Garland of the National Institutes of Health.
3942
Investigative Ophthalmology 8c Visual Science, October 1994, Vol. 35, No. 11
contain 48% PC,14 we calculated that only 8% of the
total epithelial phospholipids correspond to organelle
membranes. However, this could not explain the relatively high content of SPH in the epithelia. The contribution of intracellular organelles to the epithelial
phospholipid composition warrants further study.
Phosphatidylcholine has been reported to form more
fluid, structurally disordered membranes1115; it is then
possible that the higher levels observed in the lens
epithelium are related to a more fluid lipid environment that could support the higher enzymatic activities in epithelial membranes. The lowest levels of phosphatidylglycerol were found in the epithelial membranes for every age group. The possible structural or
functional consequences of this compositional difference between epithelial and fiber membranes are not
clear at this time.
Age Dependence of Phospholipid Composition
Merchant et al9 reported an age-related decrease in
PE from approximately 8% in young lens membranes
to approximately 5% in older ones. We did not detect
a decrease with age in fiber membranes, which were
composed of 8.4% ± 1.3% PE. This discrepancy could
be due to the great variability of PE percentage reported by Merchant et al,9 in which percentage of PE
showed the weakest statistical correlation with age.
The most significant age trends observed in fiber
membranes were the rise in percentage of SPH and
the decrease in percentage of PC. These trends were
also observed by Merchant et al9 in whole lens membranes. The only age trend in epithelial membranes
was a significant decrease in the percentage of PC.
The changes in SPH and PC would be expected to
order membranes with increasing age. The inverse
relationship between the contents of SPH and PC has
been observed in muscle sarcoplasmic reticulum16 and
in other systems.17'18 Membrane fluidity is thought to
be regulated by the ratio of these two phospholipids.18
It is also interesting in Figure 1 that the highest levels
of SPH observed for the older fiber membranes are
comparable to the percentage of SPH in the epithelial
membranes (see Table 1). Similarly, as shown in Figure 3, the minimum levels of PC observed for the
oldest age group in the epithelial membranes are comparable to the highest levels observed for the fiber
membranes. The suggestive correlations observed in
the levels of these two phospholipids with age and
region emphasize the relevance of understanding
composition-structure relationships, because they
could affect the function of cell membranes.
Key Wards
human lens membrane, phospholipid composition, epithelium, cortex, nucleus
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
References
1. Broekhuyse RM. Phospholipids in the tissue of the
eye: III: Composition and metabolism of phospholipids in human lens in relation to age and cataract formation. Ada BiochimBiophys Hung. 1969; 187:354-365.
2. Feldman GL, Culp TW, Feldman LS, Grantham CK,
Jonsson HT Jr. Phospholipids of the bovine, rabbit
and human lens. Invest Ophthalviol Vis Sci. 1964; 3:194197.
3. Gooden MM, Takemoto LJ, Rintoul DA. Evidence for
reduced lipid order in plasma membranes from cataractous human lenses. Curr Eye Res. 1983;6:367-375.
4. Cotlier E, Obara Y, Toftness B. Cholesterol and phospholipids in protein fractions of human lens and senile cataract. Ada Biochim Biophys Hung. 1978; 530:
267-278.
5. Zelenka PS. Lens lipids. Curr Eye Res. 1984; 11:13371358.
6. Folch J, Lees M, Sloane Stanley GH. A simple method
for the isolation and purification of total lipids from
animal tissues. /Biol Chem. 1957;226:497-509.
7. Meneses P, Greiner JV, Glonek T. Comparison of
membrane phospholipids of the rabbit and pig crystalline lens. ExpEyeRes. 1990;50:235-240.
8. Meneses P, Glonek T. High resolution 3IP NMR of
extracted phospholipids. / Lipid Res. 1988; 29:679689.
9. Merchant TE, LassJH, Meneses P, Greiner JV, Glonek
T. Human crystalline lens phospholipid analysis with
age. Invest Ophthalmol Vis Sci. 1991; 32:549-555.
10. Zelenka PS. Phospholipid composition and metabolism in the embryonic chicken lens. Exp Eye Res.
1978;26:267-274.
11. Lamba OP, Borchman D, Garner WH. Infrared study
of the structure and composition of rabbit lens membranes: A comparative analysis of the lipids of the
nucleus, cortex, and epithelium. Exp Eye Res. 1993;
57:1-12.
12. Ramesh B, Adkar SS, Prabhudesai AV, Viswanathan
CV. Selective extraction of phospholipids from egg
yolk. JAm Oil Chem Soc. 1979;56:585-587.
13. Shigyo F, Takeuchi M. Fat in yeast growing on hydrocarbon substrates: II: The complete extraction and
separation of lipids. Nippon Nogei Kagaku Kaishi.
1972; 46:27-33.
14. Robinson GB. The isolation and composition of membranes. In: Parsons DF, ed. Biological Membranes. Oxford: Clarendon Press; 1975;8-32.
15. Marsh D. Handbook of Lipid Bilayers. Boca Raton: CRC
Press; 1990:121-295.
16. Borchman D, Simon R, Bicknell-Brown E. Variation
in the lipid composition of rabbit muscle sarcoplasmic
reticulum membrane with muscle type. J Biol Chem.
1982;257:14136-14139.
17. Lentz B, Hoechli M, Barenholtz Y. Acyl chain order
and lateral domain formation in mixed phosphatidylcholine-sphingomyelin multilamellar and unilamellar
vesicles. Biochemistry. 1981;20:6803-6809.
18. de Gier J, van Deenen LLM. Some lipid characteristics
of red cell membranes of various animal species. Ada
Biochim Biophys Hung. 1961;49:286-296.