v; .. .....I infit p...
2 4: 2.1.311

         

c ,
...?IP

in.
x.

       

.x...
f?

 

     

..
3.433;?
.....p

....o

     
   

   

        
 

V E
. I.
‘ v
.
1El’i”

 

..‘qfl
,
A
! v '

    

        

TD. .. 1L .

w.. .. . . M. ,.,.
_. . _. 2 - .
.w. . ..w.” ., ... .3”.
ff. m...

BR:

 

'4 49.4

 

A;

       

F, M

  

 

«an

 

2-5

stfér

 

15

 

has"

 

M;

 

 

       

 

an; , . . . ... V
. , , v v _ y . V.‘ . v . H. V .
v ,. . . . . c : ....
. V} ‘ . ,. . . ‘ . . .
V V . . . . . . . r. .M
— . _ ‘ V V V . . ..
y r . . u . ,t V
R. I . V . ,
V. , A V . . . . A; y
_ .V4 . ‘ ._ .
V V l y \‘ M .
V
A ‘ V I . .
_
M o
v . v
i .. M . _ V. v .
. . . , - .‘
v v . _ N n .‘
M.
. .
V I u n u . .. .»

 

.. ..Y .o....o.§..o.......Vo...ot..v.lXVo-alla :1 . .Anr.

 

.' .vt II.M'Qc=I-i! uV4L...V V . . r... A ....crxuu..vt..2.€su..

‘ V ; . . ... ,
. V 4
Iv I. .FrTuJXBhU o .1 FLILLLLI L..V..(s‘scv0r.sl [.1111311..\v. ...r .MV ...wovsogr.1ir L..r.1IJoI Jilanhi. U. . tirl:.h.1.1.....flvll....\r.m4-or.s}l.:.(Zuni ......u2.n..v...r. c}! 1P§§.En:|.éml. .

 

(fr-at. v r - ' .x.‘ '~.J«D/htl’

   
    

I 77‘?” 1’ Y‘ "
.41 V ht

I r l ‘ .-
Ml“. ; ~ .- do":
U "

£13 I . .

TH E’ifi

This is to certify that the

thesis entitled

CHARACTERISTICS OF MEMBRANE LIPOPROTEIN FRACTIONS
FROM cow's MILK

presented by

Fred C. Swope

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in FOOd Science

      

Major professor
Dr. J. R. Brunner

Date JUly 25, 1968

 

0-169

ABSTRACT

CHARACTERISTICS OF MEMBRANE LIPOPROTEIN
FRACTIONS FROM COW'S MILK

by Fred C. Swope

The fat globule membrane is a complex lipOprotein
covering small spheres of lipids in milk. Presumably,
its main function is to stabilize the lipid phase. The
.membrane accounts for slightly more than 1% of the total
weight of the globule and consists primarily of glycerides,
phospholipids, cholesterol, :1 structural-like protein,
various enzymes and metals.

This investigation was directed toward elucidating
the membrane structure more completely, using both physical
and chemical methods of analyses. The fat globules were
separated from other milk components and churned to frag-
ment the membranes into various size particles. To fully
utilize density and size variations, differential sedi-
mentation was employed to separate the eroded membranes.
‘Three pellets with approximate minimum sedimentation
coefficients of 7,5008, 2308 and 358 were collected and
studied.

Lipid analyses showed that the 7,5008 pellet con-
tained approximately 18% total lipids while the 2308 and
358 pellets increased to 30% and 55%, respectively, The

INNDSpholipids, however, accounted for 63% of the total

Fred C. Swope

lipids of the 7,5008 pellet but only “2% of the 358 frac—
tion. Likewise, cholesterol decreased in concentration
from the 7,5008 to the 358 pellet. Lipid micelles con-
taining surface-orientated”bimodal«moleCules~would'ade—'
quately explain these observations.

The membrane protein(s) are similar in amino acid
content to other milk proteins except for a lower glutamic
acid and proline and a higher arginine content. When all
three fractions are compared on mole basis, the 358 amino
acid content exhibits more variation than the other two
pellets.

The predominate characteristic feature of these
lipid-extracted fractions is their insolubility in the
usual aqueous systems employed to carry proteins. The
three protein fractions contain carbohydrates ranging in
value from 6.6% for the 7,5008 to 10.1% for the 358 pellet
protein.

A tentative model for the basic membrane structure
has been theOrized from compositional-and electron photo-
micrographic data. A stratumélike system appears to be ‘
preSent in the membrane consisting Of a'layer ofnstructural

protein with glycolipoprotein particles anchored to it.

ӎ"-
. .' t .. l

 

 

CHARACTERISTICS OF MEMBRANE LIPOPROTEIN

FRACTIONS FROM COW'S MILK

\-

By
\l.

\

\

Fred C: Swope

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Food Science

1968

If I.
O 4.
J .
«l I ' 7'
M“ \d r 7..
I . _
’ o

DEDICATION

This manuscript is dedicated to the memory
of my brother,

Ernest A. Sw0pe, Jr.

August 11, l926-October l“, 1961

ii

ACKNOWLEDGMENTS

The author wishes to express his sincere apprecia-
tion to Dr. J. R. Brunner for his valuable direction and
suggestions during the course of this study, and for his
assistance in the preparation of this manuscript. He is
also grateful to the other members of his guidance
committee, Dr. B. S. Schweigert, Dr. A. M. Pearson, Dr.

A. J. Morris and Professor L. J. Bratzler.

Special thanks are due to Mr. R. J. Carroll, U.S.D.A.,
Eastern Utilization Research and DevelOpment Division, for
kindly providing excellent electron microscopic analyses.

In sympathy the author wishes to extend an expres-
sion of gratitude to the late Dr. G. B. Wilson, Department
caf Botany and Plant Pathology for his timely counsel.

Lastly, the author is especially grateful to his
vvife, Agnes, for her understanding and encouragement

tliroughout his graduate program.

iii

TABLE OF CONTENTS

Page
ACKNOWLEDGMENTS . . . . . . . . . . . . . iii
LIST OF TABLES . . . . . . . . . . . . . . vi

LIST OF FIGURES . . . . . . . . . . . . . Vii

INTRODUCTION . . . . . . . . . . . . .
REVIEW OF LITERATURE . . . . . . . . . . . .
Historical Survey of the Fat Globule Membrane . . .
Isolation and Amount of Membrane' . . . . .

Chemical Composition of Membrane . . . . . . .

O
\l \0 U1 UK) LA) |-‘

Physical Properties of Lipoprotein Complexes . . . 1
Comparison of Several Membrane Systems . . . . . 22
Origin of the Fat Globule Membrane . . . . . . 27
IEXPERIMENTAL . . . . . . . . . . . . . . 31
Apparatus and Equipment . . . . . . . . . . 31
Chemicals and Materials . . . . . . . . . . 32

Chemical Methods .
Nitrogen .

Phosphorus . . . 3“
Hexose . . 35
Hexosamine .

Sialic Acid
Tryptophan .
Amino Acids
Total Lipids .

Total Cholesterol

0 O O O O O 0 O 0
LA)
0“

iv

Physical Methods . . . . . . . . . .
Starch—Urea Gel Electrophoresis . . . . .
Electron Microscopy . . . . . . . . .
Preparative Procedures for the Fat Globule

Membrane Fractions . . . . . . . .

RESULTS AND DISCUSSION . . . . . . . . .

Preparative Procedures . . . . . .
The Effects of Washing the Fat Globules . .
Pellet Separation by Ultracentrifugation . .

Chemical Analyses . ' . . .
The Protein-Lipid Relationship of the
Various Fractions . . . . . . . .
The Protein Composition of the Different
Fractions . . . . . . . . . . .

Proposed Structure for the Fat Globule Membrane
Supporting Evidence for a Protein-Lipoprotein
Structure . . . . . . . . . . .
Interpretation of Electron Micrographs of
7,5008, 2308 and 358 Pellets . . . . .

SUMMARY 0 O O O O O O O Q 0 O O O O
IBIBLIOGRAPHY . . . . . . . . . . . .
APPENDIX 0 O O O O O O O O O O O O O

Page
Al
A3
A3
50
50
50
52
53
53
55
57
57
60
7A
75
91

Table

LIST OF TABLES

Appearance and composition of the various
membrane pellets obtained from fifteen
hundred milliliters of fat globules . .

Relationship of protein, cholesterol and
phospholipid content to the total lipids
for each fraction . . . . . . . .

Composition of the lipid—extracted pellets

Amino acid composition of the protein
fractions . . . . . . . . . .

Amino acid composition of the protein
fractions expressed in moles . . . .

vi

Page

64

65
66

67

68

LIST OF FIGURES

Figure Page

1. Diagram for the isolation of the membrane
pellets from eroded fat globules of
thrice-washed cream . . . . . . . . . 49

2. Starch-urea gel electrophoretograms of
skimmilk and cream washings . . . .z . . 69

3. Yield and gross composition of the aqueous
membrane fractions recovered from aliquots
of a two to five times washed cream . . . . 70

u. Electron micrograph and a corresponding
schematic interpretation of a torn fat
globule membrane ghost isolated from the
7,5008 pellet . . . . . . . . . . . 71

5. Electron micrograph and a corresponding
schematic interpretation of a fragmented
fat globule membrane isolated from the
2308 pellet . . . . . . . . . . . . 72

6. Electron micrograph and a corresponding
schematic interpretation of small
membrane fragments isolated from the
358 pellet. . . . . . . . . . . . . 73

Vii

INTRODUCTION

The fat globule membrane is a lipid-protein complex
covering the surface of fat globules in milk. Its emul-
sifying properties allow the lipids of milk to remain
dispersed without phase separation. Upon partial or
complete removal of this surface active material, the
small cores of lipid coalesce to form butter. Age and
addition of electron-seeking metals such as copper and
iron cause objectionable flavor changes in milk which are
partially attributable to the oxidative deterioration of
the membrane system.

Considerable information is available on the chem-
ical and physical prOperties of this heterogeneous sub-
,stance. However, little is known about the physical
arrangement of the chemical subunits which comprise the
membrane. Even less information is available on the
nature of the proteins which make up approximately 50%
of it.

This study was undertaken for the purpose of inves-
tigating the membrane structure with particular emphasis
on the nature of the membrane proteins. The membranes were
fragmented and separated according to density by ultra-
centrifugation. This approach was used for several

reasons. For example, if the membrane consisted of a

bimolecular leaflet as ascribed to other biological meme
branes, the chemical composition of the various membrane
pellets should be nearly identical. The sedimentation
characteristic would depend primarily on the surface
area to volume ratio of each particle--assuming that all
the particulates have a globular configuration. However,
if the fragmented particles differed quantitatively in
their composition, then each corresponding pellet would
separate by density change and by size.

The fragmented membranes were separated arbitrarily
into three pellets. The parameters for the pellets were
defined in terms of Svedberg units. Each separation was
designed to give equal distribution among the pellets.

In turn, each pellet was analyzed for carbohydrates,
total lipids, phospholipids, cholesterol, and amino acids..
Physical data were collected on the membrane complex
through electron micros00py.

A model for the structure of the fat globule mem-
brane is prOposed with aid of the data collected from this

series of experiments.

REVIEW OF LITERATURE
Historical Survey of the Fat
* Globule Membrane

Fat exists in milk as small spherical globules
ranging from 0.1 to 12 u in diameter with the bulk of the
fat in globules of 2 to 5 u (Sommer, 1951). More than
one hundred years ago Ascherson (1840) proposed the
haptogen membrane theory for the material that covers
these globules. He considered this material to be a
condensation of albumin and other particles attached to the
fat surface. At the turn of the century, V61tz (190A)
stained this membrane material and observed it under the
light microsc0pe. The different staining intensities as
well as varying degrees of thickness led him to the con-
clusion that this substance was heterogeneous. Bauer
(1911), using improved staining techniques, supported the
work of VBltz (190A); mainly, that the membrane stained
differentially and appeared to vary in thickness.

Babcock (1885) was the first American worker to
investigate this fat surface. Later, he (1889) presented
evidence for small quantities of lacto-fibrin to serve as
a coating for the globules. Although his conclusions were
proved incorrect, his work served to stimulate a great
amount of interest in the fat globule membrane which in

itself was a major contribution.

3

In 1897 Storch, realizing the need for improved
methodology for the study of fat globules, devised a
method in which the fat globules were washed to remove the
milk plasma. Cream was repeatedly washed with distilled
water employing a cream separator. The fat was solubil-
ized from the membrane material with organic solvents
with the resulting residue being reported totally
different from known milk constituents.

v51tz (190A) and Abderhalden and V61tz (1909), using
a different approach, allowed the fat globules in milk to
rise through columns of distilled water or salt and sugar
solutions. Following hot ether extraction of the washed
cream, the ether-insoluble material remaining was reported
to be casein. This gravity separation technique was also
used by Titus gt §l° (1928). They also, concluded that
the protein portion of the membrane material was casein.
Nevertheless, they noted that the solubility character-
istics were not like casein.

'Palmer and his associates (192A, 1933, 1935, 1936,
19U0, 19AM, l9u5a, 1945b) completed a series of brilliant
studies on the nature of the fat globule membrane. Unlike
earlier workers who separated the membrane from the internal
fat of the globules by ether extraction, Palmer and co-
workers (1924-1945) churned the washed cream and recovered
the membrane material from the buttermilk and the residue

from the melted butter. In addition to finding phospholipids

and a high-melting triglyceride fraction, they (1933)
discovered that the protein was not like any of the other
known milk proteins. They (l9u5a) also noted that the
composition of this lipoprotein was not the same in
buttermilk and in butter plasma. In the former the
protein/phospholipid ratio was in the range of 2.A-3.8
while the latter contained 1.0 to 2.0.

A more complete historical development of this

subject is found in a treatise by King (1955).

Isolation and Amount of Membrane

One of the first serious problems to confront
researchers working in this area was how to adequately
isolate and concentrate the membrane materia1--free from
milk p1asma--for chemical and physical studies. Since the
amount of membrane isolated depends upon such factors as
globule size, temperature of milk, stage of lactation and
variations in analytical procedures, little wonder that
many different values have been reported (King, 1955).

In 192“, Palmer and Samuelsson were first to isolate
and partially identify the globule membrane employing the
churning method. Cream was repeatedly diluted with dis-
tilled water and recovered by centrifugal separation. The
washed cream was churned and the buttermilk was combined
with washings from the melted butter. The following year

Hattori (1925) utilized a different isolation scheme. Milk

was treated with water-saturated chloroform. A partition
occurred with the swollen fat globules being removed from
the lower layer. After ether extraction, the proteinaceous
residue was dried and called haptein. The major difference
between the churning method of Palmer and the extraction
procedures used by earlier investigators, particularly
those employing gravity separation, was that the former
method allows for the study of lipoprotein particulates

of the membrane; whereas, 11: the latter case the lipids
are usually removed by organic solvents.

Palmer and Wiese (1933), employing the washing and
churning technique, found the crude membrane to range
between 0.66 to 0.89 g per 100 g of fat. Jenness and
Palmer (19A5a) later reported the protein content to range
between 0.A6 to 0.86 g per 100 g of fat. More recently,
the total membrane from Jersey milk has been estimated
between 1.3 to 1.6 g per 100 g of fat (Chien and Richard-
son, 1967a; Swope and Brunner, In Review).

Mulder (1957) objected to the washing technique on
grounds that constituents were removed from the surface
layers of the globules. He recommended starting with
uncooled fresh milk and allowing it to separate by
gravity. Based on the distribution of any component
between the milk, cream and skimmilk, its concentration
at the fat surface can be calculated. Using this method,

Mulder and Menger (1958) calculated 0.8 g of protein per

100 g of fat. Roland (1956), using an equation derived by
considering the protein and fat content of milk, cream and
skimmilk, estimated that 0.“ to 2.2 g‘of protein were
present in 100 g of fat. Both methods suffer greatly from
the fact that they are inaccurate and also that they
estimate the loosely bound plasma proteins which are
carried to the cream layer.

Unquestionably, the method of Palmer and Samuelsson
(192“) for churning washed cream has been the most widely
accepted procedure for the isolation and study of the
membranous material. This method affords large quantities
of fat globules and rapid isolation. The resulting mem-
brane sol can be conveniently concentrated by pervapor—
ation as suggested by Palmer and Tarassuk (1936) or by
acid precipitation at pH A.0. These concentrates are
easily preserved by drying in the frozen state in a
vacuum chamber.

Isotonic solutions or distilled water have generally
been used as the washing media. It is particularly advan-
tageous to employ an isotonic wash solution when labile
systems such as membrane-orientated enzymes are under
investigation (Dowben 23 a1., 1967). On the other hand,
Inembrane preparations made from these carbohydrate—rich
washes are very susceptible to bacterial destruction if
lengthy investigation procedures are used. In two sepa-

rate studies the membrane yield was shown to increase with

isotonic washes as compared with distilled water (Erickson
3: al., 196“; Swope and Brunner, In Review). Although
those losses have not been adequately studied, it would
appear that the increased viscosity of the isotonic wash
solution in part protects the membrane against erosion
during centrifugal separation.

The actual effects of repeated washing on the mem-
brane are difficult to assess from the standpoint of how
many washings are necessary to remove other milk constit—
uents without excessively eroding the membrane material.
Since repeated washing is no more than a dilution process,
several are needed to rid the membrane from milk plasma.
Tarassuk EE.§£' (1959) washed cream as many as fifteen times
with warm distilled water without greatly affecting the
emulsion properties of the membrane. This, however, does
not suggest that excessive erosion is not taking place. To
study the loss of membranes by repeated washing, Zittle
22.2l' (1956) measured the activity of membrane—orientated
enzymes (xanthine oxidase and alkaline phosphatase). Based
on the amount of activity in the original cream, xanthine
oxidase and alkaline phosphatase activities recovered in
buttermilk obtained from cream washed four times were 1“
and 15%, respectively. Earlier, Rimpila and Palmer (1935)
found approximately one-half of the alkaline phosphatase

of fresh cream removed after six washings.

Chemical Composition of the Membrane

The heterogeneity of the membrane has been recognized
for many years. The first investigators, however, concerned
themselves primarily with the protein component. This is
not surprising since proteins were known to exhibit emul-
sifying prOperties and since fat globules formed stable
emulsions. After Palmer and Samuelsson (192N) introduced
the churning procedure, other membrane components were
quickly discovered and investigated. Of these, the membrane
lipids have been actively studied. Cholesterol and other
sterols also appear to reside in the membrane as well as
small quantities of heavy metals.

Although proteins represent a major part of the
membrane (Thompson EE.2l°: 1961; Swope and Brunner, In
Review), they are probably the least well-defined of all
the components. With the exception where the membrane
residue was characterized as casein (Abderhalden and
V51tz, 1909; Titus 32 a1., 1928), it has generally
exhibited prOperties quite different from other known milk
proteins. Casein and albumin have been serologically
detected on washed membranes (Mulder and Menger, 1958);
however, they appear to be only in minute quantities
and do not deter characterization studies to any impor-
tant degree (Sasaki an Koyama, 1956).

Brunner and co-workers (1953a, 1953b, 1953c),

studying the membrane of both non-homogenized and

lO

homogenized milk, showed that significant changes occur

at the surface of the fat globule during processing. These
studies indicated that the main protein constituents of
milk plasma form a large percentage of the membrane-protein
complex in homogenized milk. This would be anticipated
since homogenization considerably increases the surface
area to volume ratio of the fat globule; thus, creating a
need for more interfacial material. When the amino acids
of the native membrane were compared with those of milk
plasma proteins (Brunner 33 al., 1953a; Hare 22 a1., 1952),
the membrane was richer in arginine, glycine and phenyl-
alanine, but contained less aspartic acid, glutamic acid
and leucine.

In subsequent experiments (Herald and Brunner, 1957;
Brunner and Herald, 1958), the membrane was concentrated
by salting <iut in 2.2 M ammonium sulfate. After lipid
extraction, the proteins were dispersed in water and
classified into two distinct fractions by centrifugation.
The protein pellet, which represented approximately half
of the membrane proteins, was reddish brown in color and
quite insoluble in aqueous buffer systems. The soluble
fraction showed a higher concentration of alkaline phos-
phatase and less xanthine oxidase than the insoluble or
pellet fraction. The soluble proteins were studied
electrOphoretically in various buffer systems ranging from

pH 2.05 to pH 8.79. One small and two large migrating

ll

boundaries were observed. Employing strong reducing
agents and detergents, the pellet was partially
solubilized. Electrophoretic mobility patterns for
sodium sulfide—treated samples exhibited a single,
homogeneous boundary, whereas, the detergent-treated
fractions were heterogeneous. These solubility charac-
teristics indicate that part of the membrane proteins
are structural in nature. On the other hand, glycopro-
teins have been identified in the membrane (Thompson and
Brunner, 1959; Jackson gt al., 1962) which implies a
functional purpose.

A variety of minor proteins are known to exist in
milk (Whitney, 1958). Of these, the milk enzymes have
created considerable interest. The origin of these
enzymes is poorly understood at present and is compli—
cated by the fact that some may be derived from bacteria
or leucocytes (McMeekin and Polis, 1949). One of the
more intriguing questions left unanswered is whether they
originate from the vascular system or actually represent
secreted mammary gland enzymes. At present there is
evidence to support both routes. Aldolase, a membrane-
associated enzyme, has the same order of activity in
blood as in milk (Polis and Shmukler, 1950). Bingham
and Zittle (1964) found that ribonuclease A of milk is
identical to pancreatic ribonuclease A which seems to

swiggest the possible "leakage" of this enzyme from the

12

blood stream. To the contrary, Folley and White (1936),
studying the effects of thyroxine on milk secretion and
particularly on phosphatase activity, could find no
evidence to support the theory that this enzyme came from
the blood serum. Consequently, they considered phospha-
tase a true mammary gland enzyme.

The fact that many of the milk enzymes are asso—
ciated with the fat globule membrane is well documented
(Brunner, 1965). Other than xanthine oxidase and alkaline
phosphatase, the two major membrane enzymes, acid phos-
phatase (Bingham 33 a1., 1961), aldolase (Polis and
Shmukler, 1950), phosphodiesterase (Matsushita 33 $1.,
1965), lipase (Tarassuk and Frankel, 1957), DPNH-
cytochrome c reductase and DPNH-diaphorase (Morton,
1953c) have been reported as constituents of the fat
globule membrane. More recently Dowben 32 a1. (1967)
have identified (Na+-K+-Mg2+)-activated ATPase in the fat
globule membrane, an enzyme believed to be involved in
"active transport" across cell membranes.

Milk and particularly cream are rich sources of
xanthine oxidase, a purine and aldehyde activating enzyme.
As early as 1932, Tayama published an isolation procedure
for it. Sharp and Hand (1940) recognized the presence of
this enzyme on the surface of fat globules but referred

to it only in terms of a flav0protein. They noted that its

presence accounted for the reddish color of buttermilk and

i.
('A)

that upon heating to 176°F the flavin was released with a
subsequent loss of this characteristic color. Its molec-
ular weight has been estimated at 275,000 (Andrews 32 31.,
196A) with the crystalline enzyme average ratios for
protein, flavin adenine dinucleotide (FAD), molybdenum and
iron of l:2:(l.3-l.5):8, (Avis 33 31., 1956). Using the
molybdenum, iron and FAD content as a criterion for esti-
mating the amount of this enzyme, Swope 33 31. (1965),
Swope and Brunner (in review)calculated that it makes up
approximately 4% of the total weight of thrice-washed
membranes. It is also interesting to note that the ribo-
flavin content of human, mares', cows’, sheep and goats'
milks parallel the activity of xanthine oxidase found
therein (Owen and Hart, 1962; Long, 1961; Rodkey and

Ball, 19A6). Since FAD is loosely bound and may be sus-
ceptible to enzymatic degradation, it could well be a
source of riboflavin in milk.

Graham and Kay (1933) first demonstrated that phos-
phatase is concentrated on the fat surface of milk. This
enzyme, also known as alkaline phosphomonoesterase,
catalyzes the hydrolysis of organic phosphates. Morton
(1953a, 195A) highly purified the enzyme and estimated
that normal milk contains approximately 6 mg per liter.
Of the total,30-AO% of it is adsorbed to the fat globules
(Morton, 1953b). Since the enzyme is not present in true

solution and can be released by treatment with butanol,

1A

the remainder must be associated with small fat globules
or lipOprotein complexes present in skimmilk. In agar—gel
electrophoresis the enzyme from raw cream shows a pattern
of at least three isozymes (Peereboom, 1966).

In 192“ Palmer and Samuelsson qualitatively identi-
fied phospholipids as a membrane component. Palmer and
Wiese (1933) isolated and tentatively classified these
lipids as lecithin, cephalin and sphingomylin. Calculated
on the basis of non-dialyzable phosphorus, these three
fractions accounted for approximately 19% of the total
membrane weight. Later, Rimpila and Palmer (1935)
reported that more than 50% of the membrane actually con-
sisted of a non-phospholipid ether-extractable fraction.
This fraction when recrystallized from ethanol had iodine
numbers of 5.9-7.1, saponification numbers of l98.8-20A.0
and melting points of 52—5300 (Jenness and Palmer, 1945b).
They (19U5b), therefore, concluded that the properties of
this high—melting glyceride fraction were very similar to
one found in butterfat and may actually represent interior
lipids clinging to membrane phospholipids.

Mulder and associates (1957, 1958), in repeating
paii.of these experiments found twice as much phospholipids
(400-500 mg per 100 g of fat) with the surface layers of
tJie fat globules as reported earlier (Palmer and Wiese,
21933). They also reported 36 mg of cholesterol per 100 g

(of fat globules of which 15% was present in the esterified

15

form. It is rather difficult to compare Palmer's data_
with those of Mulder since separation procedures were used
which would account for a larger amount of surface material
in the latter case.

In 1961, Thompson 33 a1. extracted lipids from the
fat globule membrane and resolved them by gradient elution
from a silicic acid chromatographic column. The following
lipid classes were identified as components of the membrane:
carotenoids, a squalene-like compound, cholesterol esters,
cholesterol, an unique triglyceride mixture similar to the
high-melting triglyceride fraction, mono-diglycerides,
phospholipids, and unesterified fatty acids from some
preparations.

Rhodes and Lea (1958), in a detailed analysis of
milk phospholipids, found no major difference between
those of skimmilk, buttermilk or butter. The membrane
phospholipids contained (in moles %) phosphatidyl
ethanolamine 29, phosphatidyl serine 10, phosphatidyl
choline 33, and sphingomyelin 19. The content of lipid-
bound inositol and plasmalogens were 5 and 3 moles %,
respectively. They (1958) reported that unsaturated
fatty acids were present in both c' and 8 positions of the
Inilk glycerOphospholipids while other animal sources
usually have unsaturated acids in the a' position and

saturated acids in the 8 position.

16

The high-melting triglyceride fraction has now been
characterized by several groups (Patton and Keeney, 1958;
Thompson 33 31., 1959; Wolf and Dugan, 196“). Its infra-
red spectrum is indistinguishable from the spectra of
tripalmitin and tristearin. Wolf and Dugan (196u) found
that the trisaturated glyceride content of the membrane was
71.2%. Of the remaining triglycerides, 26% and 2.8% were
of the GS2U and GSU2 types, respectively.

At least two vitamins may be concentrated at the sur-
face of the fat globule. Skimmilk appears to be signifi-
cantly higher in vitamin A per g of fat and also in the
globule surface to volume ratio than whole milk (White 33 31.,
195“). But the work of Kon 32.31. (194A) suggests that vita-
min A is actually in true solution in the fat rather than
located at the surface.) Erickson 33 31. (196A) reported that
tocopherol concentration was at least three times higher in
the lipids of the fat globule membrane than inside the fat
globule.

Spectrographic analysis of the membrane proteins
showed that numerous metals are present (Herald 33 31.,
1957). However, to quantitate these metals with a
reasonable degree of precision is indeed difficult. The
partition ratio of metals found in milk versus cream
indicates that molybdenum, copper and iron are selectively
located in the membrane (Swope and Brunner, In Review).

Of the metals associated with membrane, copper has
been studied intensively (King 33 31., 1959; Koppejan and

Mulder, 1953; Davies, 1933). Its role as a catalytic

17

agent in oxidative deterioration of phospholipids is well
substantiated (Swanson and Sommer, 19u0; King and Dunkley,
1959; Tarassuk and Koops, 1960; Smith and Dunkley, 1961)°
Approximately one-quarter of the natural cOpper of milk

is bound to the surface of fat globules, probably as a
copper-protein complex (King et al., 1959). Consequently,
this makes the membrane the highest copper-containing

protein thus far identified in milk (Samuelsson, 1966).

Physical Prgperties of Lipoprotein Complexes

The theoretical interpretation of the molecular
structure of membranes in general comes from chemical
analysis, different physio-chemical properties of living
cells, x-ray diffraction of multimembranous structures,
electron microscopy and particularly Umestudies of mono-
molecular films (DeRobertis 33 31., 1965). By inte-
grating some of these observations, Danielli and Davson
(193A) proposed that membranes are composed of a lipid
layer with protein adhering to both interfaces. The
lipid layer is bimolecular with the polar groups extending
outwardly while the nonpolar groups are adjacent to each
other. In recent years detailed examinations of membranes
have resulted in a more definitive explanation of this
lipoprotein interaction. All membranes have large amounts
of an amphipathic molecule, usually phospholipids. These
bimodal molecules will aggregate in water to form micelles

which will interact with proteins (Green and Tzagoloff,

18

1966).. Lucy and Glauert (196A), studying the structure
and assembly of macromolecular lipid complexes made from
various mixtures of lecithin, cholesterol and saponin,
postulated that biological membranes have a globular con—
figuration in which phospholipid micromicelles form the
basic structure of the membrane. According to this model,
protein molecules are located around the repeating phos-
pholipid micelle. Experiments with model systems indi—
cate that these protein—lipid complexes are held together
by electrostatic forces, hydrogen bonding and in some
cases probably by hydrophobic bonds (Cornwell and Horrocks,
196“; Payens, 1959; Eley and Hedge, 1956). These inter-
actions and their relation to the stability of concen-
trated milk has been reviewed by Brunner (1962).

Several hypotheses have been proposed for the
structure of the fat globule membrane. King (1955)
visualized a monolayer of radially orientated phospho—
lipid and cholesterol molecules anchored in the fat phase
with their hydrophilic ends pointed toward a perpendicular
layer of protein. The protein is situated in such a manner
that its hydrophilic groups are electrostatically bound to
the phospholipids on one side and hydrated in the aqueous
phase on the other. In this scheme most of the hydro-
phobic side chains of the proteins are concealed by
structural folding. On the other hand, Morton (195A)

favored the View that the fat globules in milk are

l9

surrounded by a continuous protein membrane to which micro—
somal lipoprotein complexes are adsorbed. The microsomal
concept (Bailie and Morton, 1958a, 1958b) was based on the
fact that discrete lipoprotein particles with similar
enzymatic activities were found in mammary gland tissue and
on the fat globule surface. More recently Dowben 33 31.
(1967) have theorized that the globule membrane is a true
biological membrane. Evidence to support this hypothesis
is found in the ability of fat globule membrane antisera

to agglutinate and hemolyze bovine erthrocytes; the per—
meability of the membrane to potassium; the location of

the "unit" membrane structure; and detection of "micro—
somal" enzymes.

There is ample evidence that the membrane of fat
globules is a heterogeneous combination of lipids and
proteins. In 1961 Alexander and Lusena suspended material
obtained by freezing washed cream in a 2% sodium desoxy-
cholate (DOC) solution. The membrane suspension was
sedimented into five fractions by differential centrifu-
gation. These different fractions were distinguishable by
sedimentation behavior, enzyme content and total lipid and
phospholipid content. In a related but more thorough
study, Hayashi and Smith (1965) selectively solubilized
the intact fat globule with sodium desoxycholate. Water-
soluble lipoprotein particles accounting for 45% of the

total membrane weight were released. Chemical analysis

20

showed that the lipids and proteins were present in
approximatley equal amounts and that 76% of the lipids was
phospholipids. Comparison of these DOC-released particles
and those obtained from untreated preparations of membranes
indicated that the DOC treatment did not significantly
alter the physical structure of the complexes. Ultra-
centrifugal data showed that the particles were physically
heterogeneous (Hayashi 33_313,1965). It was therefore
concluded that the membrane consisted of two types of
lipid-protein complexes. These complexes are an insoluble
lipOprotein which borders the triglyceride core and water-
soluble lipoprotein particles which are adsorbed to it.
Chien and Richardson (1967a, 1967b) studied the gross
composition of the membrane by agitating washed fat
globules in distilled water rather than using a solu-
bilizing agent. The desorbed particles as well as the
interfacial material released from phase inversion were
also classified by sedimentation characteristics and
chemical analysis. The data in general support the con-
clusions of Hayashi and Smith (1965) that the globule
membrane is, indeed, a complex lipOprotein system comprised
of several particulates which differ quantitatively in both
lipids and proteins. Morton (1954) has estimated these
lipoprotein fragments to comprise 5% by weight the total

milk protein and range in size from less than 30mu to 200mu.

21

Upon separation of milk into cream and skimmilk a
distribution of phospholipids occurs with approximately
40% of the lipid phosphorus remaining with the skimmilk
(Perlman, 1935). These phosphorus compounds are presumably
derived from small unseparated fat globules and lipo-
proteins macromicelles which may or may not originate from
fat globules. The migration of loosely-bound lipid com-
plexes from the fat globule is a common occurrence when
processing treatments such as agitation, heating or
homogenization are employed. For instance, Greenbank and
Pallansch (1961) calculated that 29.5% of membrane phos-
phorus migrates to the aqueous phase upon stirring fat
globules for one minute at 20°C. Heating of milk causes
about 20% of the phospholipids to be desorbed from the
globule surface (K00ps and Tarassuk, 1959).

Both the membrane lipoprotein and other lipoproteins
prepared from unwashed cream show alkaline phosphatase
activity (Sasaki and Koyama, 1959). Zittle 33 31. (1956)
.suggested that alkaline phosphatase occurs in skimmilk
primarily in the same complexes as it does in cream but
that xanthine oxidase may occur in smaller and less easily
sedimented units in the skimmilk. Within the last several
years Patton 33 31. (1964, 1965) have discovered a lipogenic
agent in milk which has the capacity for glyceride biosynthe-
sis. This material has physical and chemical similarities to
the fat globule complexes but its exact relationship is
unclear since the biosynthetic activity appears to be confined

almost entirely to milk plasma.

22

The possibility of a lipid-casein Specie has been
raised by Cerbulis and Zittle (1965). They stated that
acid precipitated casein contains 4-7% of complex lipids
which can be differentially extracted with organic
solvents. Morton (1953a) made a similar observation
when acids or organic agents were used to remove casein
from milk. However, he considered this material nothing
more than entrapped and adsorbed lipoproteins of milk.

From the above discussion it is clear that lipo-
protein micelles are easily adsorbed and desorbed from the
fat globules. Whether these membrane—associated particles
represent all lipOproteins of milk cannot be answered
with certainty at this time. From enzymatic and compo-
sitional studies it appears that these particles may be
similar in certain aspects while different in others.

The data presented reject the concept of a well organized
membrane structure for the fat globule. Perhaps in places
an intact "unit" membrane does exist. By and large,
however, lipoprotein macromicelles surround the fat
globule. This material might be structurally the subunits

of membranes aggregated in a random fashion.

Comparison of Several Membrane Systems
The complexity of membranes is such that a compara-
tive analysis is very elementary at best. Their molecular
interpretation has been based primarily on differential

staining in conjunction with electron microscopy. Various

23

heavy metal stains have been employed for this purpose.
Osmium tetroxide, which appears to be deposited at polar
interfaces, has been used extensively to show the linear
relationship of proteins in membranes. For contrast,
tissue is fixed with formalin dichromate for electron
staining of“phospholipids. Isolation of the actual mem-
brane for chemical analysis, however, is a very formidable
task. It is exceedingly difficult to concentrate these
structures free from other cellular debris, and for this
reason the precise characterization of chemical compo-
nentseun.enzymatic activities is questionable.

Most information on the chemical composition of
membranes comes from the study of erythrocyte ghosts
(Parpart and Ballentine, 1952). After hemolysis of the
red blood cell a thin membrane (approximately 1% of the
red cell weight) is left which is composed of lipids and
proteins. The major protein component is fibrous and has
a high molecular weight. Most of the lipids are phospho—
lipids with cephalin representing nearly three quarters
of them. These lipids have different degrees of binding
and have been divided into classes by action of solvents.
It has long been recognized that strongly acidic groups
are responsible for the net negative charge of the erythro—
cyte possibly originating from phospholipids. Recent
evidence indicates that the electrokinetic behavior may

come from another source (Cook 33 31., 1961). They showed

24

that a decrease in erythrocyte charge occurred after treat-
ment with crystalline neuraminidase which released an
acetylated neuraminic acid and reduced the charge approx-
imately 80%.

In general, plasma membranes have not been adequately
characterized due to limitations in the isolation proce-
dures. Nevertheless, a dominant feature of these membranes
is a tripartite structure ranging in thickness from
75-100 X (DeRobertis 31; 31., 1965). Emmelot 33 31. (1964),
studying the chemical and enzymatic composition of plasma
membranes isolated from rat liver, found for the most part,
those components which appear to exemplify all membranes.
They stated that per 100 molecules of plasma membrane
phosphorus about 50 molecules of choline, 40 molecules of
cholesterol, l7 molecules of hexose, l6 molecules of
hexosamine, 9 molecules of sialic acid and 3,000 molecules
of amino acids are present. Membrane enzymes, such as
phosphatases and ATPase, were particularly evident. Using
neuraminidase, Weiss (1963) showed that this enzyme pro-
duced significant weakening <of cell surfaces. It appears,
that here too, a glyco- or muc0protein is a component of
plasma membranes. This mucoid substance may effect cell
aggregation and determine their antigenic properties.

After examining fat globules with the electron
microsc0pe, Schwarz (1947) reported that this membrane

also consisted of three layers. The inner layer was thin

25

but Optically dense; the middle layer appeared to contain
a ring of phospholipid droplets; and the outer layer was
thick and presumably made of protein. Knoop 33 31. (1958)
estimated this structure to range from 5-10 mu in actual
thickness. As with other types of membranes, the globule
membrane is negatively charged having an isoelectric

point at pH 4.3 (Jack and Dahle, 1937). An increase in
globular mobility occurs as the number of washings of

the cream is increased (Tjepkema and Richardson, 1966).
This is probably due to erosion of material from the polar
ends of phospholipids.

One of the more interesting facets of the fat
globule membrane is the phenomenon of creaming. This sub-
ject has been treated in detail by Dunkley and Sommer
(1944). They concluded that globular clustering is an
agglutination mechanism common to both fat globules and
bacteria with a globulin being essential in each case.

The removal of this material or heat denaturation of it
will hinder this clustering effect. Payens 33 31. (1965)
presented evidence to support this hypothesis in the case
of fat globules. They suggested that agglutination is
brought about by an euglobulin bridge between the indi-
vidual globules.

Admittedly, the available data for membrane compar-
ison are fragmentary and should be used only as a guide-

line. Much additional work is required to clarify this

26

area of research. For instance, the glyceride compounds
of the fat globule membrane may not be an actual part of
the true membrane system. Vasié and deMan (1966a, 1966b)
suggested that the high-melting glycerides are in reality
an artifact resulting from crystallization in the peri-
pheral layer of the globule and should not be considered
a true constituent of the membrane.

It is, indeed, interesting to compare certain
aspects of these membrane systems. Several enzymes appear
to be common to all--the esterases and ATPases. They are
also characterized by high phospholipid and cholesterol
content. The proteins are essentially of a high molecular
weight and aggregate into large complexes which are
insoluble with normal treatment. It is apparent that a
carbohydrate-containing material is an integral part of
these systems and probably determines the agglutinating
and antigentic properties of the membrane surfaces and
in part regulate their charge.

0n the other hand, many dissimilarities exist in
these structures. The fat globule membrane is consider-
ably thicker and contains many loosely-bound lipoprotein
complexes not common to all membranes. Other subtle
differences are detectable and probably reflect special—
ized functions. For example, membrane ghosts from 333333
tococcus faecalis protOplast contain a considerable amount

of a lipid resembling glycerides which is not normally

27

found hiother biological membranes (Ibbott and Abrams,
1964). In addition, the skeletal muscle cell membrane has
a very low lipid to protein ratio (approximately 1:4)

as compared with other membranes (Kono and Colowick, 1961).
Emmelot 33 31. (1964) reported that the enzymes esterase,
glucose 6-phosphatase and NADH-cytochrome c reductase are
4—9 times more active in microsomal than plasma membranes
preparations. This observation is particularly noteworthy
since there is continuity between these two membrane
systems at certain sites. Novikoff 33 31. (1953) have
also demonstrated marked chemical and enzymatic hetero-
geneity among microsomal as well as mitochrondrial frac-
tions. Thus, the chemical and physical differences serve
to place certain restrictions on any definitive inter-

pretation of membranes at this time.

Origin of the Fat Globule Membrane

A considerable effort has been put forth to learn
the origin of the fat globule. The technique of radio-
active labeling has elucidated several of the synthetic
pathways. Presently, the most obscure area left for
investigation is the transportation or movement of these
materials into the mammary gland and their organization
and subsequent secretion.

The milk fat lipids are primarily derived from the
lipids of the B-lipoproteins (Glascock 33 31., 1964, 1966)

while the glycerol for these lipids appears to originate

28

from blood glucose (Wood 33 31., 1958; Luick, 1961).

Once these materials are made available to the cell they
are rearranged in various ways by intracellular enzymes
(McCarthy 33 31,, 1965). The phospholipids of blood,
according tn) Graham 33 31. (1936), do not account for any
appreciable quantity of phosphorus compounds of milk.

Such compounds are derived mostly from inorganic phos-
phates of blood plasma.

Hayashi and Smith (1965) have raised the possibility
of blood lipoproteins making up in part the loosely-
adsorbed material that surrounds the fat globule. This
suggestion is plausible since certain milk proteins do,
indeed, enter the milk performed from the blood stream
(Coulson and Stevens, 1950; Polis 33 31., 1950; Larson and
Gillespie, 1957). Larson (1965) reported that approxi-
mately 10% of milk proteins are blood proteins while the
remainder are synthesized in the mammary gland. Using
fluorescent antibody gamma globulin, Dixon 33 31. (1961)
found that the acinar epithelieum allows selective passage
oi‘certain blood proteins while restricting others. These
proteins pass through the cytoplasm and are discharged into
the acinar lumen by an undisclosed process. In contrast,
proteins which are synthesized by the mammary tissue seem
to be packaged and released to the lumen by discrete

action of the Golgi apparatus.

2 9 ‘

However, little evidence is presently available to
support the view that blood lipoproteins provide any
sizable amount of the material coating the fat globule.
Von Muralt 33 31. (1964), using antigenic properties to
determine the relationship of milk and blood proteins of
human milk, reported that at least 13 to 18 proteins
correspond to known blood serum proteins but only traces
of blood lipoproteins were present.

The mechanism by which fat droplets are formed in
the secreting cell is nothing more than speculation at
present. Bargmann and KnOOp (1959) and Bargmann 33 31.
(1961) have shown certain events which appear to be
sequential during the secretion process. Formation of
lipid drOplets occurs in the basal region of the secreting
cell, probably in vacuoles of the ergastoplasm. Although
the regulation of globular growth is completely obscure,

a limiting membranous interface is evident during globular
intracytoplasmic movement. At the apical area of the

cell the fat globule protrudes into the lumen of the
alveolus and becomes constricted off carrying with it
additional cellular material. With this scheme in mind

it is possible to designate at least two membranous
layers--one originating from the interior of the cell and
tightly bound and another capping the globule upon release

from the cell.

30

There is ample evidence to support "selective leakage"
or perhaps aprocine secretion for fat globules. Many
cellular components have been identified in the membrane
while others are conspicuously absent. For example,
mitochondrial enzymes are missing (Morton, 1954). The
presence of small amounts of RNA have been detected
(Swope and Brunner, 1965) and would seem to suggest that
the secreting cell is not losing to any great extent its
machinery for synthesis and that the material being lost
can readily be resynthesized. On the other hand, many
secreting cells are sloughed off during the secreting
process with the subsequent release of autolyzing enzymes
and cell fragments (Paape and Tucker, 1966). This,
indeed, makes it exceedingly difficult to ascertain the
true nature of the fat globule membrane. Although such
cellular degradation should not be viewed as the major
route of secretion, it must be taken into consideration for
the interpretation of the results of the secretion phenom-

enon o

 

EXPERIMENTAL

Apparatus and Equipment

The milk used in this study was collected in stain—
less steel cans and separated with a DeLavel disc-type
separator (Model 9). Low—speed centrifugations were
performed with the International Model V, size 2 centri-
fuge. Intermediate-speed separations were performed with
a Sorvall RC2-B refrigerated centrifuge, using a type
SS-34 rotor. The Beckman Model L-2 preparative ultra-
centrifuge, equipped with a type 30 fixed-angle rotor,
was used for high-speed runs.

Cream was stored in Erlenmeyer flasks and churned
on an Eberbach rotary shaker.

For weighings, a tOp-loading, direct reading,
Mettlerigme K-7, a Mettler type H-16, and a Sartorius
series 2400 balances were used.

Amino acid analyses were performed with a
Beckman/Spinco Model 120C amino acid analyzer.

Membrane preparations were dried from the frozen
state by a laboratory—constructed 1y0philizer. Final
drying for preparations as well as chemicals was
accomplished in a temperature—controlled Cenco vacuum
OVen in which the vacuum was obtained with a Cenco

Pressovac-4 pump.

31

32

A Beckman DK-2A ratio recording spectrophotometer,
equipped with quartz cells, was used for all colorimetric
determinations.

A laboratory-constructed Plexiglass electrophoretic
cell was employed for starch-urea electrophoresis with a
Heathkit Model IP-32 supplying the power. Excess stain

was removed with an electrolytic destainer.

Chemicals and Materials

The principal chemicals used in this research with
the name of the suppliers are listed below.

Sigma 121 (tris-hydroxymethyl aminomethane) was
used as a primary standard and obtained from the Sigma
Chemical Company. Phenol was purchased from Mallinckrodt.
Acetylacetone was acquired from Fisher Scientific Company
and redistilled before use. The p—dimethylaminobenzalde-
hyde and 2-thiobarbituric acid were purchased from Eastman
Organic Chemicals and recrystallized. Specially prepared
hydrolyzed starch was obtained from Connaught Medical
Research Laboratories.

Mannose was purchased from Fisher Scientific
Company and galactose from Pfanstiehl Laboratories, Inc.
(Germany). Glucosamine hydrochloride, galactosamine and
penicillin "G" were acquired from Nutritional Biochemicals
Corporation, while N-acetyl neuraminic acid, tryptophan,

streptomycin sulfate and the amino acid standards were

33

obtained from Calbiochem. Applied Science Laboratories,
Inc. was the supplier of cholesterol.
All other chemicals used in this research were of

reagent grade.

Chemical Methods

 

Nitrogen

A micro-Kjeldahl apparatus was used for the nitrogen
analysis. The digestion mixture consisted of 5.0 g
CuSOu-5H20 and 5.0 g of SeO2 in 500 m1 of concentrated
H280“. Ten to 15 mg of dried protein in duplicate for
each sample were digested with 4 m1 of the digestion
mixture over a gas flame for one hr. After cooling,
1 m1 of 30% H202 was added to each flask and digestion
continued for an additional hr. Finally, the digestion
flasks were rinsed with 10 m1 of deionized water. Each
digestion mixture was neutralized with approximately 25 m1
of a 40% NaOH solution. The ammonia was distilled into
15 m1 of a 4% boric acid solution to which 3 drops of
indicator had been added. The indicator consisted of
400 mg bromocresol green and 40 m1 of methyl red in 100 m1
of 95% ethanol. The distillation was continued until a
volume of 50 m1 of the solution was present in the receiving
flask. The ammonia-borate complex was titrated with
0.020 N HCl.

The average recovery of a tryptophan standard was

97.4%.

34

Phosphorus

Phosphorus was determined by modifying a method
described by Sumner (1944). All the analyses were performed
in duplicate. For membrane protein, 10 to 16 mg were
digested with 2.2 m1 of 50% H280“ in test tubes. The
digestion was carried out on a sand bath heated by an
electric heater at a temperature of 160 to 170°C until the
protein was thoroughly charred. After cooling, 8 drops of
30% H2O2 were added to each test tube and heating resumed
for 15 min. This step was repeated until the solutions were
colorless. It is important to drive off all the H2O2 before
continuing with the determination. After final cooling,
the digestion mixtures were transferred to 25 m1 volumetric

flasks. Five ml of a 6.6% (NHu)6Mo 024-4H20 solution and

7
enough water were added to give a volume of approximately
15 ml. Then, 4 m1 of a solution of 5 g FeSOu°(H2O)7 in 50
ml of water plus 1 m1 of 7.5 N sulfuric acid were added
and mixed. The ferrous sulfate was prepared immediately
before use. The volumetric flasks were made up to volume
with deionized water and mixed. After standing for 30 min
to allow complete color formation, the transmittance (%)
of the solutions were read at 660 mu with a spectrophoto-
meter.

For lipid phosphorus, aliquots of chloroform-methanol

extracts (4 to 7 mg of lipid) were evaporated to dryness

in a water bath. The digestion and color formation were

35

carried out in the same manner as described previously
except that 50 ml volumetric flasks were used in place of
the smaller ones.

Ten m1 of a stock solution containing 1.3609 g of
KH2POu dissolved in 1000 ml were diluted to 100 m1
(0.031 mg/ml). A standard curve covering the range of

0.0 to 0.28 mg of phosphorus was prepared.

Hexose

The hexose analyses were performed by the method
reported by Dubois 33 31. (1956).

Two to 4 mg of protein were placed in a test tube
and 3 ml of deionized water were added. An equal volume
of 5% phenol solution was added and mixed. Finally, 15
ml of concentrated sulfuric acid was added rapidly. As
the acid entered, the test tube was rotated with a Mini-
Shaker to obtain maximum mixing and heat production.
After standing for 10 min, the tubes were again stirred
and then placed for 20 to 30 min in a water bath at 25 to
30° C. The addition of acid and the stirring of the
reaction mixture in the prescribed manner are necessary
for reproductibility in this test. The transmittance (%)
was read at 490 mu with a spectrophotometer.

A standard curve was constructed covering the range
of 0.0 to 200 ug galactose—mannose (1:1 w/w) per 3 m1

of water.

36

Hexosamine

 

The hexosamine content of the various protein frac-
tions was analyzed according to the procedure described
by Johansen 33 31, (1960).

Samples weighing 1 to 2 mg were directly placed
into 5 m1 ampoules to which 1 m1 of 4N HCl was added. The
samples were frozen in a dry ice-ethanol bath, evacuated,
refrozen and sealed under vacuum. Standard glucosamine-
galactosamine solutions (25 ug) were treated in the same
manner as the samples to serve as recovery standards.
Blanks containing all the reagents were treated identi-
cally to the samples and served as controls.

Hydrolysis of the protein took 6 hr at 100° C in a
hot air oven. After cooling, the hydrolyzed samples were
transferred to distillation flasks. The empty ampoules
were rinsed with 1 ml of 4N NaOH, followed by two 1 ml
rinsings with deionized water.

One milliliter of freshly distilled acetylacetone was
added to 25 ml of M Na2CO3 solution plus 20 m1 of water.
The pH was adjusted to 9.8 and the volume was made up to
50 m1. This solution was used within 30 min. Ehrlich's
reagent was prepared by dissolving 2 g of recrystallized
p-dimethylaminobenzaldehyde (Rondle and Morgan, 1955) in
absolute ethanol containing 3.5% of concentrated HCl.

This solution was diluted to a final volume of 250 ml.

37

Five and five-tenths milliliters of the acetylactone
reagent were added to each flask containing the hydrolyzed
sample. The final pH of this solution must be maintained
between 9.5 and 10.0. The flasks were tightly stoppered
and heated in a boiling water bath for 20 min. Then, each
cooled flask was heated with a Bunsen burner until 2.5 m1
of the distillate containing the steam—volatile chromogen
was condensed in 7.5 m1 of Ehrlich's reagent. The volu-
metric flasks were stoppered and mixed. After 30 min, the
transmittance (%) of the solution was read with a spectro-
photometer at a wave-length of 548 mu.

Equal amounts of glucosamine and galactosamine were
combined and used to construct a standard curve. The
carbohydrate concentration ranged from 0.0 to 50 pg per

3 ml.

Sialic Acid

 

Warren's thiobarbituric acid method (1959) as
modified for milk proteins by Marier 33 31. (1963) was
adopted for the sialic acid determination. The solutions
used in the analysis were as follows: 0.1 N H280“, 0.2 M
sodium periodate (meta) in 9M phosphoric acid, a solution
of 0.6 M sodium sulfate in 0.1 N H280“ to which was added
sodium arsenite at a rate of 10% in terms of final concen-
tration and a 0.5 M sodium sulfate solution to which
recrystallized 2-thiobarbituric acid was added at the

rate of 0.6%.

38

Samples of dried protein ranging in weight from 1.0
to 1.6 mg were placed in test tubes followed by the addi-
tion of 0.5 m1 of 0.1 H280“. Sialic acid was released by
digestion at 80°C for 40 min. To the hydrolyzed samples,
0.1 ml of the periodate solution was added. After mixing,
the tubes were incubated at 25° C for 20 min. Following
the addition of 1 m1 of the arsenite solution, the tubes
were shaken until the solution became colorless. Three
milliliters of thiobarbituric acid solution were added and
the tubes were shaken, tightly covered and placed in
boiling water for 15 min. Upon cooling, 7 ml of cyclo-
hexanone were added and the tubes shaken vigorously.
Finally, the tubes were centrifuged in a clinical centri-
fuge for 10 min. The clear cyclohexanone phase was
removed and read at 549 mu in a spectrophotometer.

A standard curve was constructed using N-acetyl neura-

minic acid in the range of 0.0 to 50 ug in 0.5 m1 of water.

Tryptophan

This amino acid is labile to acid hydrolysis and must
be determined independently from the rest.

In this study tryptophan was chemically analyzed by pro-
cedure E as described by Spies (1967). A 2~to 4 mg sample of
protein was weighed directly into a small glass vial. To
each sample was added 100 pl of Pronase solution which con-
tained 10 mg of Pronase per ml of 0.1 M phosphate buffer,
pH 7.5. Solutions were prepared on the day they were used.

The vials were stOppered and incubated for 24 hr at 40° C.

39

The cooled vials were set in 25 ml Erlenmeyer flasks
containing 9.0 m1 of 21.2 N sulfuric acid and 30 mg of
p-dimethylaminobenzaldehyde. To each vial was added 0.9
m1 of 0.1 M phosphate buffer solution. The vials were
tipped over and quickly mixed by swirling. The flasks
were placed in the dark at 25° C for 6 hr. Then, 0.1 m1
of 0.045% sodium nitrite was added. After 30 min,
transmittances (%) were read at 590 mu.

Duplicate samples of the Pronase solution, without
protein, were treated and analyzed as described above.

The tryptophan content of the Pronase solution was sub-
tracted fronlthetotal tryptOphan value. The blank solu-
tion contained everything but protein and Pronase.

A standard curve was constructed using 0.0 to 120
ug of tryptophan in phosphate buffer solution as described

in procedures 3 and E by Spies and Chambers (1948).

Amino Acids

 

The amino acid analyses were performed on 20 and
70 hr acid hydrolysates employing a Beckman Amino Acid
Analyzer, Model 120C (Moore, 33 31., 1958). Approximately
4 mg of dried protein were weighed into 10 ml ampoules.
Six milliliters of 6 N HCl were added to the ampoules.
The contents of the ampoules were frozen in dry ice-
ethanol, evacuated by vacuum, allowed to melt slowly to
remove gases, refrozen and sealed with an air-propane

flame. The sealed ampoules were placed in an oil bath

40

which was heated by a recirculating oven regulated at
110° C. The sets of duplicate samples were hydrolyzed for
either 20 or 70 hr.

The hydrolysates were evaporated to dryness in a 25
ml pear-shaped flask fitted to a rotary evaporator. Each
residue was taken up in small amounts of deionized water
and redried until all the HCl had been removed. Finally,
5 ml of a citrate-H01 buffer (pH 2.2) was added to the
dried hydrolysate. An aliquot of 0.2 ml was removed

from each sample for amino acid anlaysis.

Total Lipids

 

The lipids were extracted from the lipoprotein
particles by the procedure of Folch, 33 31. (1957).

One gram samples were mixed with 100 m1 of chloroform-
methanol (2:1 v/v). After 20 min of stirring, 100 ml of a
2% KCl solution were added and stirred for 10 min. This
mixture was centrifuged for 20 min at 2,000 X g. The
lower lipid-containing layer was carefully removed by
syringe. Anhydrous sodium sulfate was added to remove the
residual moisture from the lipid extracts. Following
filtration through Whatman No. 42 paper, 50 ml aliquots
were evaporated in aluminum containers to determine the
lipid content. Controls were analyzed to determine

residual content of the solvent systems.

.4.“

M
l“
I

41

Total Cholesterol

 

This sterol was determined by modifying a method
described by Bowman and Wolf (1962).

Aliquots of lipid extracts (l to 3 mg) in chloroform-
methanol were evaporated to dryness in a hot water bath.
The lipids in each test tube were dissolved in 2 m1 of
absolute ethanol. The color was developed by adding 2 m1
of a color reagent containing 8 m1 of an iron stock
solution diluted to 100 ml with concentrated sulfuric
acid. The iron stock solution was prepared by dissolving
2.5 g FeC13-6H20 in 100 ml of 87% H3P0u. After vigorous
stirring with a Mini-Shaker, the tubes were cooled with
the transmittance (%) read at 550 mu after 20 min.

The standard curve was constructed using choles-
terol (99% minimum purity) in the range of 0.0 to 55.0

pg per 2 m1.
Physical Methods

Starch-Urea Gel Electrophoresis

The procedure and equipment used in preparing and
performing this type of zonal electrOphoresis were adapted
from the method described by Wake and Baldwin (1961).

The gel was prepared as follows: to 70 ml of stock
tris-citrate buffer (92.0 g of tris-hydroxymethyl amino-
methane and 12.1 g of citric acid per liter), 170 m1 of

deionized water and 45 g hydrolyzed starch were added with

42

constant stirring and heated over an cpen flame to 65° C.
Next, 147 g of urea were slowly added with continued stirring
and heating until 90° C was reached. The hot solution was
poured into a two-liter filter flask and degassed by applying
vacuum. The deaerated solution was poured into the gel bed
and a slot former placed near the cathodic end of the cell.
After aging for approximately 24 hr, the gel was prepared
for the analysis.

The protein solution was prepared by adding 20 mg of
protein to five-tenths milliliter of a buffer consisting
of 70 m1 of tris-citrate stock buffer, 170 ml water and
147 g urea. Approximately 150 pl of each sample were intro-
duced into each gel slot. The slots were covered with a
layer of mineral oil.

The buffer tanks, connected to the gel through filter
paper wicks (Eaton and Dikeman No. 652), were filled with
boric acid-sodium hydroxide buffer. This stock buffer
contained 881 g boric acid and 190 g sodium hydroxide in
19 liters of water, adjusted to pH 8.6. The stock buffer
was diluted 1:1.5 with deionized water before use. Approx-
imately 1,200 ml of the diluted buffer were added to each
buffer tank. Platinum electrodes were inserted in each
buffer tank with 200-250 volts (d.c.) applied across the
system for 16 hr, or the time necessary for the amino
black dye marker to move 14 cm past the starting slots.

This system represents discontinuous gel electrophoresis.

43

The gel was sliced transversely with a thin wire
after the electrOphoretic run was completed. The middle
slice was stained for 10 min with an Amido Black staining
solution containing 250 ml of water, 250 ml mechanol, 50
ml glacial acetic acid and 2 g amido black. After
removing the excess stain with water, the unbound stain
was removed from the gel in an electrophoretic destaining

cell containing a 7% acetic acid solution.

Electron Microscopy

The samples were fixed in 1% glutaraldehyde (0.2 m1
of sample containing approximately 1% membrane to 2 m1 of
fixative) for 15 min, diluted with distilled water,
dispersed on glass slides by dipping. The samples on the
glass slides were post-fixed in 050“ vapors for 15 min and
air dried. Finally, the material was shadowed with plat-
inum at a 3:1 angle (arc tan 1/3). The carbon film with
the sample material was floated onto a water surface and
picked up on copper grids. Electron microscopy was per-
formed with a RCA EMU 3 G scOpe operating at 100 kV and
using a 25 u objective aperture.
Preparative Procedures for the Fat
Globule Membrane Fractions

The milk used in this study was obtained from the
Michigan State University dairy herd which consisted of
Brown Swiss, Jersey and Holstein cows. The milk was

collected at the evening milking and separated while warm.

44

Preparation of membranes for washing and erosion
studies.--After the cream was separated, it was diluted
with warm (40° C) deionized water (1:3 v/v) and reseparated.
This process was repeated five times with aliquots of the
cream and corresponding washings collected after each
separation for subsequent analyses. The washed cream
samples were standardized to a volume of 600 ml of 40%
fat, cooled overnight to 5° C and churned in 2,000 ml
Pyrex Erlenmeyer flasks with a rotary shaker operated at
room temperature. Under these conditions the fat emulsion
broke in approximately 45 min. Agitation was continued
until the fat granules were gathered into ball-shaped
masses about one inch in diameter. The membrane material
in the buttermilk was designated as aqueous membrane
whereas the material trapped in the butter granules was
designated as serum membrane. The resulting aqueous
membrane suspensions were filtered through a triple layer
of cheese cloth to trap the small granules of butter and
finally centrifuged for 0.2 hr at 2,000 X g to remove the
unchurned fat globules. Each membrane sample was 1yophil—
ized and dried in a vacuum of 25 inches of Hg and over
P205 at room temperature for 24 hr prior to weighing.

Total lipids were extracted from weighed aliquots of
the dried membrane preparations. For this purpose four
extractions with ethanol/diethyl ether mixtures of the

following compositions were employed at room temperature:

45

40/60, 20/80, 10/90, and 0/100% (v/v)° The residual
proteins were dried as previously described prior to
weighing. Protein percentages were estimated from weight
differences.

Samples were taken from skimmilk and each cream
washing and dialyzed for 24 hr with constant stirring to
remove the milk salts. After these samples were dialyzed
and 1y0philized, 4% suspensions of the residues were
analyzed for plasma proteins by starch-urea gel electro-
phoresis according to the method described by Wake and

Baldwin (1961).

Preparation of membranes for physical and chemical
studies.--The fragmented membranes were prepared from
washed cream in the same manner as described in the
preceding paragraphs except for the following modifica—
tions. The cream was washed 3 times, and the final fat
content was adjusted to 50% before churning for the
purpose of concentrating the membrane suspension. Three
pellets were removed from the aqueous membrane suspension
by centrifugation. These pellets were resuspended,
lyophilized and dried prior to chemical analyses. For the
various lipid determinations the Folch 33 31. (1957)
extraction procedure was used to remove the total lipids
from the pellets. In preparing the proteins for analysis,
however, the pellet lipids were extracted with chloroform-

methanol (2:1 v/v) at the rate of 100 ml of solvent per g

46

of membrane. These mixtures were stirred for 20 min and
centrifuged for 30 min at 2,000 X g to collect the insoluble
proteins. This procedure was repeated three times, and
finally the protein pellets were washed with acetone before
drying.

For electron microscopic examination, the wet
unextracted pellets were resuspended in deionized water at
1% concentration and preserved with streptomycin and
penicillin "G". The antibiotics were added at the rate of

50 mg/liter and 5,000 units/liter, respectively.

Preparation of the 7,5008, 2308 and 358 pellets by

centrifugation.—-A scheme for fractionating the aqueous

 

membrane suspension is presented in Figure l with the
pellet values expressed in Svedberg units. The minimum
sedimentation coefficients of the pellets were chosen so
that the amount of each pellet would be sufficient for
analysis and at the same time provide adequate differential
fractionation. The 7,5008 and 2308 pellets were redis-
persed in 200 to 1 (v/w) deionized water before final
centrifugation. This step allowed particles trapped

during the initial sedimentation with coefficients less
than 7,5008 and 2308 to be removed from the final pellets.
The 358 pellet was not washed since approximately two-
thirds of the membranes had already been removed in the
other two pellets. Additionally, the final supernatant con—

tained only 3 to 4% of the total membrane material.

47

Sedimentation rates of the pellets were determined

in accordance with the equation set forth by Trautman

 

(1961):
1013 1n (r /r )
_ a b
s - 2 ,
W T
where s = the sedimentation coefficient is expressed in
Svedberg units (S),
r3 = the radius from the center of rotation (cm)
to bottom of liquid,
rb = radius from the center of rotation (cm) to
air-liquid meniscus,
T = separation time in seconds, and

W = the angular velocity in rad/sec = 2n (r m) .
50

For example, the calculations of the 7,5008 pellet
employing the Sorvall RC2-B with a type SS-34 rotor with

the following parameters

ra = 10.8 cm,
rb = 5.7 cm,
rpm = 10,000, and
T = 756 sec

would yield

1013 ln (10.8/5.7)

2n (10,000)? . 756
60

s = = 7,5008 pellet.

 

48

The 2308 pellet was obtained using the same equipment
as above but with an increase in (T) and (rpm) as shown in
Figure l. The Beckman Model L-2 ultracentrifuge with the
type 30 fixed—angle rotor was used to collect the 35 S
pellet. The ra and r values for this rotor are 10.5 and

b
5 cm, respectively.

 

49

AQUEOUS MEMBRANE SUSPENSION

(0.3 hr at 2,000 x g)

 

Unchurned Fat

 

 

 

 

 

 

Globules
Discarded
MEMBRANE-CONTAINING SUSPENSION
(0.2 hr at 12,100 X g at 4° C)
7,500 S Floatation Layer
Pellet Discarded
SUPERNATANT
(2.7 hr at 30,900 x g at 4° 0)
230 S Floatation Layer
Pellet Discarded
SUPERNATANT
(6.3 hr at 92,000 X g at 4° C)
35 S FINAL Floatation Layer

Pellet SUPERNATANT Discarded

FIGURE l.--Diagram for the isolation of the membrane pellets
from eroded fat globules of thrice-washed cream. The 7,5008
and 2308 pellets were washed with deionized water (200:1
v/w) before final centrifugation.

RESULTS AND DISCUSSION

Preparative Procedures

 

The Effects of Washing the
Fat‘Globules

 

 

The removal of the milk plasma from the fat globules
before isolating the membranous material is of paramount
importance. As mentioned previously in the procedural
section, this separation is accomplished best by taking
advantage of the density difference between the plasma and
fat globules. With successive dilutions of cream with
water (1:3 v/v) and reseparating, the plasma is adequately.
removed. This removal was monitored by taking dried
samples of each washing and applying them to starch-urea
gel as shown in Figure 2. Although equal amounts of a 4%
concentration of the dried material from the first through
the fifth washing were applied to the gel slots, the band
intensities decreased progressively toward the fifth
washing. Presumably, membrane lipoprotein complexes make
up the majority of the material in the latter slots which
do not stain well with Amino Black or migrate in starch—
urea gel. Undoubtedly, these particulates are erosion

products from the membrane proper.

50

The wash medium used in these studies was warm
(40° C) deionized water. By employing warm water, the
separations were essentially performed at the temperature
of freshly secreted milk; thus, minimizing induced lipolysis
(Thompson, 33 31., 1961). Actually, a saline-sucrose wash
solution preserves the membrane integrity better than
deionized water (Swope and Brunner, In Review). However,
the need for prolonged dialysis and the threat of bacterial
growth in these suspensions precluded its use.

In Figure 3 the results of washing the fat globules
before isolating the membranes are presented. The concen-
tration of the aqueous membrane material isolated from
samples of cream washed from two to five times decreased
from 1.05 g to 0.89 g per 100 g of fat. A loss of
approximately 4% of the membrane occurred per washing.
Also, the membrane protein decreased substantially from
the second to third washing and then stabilized over the
next several washings. At least part of this protein
found in the twice-washed sample is attributable to
plasma proteins since the gel electrOphoretograms showed
that it took three washings to remove them satisfactorily.
The percentage of protein in the membrane is considerably
less in Figure 3 than was found in the various pellets.
However, in these washing experiments the total aqueous
membrane suspension was analyzed which included high lipid
fractions such as the floatation layers; consequently, the

indicated amount of protein was lower.

52

The data show that the membrane is labile to erosion
when employing dilution techniques to remove plasma
material. However, washing fat globules three times with
deionized warm water (1:3 dilution) will adequately remove
the plasma proteins and keep the erosion at acceptable

levels.

Pellet Separation by Ultracentrifugation

The fractionation scheme shown in Figure 1 separates
membrane particles of different sizes and densities into
various fractions by centrifugal force. By no means does
this suggest that each pellet is distinctly different from
the other. In fact, only an enrichment of the various
particulates occurs from one pellet to the next with
differential centrifugation. The pellets collected,
however, were sufficiently different to warrant chemical
and physical analyses, and to a reasonable degree, an
interpretation of the native structure.

Other workers have employed centrifugal separation
techniques to study the globule membrane in conjunction
with solubilizing agents (Alexander and Lusena, 1961;
Hayashi and Smith, 1965). These chemicals were avoided
in this study to decrease the possibility of introducing
artifacts.

Approximately 10—15% of the starting membrane material

was lost in the washing of pellets, the final supernatant

53

and the floatation layers. The washings from the 7,5008
and 2303 pellets were assumed to be representative of the
pellets in general and were discarded. The material in
the final supernatant (approximately 4%) could not be
sedimented in sufficient quantities to analyze. Perhaps
of greater interest is the nature of the floatation layers.
This material has physical and floatation characteristics
similar to plasma chylomicrons (Wathen and Levy, 1966).
Accordingly, Zittle, 33 31. (1956) reported that a similar
substance in milk contained an exceedingly high lipid
content (95%) which indicates a very low density lipo-
protein. It must be assumed that these lipid-rich par—
ticles are orientated to the fat globule in an unspeci-

fied manner and desorbed during the churning process.

Chemical Analyses

The Protein-Lipid Relationship
of’the'Various Fractions

The compositional data of the three pellets are
reported in Table 1. The yield for the 7,5008 pellet is
considerably lower than the other two fractions. Using
143w centrifugal force only, the larger membrane fragments
Euid intact membrane ghosts are sedimented. As this force
218 increased to 2308 and finally 358 speeds, the smaller
Ennd less dense particles are removed from suspension.

'This latter material accounts for 80% of the total pellets.

 

54

Perhaps of more interest is the lipid distribution between
pellets. The lipids increase from 17.5% in the 7,5008
pellet to 55.3% in the 358 pellet. Such a variation indi—
cates that the pellet separation is not based entirely

on particle size but also on particle density.

The ratios of protein, cholesterol and phospholipids
to the total lipids of each pellet are reported in Table
2. The membrane pellets are high in phospholipids which
account for approximately 50% of the total lipids. Phos-
pholipids as well as cholesterol maintain their highest
level in the 7,5008 pellet which contains only 17.5%
lipids. As the lipids of the pellets increase-—29.5% for
the 2308 and 55.3% for the 3SS--the ratios for phospho-
lipids and cholesterol per.total lipid actually decline.

Therefore, the data lend support to the postulate
that the lipids are arranged in globular micelles with the
bimodal molecules of phospholipids and perhaps choles—
terol located at the surface and neutral lipids toward
the interior of the structure. Variations in micellular
size would adequately explain the changes in the
phospholipid-cholesterol/total lipids relationship of the
pellets, since the surface area to volume ratios would

also change.

55

The Protein Composition of
the Different Fractions

 

Various chemical constituents have been identified
in the lipid-extracted pellets and are reported in Table 3.
Most milk proteins contain a nitrogen content between 15.0
and 16.0%, slightly higher than the membrane fractions.
This difference is related to the carbohydrate moiety as
revealed in the hexose, hexosamine and sialic acid content
of the membrane proteins. Carbohydrates in general and
sialic acid specifically appear to be a common constituent
of membranes. It might be Speculated that their presence
in this membrane system is manifested in the aggregation
or clustering of fat globules during creaming (Dunkley
and Sommer, 1944).

Again, the enrichment of pellets with specific
materials by differential centrifugation is amply demon—
strated since the carbohydrate values range from 6.6% in
the 7,5008 to 8.5 and 10.1% in the 2308 and 358 extracted
pellets, respectively. Interestingly, the higher carbo-
hydrate content is found in the higher lipid-containing
pellets, indicating a possible relationship between these
carbohydrates and the lipid micelles.

Small amounts of phosphorus are present in the
extracted pellets, but whether this element is attributable
to residual phospholipids, a phosphorylated protein or

traces of nucleic acids is not known at present.

56

The amino acid residues of the pellet proteins are
tabulated in Table 4. These proteins, collectively, are
similar in amino acid content to other milk proteins
except for a lower glutamic acid and proline and a higher
arginine value. The predominant characteristic feature
of these fractions is their insolubility in the usual
aqueous systems employed to carry proteins. On close
examination, however, the concentration of hydrophobic
amino acid residues is no greater than other proteins
which exhibit a high degree of solubility. It is sus-
pected, therefore, that the sequential arrangement of
amino acids is a significant factor in determining the
solubility properties.

There are no striking differences in the amino acid
composition of the various pellet proteins. When all
three fractions are compared on mole basis (Table 5), the
358 amino acid content exhibit more variation than the
other two pellets. Seemingly, this change suggests a
greater selectivity in protein composition for this high
lipid-containing fraction. A meaningful interpretation of
these results is not possible, however, since the pellets
at present are undefined protein mixtures. Also, it must
be remembered that proteins do not necessarily show a
sizable variation in their amino acid content in order to
reflect considerable differences in their physical and

chemical properties. Consequently, the 7,5008 and 2308

57

pellets may be actually more unlike than their amino acid

composition would seem to indicate.

Proposed Structure for the Fat
Globule Membrane

Supporting Evidence for a Protein-
Lipoprotein Structure

 

In view of the data already presented, the fat
globule membrane appears to consist of a proteinaceous
layer(s) which serves internally as a covering for the
glyceride drOplet and externally as adsorption sites for
lipoprotein micelles. The characteristic "insolubility"
of this major membrane protein gives credence to such a
structural role. Like other structural proteins, its
intermolecular binding forces have not allowed for
definitive physical characterization (Harwalkar and
Brunner, 1965).

King (1955) postulated that these proteins were
held to the lipid surface of the globule by electrostatic
bonding by the polar ends of phospholipids. Recently,
Patton and Fowkes (1967) have proposed that the protein
envelOpment of the droplet can be adequately explained
on the basis of London-van de Waals forces which are
estimated to achieve an attraction force of one atmos-
phere when the separating distance between the membrane
and fat droplet is reduced to 20 A. Therefore, electro-

static bonding would be of little consequence under these

58

conditions. Morton (1954) gave additional evidence for

a protein layer(s) when he prOposed that a continuous

protein membrane surrounds the fat globule to which were

adsorbed microsomal lipoprotein particles. By employing

sodium deoxycholate as a solubilizing agent on the intact

globules, Hayashi and Smith (1965) discovered a water-

insoluble matrix to which water-soluble lipOproteins were

attached.

Substantial evidence is now available to verify the

presence of loosely adsorbed lipOprotein particles on the

membrane proper. The washing experiments, discussed
earlier in this investigation, show a steady erosion of
the membrane during the isolation procedures. Koops 33
(1958), examining the membrane thickness with electron
microsc0py, found that it was dependent on prior treat-
ment of the cream. The membrane was noticeably thicker
in the case of cream allowed to rise naturally. Any
form of agitation apparently causes a migration of mem-
brane phospholipids (Greenbank and Pallansch, 1961).
Even the effect of heating results in approximately 20%
lipid phosphorus dispersion to the aqueous phase (Koops
and Tarassuk, 1959). Recently, Chein and Richardson
(1967a) confirmed the presence of a substantial neutral
lipid content in these migrating lipoprotein particles.
The actual structure required for lipoprotein

stability is a matter of conjecture at this point.

59

However, experimenting with various mixtures of lecithin,
cholesterol and saponin, Lucy and Glauert (1964) observed
a variety of molecular arrangements as interpreted with
the aid of electron microscopy. They concluded that only
the formation of small globular micelles would satisfac-
torily explain these organized units. Eley and Hedge
(1956) had previously substantiated that lipid micelles
bind proteins primarily through hydrogen and electro-
static bonding. The same forces should also be applicable
to the lipoproteins of the fat globule membrane.

The chemical composition and physical prOperties of
the pellet material strongly imply the presence of
adsorbed lipOprotein of various sizes which are similar
in many respects to those found in chyle and blood serum.
Cook and Martin (1962) have surveyed over 40 lipOprotein
fractions and have discovered a definite relationship
between the neutral lipid, phospholipids and protein of
these particles. For instance, when the neutral lipid
increases in these lipoproteins, the phospholipid
decreases. As the protein content rises above 30%, the
neutral lipid phospholipids are present in similar
proportion. All this information seems to indicate the
formation of globular micelles with a surface area
relatively rich in phospholipids and proteins. In the
case of the membrane pellets, the same type composi-

tional relationship was found likely to exist.

 

:5“

60

Interpretation of Electron Micrographs
6T‘7,5008, 2308 and 358 Pellets

To pr0pose a membrane structure based entirely on
electron microscopy would be misleading and perhaps
erroneous. Korn (1966) has aptly pointed out the limi-
tations of this analytical procedure and suggests that
chemical differences among membranes are more definitive
and more significant than the similar appearance of
membranes in electron micrographs. For this reason, the
author has chosen to use three areas of support for
elucidating the fat globule membrane structure: (1) the
chemical data already presented in this investigation;
(2) research contributions of other workers as reviewed
in the literature section; (3) examination of the 7,5008,
2308 and 358 membrane pellets by electron microscopy.
Hopefully, a reasonable membrane model can be drawn by
integrating these areas. Therefore, the schematic inter-
pretation of each pellet is presented alongside an
electron micrograph of the corresponding pellet primarily
as an illustrative aid. Regions that best represent or
describe these interpretations are indicated by markers,
but they do not necessarily reflect in absolute terms
the pictorial interpretations.

A membrane ghost from the 7,5008 pellet is shown
in Figure 4. This pellet is characterized by large, low
lipid—containing fragments. A rough overlapping membrane

section is clearly visible in the electron micrograph.

61

This structure is interpreted in the schematic represen-
tation as consisting of a basal layer of proteins with a
few small and intermediate sized, adsorbed lipOprotein
micelles. The micelles would easily account for the high
phospholipid-cholesterol/total lipid ratio (see Table 2).

In Figure 5 an eroded membrane and a few irregular
shaped particles are present in the electron micrograph,
presumably originating from the membrane proper. Rough
and smooth sections are quite noticeable which may repre-
sent areas of adsorbed and desorbed lipoproteins. One
of these areas might reflect the Opposite side of the
membrane. In either case, the proposed protein-
lipoprotein complex is still tenable. The membrane frag-
ments in the 2308 pellet are generally smaller and less
dense than those found in the 7,5008 pellet. The total
lipid content rises with an accompanying change in the
phospholipid-cholesterol/total lipid ratio.

Finally in the high-speed 35$ pellet (Figure 6),
small membrane fragments exhibiting definite structural
features are found individually and in clusters. The
total lipid, phospholipid and cholesterol content
suggests that they are primarily large micelles. The
"insoluble" membrane protein is still present in this
fraction denoting the inclusion of basal material. The

carbohydrate content has the highest value of any fraction,

 

62

indicating that at least one of the proteins associated
with the micelles has a high carbohydrate moiety (see
Appendix).

In the foregoing discussion various aspects of
membrane and lipOprotein structure have been reviewed with
particular emphasis placed on the forces required for their
formation and stability. The physical behavior and quan-
titative analyses of the fat globule membrane imply that
these same forces are also responsible for its integrity.
The schematic interpretations (Figures 4-6), however,
represent the simplest possible arrangement for this
material. Visually, this membrane is portrayed as a pro-
teinaceous layer held tightly to the lipid core by
London-van der Waals forces. Externally, lipoprotein
micelles of various size and composition are adsorbed to
this layer through the influence of electrostatic and
hydrogen bonding.

Even with this plausible model of the fat globule
membrane, many questions concerning its structure and
composition remain unanswered. For example, the basal
protein might be arranged in a mosaic pattern leaving
exposed interstices by which the lipoproteins are
anchored to the lipid core. The data presented could
accommodate suchzimodel with slight modification in

chemical bonding. Additionally, the lipoprotein micelles

63

are exceedingly rich in the enzymes xanthine oxidase and
alkaline phosphatase for which there is no adequate
explanation. Finally, no distinction has been made between
the lipoproteins of the various fractions of milk. All

the lipoproteins may or may not be derived from the mem-
brane complex.

In conclusion, the fat globule membrane must be
considered a heterogeneous system which contains a
structural protein, lipoprotein micelles and other sub-
stances closely identified with the secretory cells of
the mammary tissue. The precise nature of this material
has not yet been defined in specific terms, and addi-
tional research will be required to completely unravel

the mystery surrounding it.

.—
.‘ '_,' a

64

TABLE l.--Appearance and composition of the various membrane
pellets obtained from fifteen hundred milliliters of fat
globules.a

 

 

Mass of Mass of Lipids in
Pellet pellet lipid pellet Appearance
_ (s) (g) (%)
7,500 S 1.77 0.31 17.51 white
230 S 4.44 1.31 29.50 white-brown
35 s 3.13 1.73 55.27 brown
TOTAL 9.34 3.35 35.87

 

aThree liters of 50% washed cream.

-
v

65

121.

TABLE 2.--Re1ationship of protein, cholesterol and phospho-
lipid content to the total lipids for each fraction.8

 

 

 

 

Pellet Protein Cholesterol Phospholipidb
lipid lipid lipid
7,500 S 4.71 0.041 0.632
230 s 2.39 0.030 0.532
35 S 0.81 0.028 0.421

 

8Expressed as component ratios.

bPhosphOlipid P x 25.

' ‘TABLE 3.--Composition of the lipid-extracted pellets.a

66

 

Constituent

 

7,500 S 230 S 35 S

(%)
Nitrogen 14.14 14.05 14.01
Phosphorus 0.16 0.17 0.23
Hexose 2.82 3.49 4.15
Hexosamine 2.48 3.22 4.17
Sialic acid 1.29 1.74 1.77

 

aProtein

residues.

F4

67

 

TABLE 4.--Amino acid composition of the protein fractions.a

 

Residue 73500 S 230 S 35 s
(g of residue/100 g protein)bac

 

Lysine 6.6 6.6 5.8
Histidine 2.1 2.3 2.7
Arginine 6.6 6.5 5.5
Aspartic acid 7.7 7.6 9.0
Threonine 4.7 4.9 5.8
Serine 5.4 5.6 5.2
Glutamic acid 11.3 11.0 10.6
Proline 5.1 5.4 4.3
Glycine 3.2 3.3 3.5
Alanine 3.9 3.9 4.0
1/2 Cystine 1.5 1.4 1.9
Valine 4.5 4.4 4.5
Methionine 1.9 1.4 1.2
Isoleucine 4.4 4.3 4.5
Leucine 8.2 8.1 7.4
Tyrosine 3.6 3.5 3.9
Phenylalanine 5.3 5.5 5.3
Tryptophan» 2.2 2.3 2.4
Amino acid residue

(weight %) 88.2 88.0 87.5
Total carbohydrate

(weight %) 6.6 8.5 10.1
P as H2PO3 0.4 0.4 0.6
Total (weight %) 95.2 96.9 98.2

 

aPellet protein residues following lipid extraction.

bWeight percentage of each amino acid residue based on
a nitrogen content of 14.14%, 14.05%, and 14.01% for each
respective pellet.

0Amino acid content based on duplicate analyses of 20
and 70 hr digestion with values corrected for destruction
during hydrolysis according to the equation given by Hirs,
Stein, and Moore (1954).

68

TABLE 5.--Amino acid composition of the protein fractions
expressed in moles.a

 

 

7,500 S 230 S 35 S

Re51due (moles/1000 moles)

Lysine 65.2 i 0.8 65.0 i 0.2 57.5 i 0.1
Histidine 19.6 i 0.8 21.1 i 0.4 25.2 i 0.1
Arginine 53.7 i 0.5 52.4 i 0.2 44.3 i 0.4
Aspartic aCid 84.1 i 0.1 84.1 i 0.6 98.4 i 1.6
Threonine 58.3 i 0.3 61.4 i 0.1 72.6 i 0.3
Serine 77.8 i 0.5 81.5 i 1.2 75.3 i 0.9
Glutamic acid 110.5 1 0.3 107.6 i 0.2 103.6 i 0.5
Proline 65.7 i 0.7 70.3 i 1.3 56.6 i 0.7
Glycine 71.6 i 0.6 72.5 i 0.4 78.5 i 0.4
Alanine 69.9 i 0.8 69.4 i 0.4 71.0 i 0.1
1/2 Cystine 18.8 i 0.9 17.1 i 0.3 23.7 i 1.2
Valine 57.6 i 0.6 55.7 i 0.8 57.1 i 0.3
Methionine 18.4 i 0.2 13.9 i 0.6 12.1 i 0.2
Isoleucine 49.3 i 0.5 47.9 i 0.5 50.4 i 0.7
Leucine 91.2 i 0.3 90.1 i 0.5 82.7 i 1.4
Tyrosine 28.1 i 0.5 27.2 i 0.2 30.1 i 0.4
Phenylalanine 45.2 i 0.3 47.4 i 0.2 45.1 i 0.6
Tryptophan 15.0 i 0.1 15.4 i 0.1 15.8 i 0.2

EN = 1000

 

aAmino acid content based on duplicate analyses of 20
and 70 hr digestion with values corrected for destruction
during hydrolysis according to the equation given by Hirs,
Stein, and Moore (1954). The deviation from the mean is
shown for each value. ‘

69

.mQOfipmnpcwoeoo a: ma
omfiadgm who: mOHQEwm .mwcanmmz Smoke cam xHHEEme
mo memmwouosocqompOOHo How doesnnopmpmll.m osdwfim

x_::E_xm

 

2-2.3; E:

 

m:_;mm§ ucooom
2:525 2::

 

 

 

m:_cmm31mw.:oa .me .w

9.2:; 5E ....

i. 2%.... ll: :_>:..,I-.-,

 

O
7.

.Emmpo pocwmz moEHp O>Hw On 035 m mo mposwflam Eopm Oopo>0oop

 

 

mcowpowwg OcmppEoE msomsvm one mo COHpHmOQEOO mmopw pew pamflw!!.m opzwwm
3:33 .o .3532
m e m N
_ _ - .
36 I O!!!!!!!!! 8
7.0!!!

on... .. we
.w a... - 3
n
m
was- 3%
m . ..m
M mo 9 I M
a I.
M 3 o 7 RM
7 m
flw.oogu.u «mad
m . w
..o... 5.. 7 3 m
...... m

8.. 7 cm

3.. .. 8: 228.28”. ........ mm

8.... E t 28. Bowl
8.. .. 8

 

 

 

l
7

.xooo.mm .mOHHOOfiE GfiOQOAOOQHH omnhompw wQH:HMp:oo Osmansos
mo mono cm mopmofipsa hexane one .uoaaoa moom.> on» 8099

omumfiomfi pmonw osmmnEoE wasnofiw 9mm stop a mo scanneonqsopcfi
oapdEonow WCHOQOQmonLOO m can newsw0h0fifi cosuooam!!.z wmswfim

 

72

.xooo.mm .oswhnEoE can no Noam

nuooEm cam nwson m Op mchOQ poxmms one .OOHHOQ momm on» Eosm
ompwaomfi ocmanoE wasnoaw pmu consoewwhm a mo coapmposQLOOCH
oapmamnom wGHOCOQmOAAOO m 05m namLmOLOHE copuooam!!.m osdwfim

 

 

3
7.

.xooo.mm .pmzmH onwanEmE Hanan on» maficfimpcoo

zanmnopd mmmhm oocfimoo mmoa on cam mOHHOOHE :HouopQOQfiH ponpom
lop owhmH op chOQ .zHHmsufl>HUCH .mhothe one .poafiom mmm

mnp Eonm umpmaoma mucoEthm mumanmE HHwEm mo coapmpoaQsOch
OprEosom MQfipcoammphoo m ocm ammmwOLOflE copuooam!!.m mmswfim

 

SUMMARY

Washing and reseparating fat globules three times with
three volumes of deionized water per volume of cream
will adequately remove the milk plasma.

Material was lost from the membrane proper at an

uniform rate during the preparative procedures.

The total lipids increased from the 7,5008 pellet (17.5%)
and reached their highest value in the 358 (55.3%) pellet.
The total lipids of the 7,5008 pellet contained 4%
cholesterol and 63% phospholipids, but these values
decreased to 2.8% cholesterol and 42.0% phospholipids in
the 358 pellet.

The proteins from the 7,5008 pellet contained 6.6%
carbohydrates while the 2308 and 358 contained 8.5 and
10.1%, respectively.

The pellet proteins contained more arginine and less
glutamic acid and proline than other milk proteins.

The protein pellet from the 358 exhibited better sol-
ubilizing properties than the 7,5008 and 2308 lipid-
extracted pellets.

A membrane model is proposed utilizing the physical

and chemical data collected during this study.

74

BIBLIOGRAPHY

75

BIBLIOGRAPHY

Abderhalden, E., and VOltz, W.
1909. Beitrag zur Kenntnis der Zusammensetzung und
der Natur Hfillen der Milchkfigelchen. HOppe-

Seyl. Z., 22:13.

Alexander, K. M., and Lusena, C. V.
1961. Fractionation of the lipoproteins of the fat
globule membrane from cream. J. Dairy
Sci., 22:1414.

Andrews, P., Bray, R. C., Edwards, P., and Shooter, K.V.
1964. The chemistry of xanthine oxidase. II. Ultra—
centrifuge and gel-filteration studies on
the milk enzyme. Biochem. J., 22:627.

Ascherson, F. M.

1840. On the physiological utility of the fats
and a new theory of cell formation based on
their cooperation and suggested by several
new facts. A translation by Emil Hatschek
in The Foundations of Colloid Chemistpy,

E. Benn, Ltd., London, 1925.

Avis, P. G., Bergel, F., Bray, R. C., James, W. F., and
Shooter, K. V.
1956. Cellular constituents. The chemistry of
xanthine oxidase. Part II. The homogeneity
of crystalline metalloflavoprotein fractions.
J. Chem. Soc., 1212.

Babcock, S. M.‘
1885. A study of the fat globules of milk. N. Y.
(Geneva) Agr. Expt. Sta., 4th. Ann. Rpt. 293.

Babcock, S. M.
1889. Fibrin in milk.
Wis. Agr. EXpt. Sta. 6th. Ann. Rpt., 63.

Bailie, M. J., and Morton, R. K.
1958a. Comparative properties of microsomes from
cow's milk and from mammary gland. I.
Enzymic activities. Biochem. J., 32:35.

Bailie, M. J., and Morton, R. K.
1958b. Comparative properties of microsomes from
cow's milk and from mammary gland. 2. Chemical
composition. Biochem. J., 32:44.

76

 

10.

ll.

l2.

13.

14.

15.

16.

17.

18.

77

Bargmann,"W., Fleischhauer, K., and Knoop, A.
1961. Uber die Morphologie der Milchsekretion.
II. Zugleich eine Kritik am Schema der
Sekretionsmorphologie.
Z. Zellforsch. 22:545.

Bargmann,”W., and Knoop, A.
1959. Uber die Mrophologie der Michsekretion.
I. Licht- und Electronenmikroskopische
Studien an der Milchdrfise der Ratte.
Z. Zellforsch. 22:344.

Bauer, H.
1911. Untersuchungen fiber Oberflachenspannungs-
verhaltnisse in der Milch und fiber die
Natur der Hfillen der Milchfettkfigelchen.
Biochem. Z., 22:362.

Bingham, E. W., Jasewicz, L., and Zittle, C. A.
1961. Acid phosphatase in cream and skimmilk.
J. Dairy Sci., 22:1247.

Bingham, E. W., and Zittle, C. A.
1964. Ribonuclease of bovine milk: Purification
and properties. Arch. Biochem. BiOphys.,
l2§:235.

Bowman, R. E., and Wolf, R. C.
1962. A rapid and specific ultramicro method for
total serum cholesterol. Clin. Chem., 8:302.

Brunner, J. R. .
1962. Protein-lipid interactions and their
relation to the physical-chemical stability
of concentrated milk. A review.
J. Dairy Sci., 22:943.

Brunner, J. R.
1965. Physical equilibria in milk: The lipid
phase. In Fundamentals of Dairy Chemistry
(B. H. Webb and A. H. Johnson, eds.).
Avi Publishing Co., Inc., Westport, Conn.

Brunner, J. R., Duncan, C. W., and Trout, G. M.
1953a. The fat-globule membrane of nonhomogenized
and homogenized milk. I. The isolation and
amino acid composition of the fat-membrane
proteins. Food Res., 28:454.

.254

3:.- ~
-

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

78

Brunner, J. R., Duncan, C. W., and Trout, G. M.
1953c. The fat-globule membrane of nonhomogenized
and homogenized milk. III. Differences in

the sedimentation diagrams of the fat—

membrane proteins. Food Res., 22:469.

Brunner, J. R., and Herald, C. T.

1958. Some electrOphoretic characteristics of two
protein components of the fat—globule mem-
brane of normal milk.

J. Dairy Sci., 2i:1489.

Brunner, J. R., Lillevik, H. A., Trout, G. M., and
Duncan, C. W.
1953b. TWKB fat-globule membrane of nonhomogenized
and homogenized milk. II. Differences in the
electrophoretic patterns of the fat-
membrane proteins. Food. Res., l§:463.

Cerbulis, J., and Zittle, C. A.
1965. Lipids associated with acid-precipitated
casein. J. Dairy Sci., 22:1154.

Chien, H. C., and Richardson, T.
1967a. GIWDSS structure of the fat globule membrane
of cow's milk.. J. Dairy Sci., 22:451.

Chien, H. C., and Richardson, T.
1967b. Gxnoss structure of the fat globule membrane
of cow's milk. II. Distribution of amino
acids, hexosamines, ribonucleic acid, copper
and iron. J. Dairy Sci., 22:1868.

Cook G. M. W., Heard, D. H., and Seaman, G. V. F.
1961. Sialic acids and the electrokinetic charge
of the human erythrocyte. Nature, 191:44.

Cook, W. H., and Martin, W. G._
1962. Composition and prOperties of soluble
lipoproteins in relation to structure.
Can. J. Biochem. Physiol. 22:1273.

Cornwell, D. G., and Horrocks, L. A.
1964. Protein-lipid complexes. In Symposium
on Foods: Proteins and Their Reactions
(H. W. Schultz and A. F. Anglemier, eds.),
The Avi Publishing Co. Inc., Westport, Conn.

Coulson, E. J., and Stevens, H.
1950. The serological relationship of bovine whey
albumin to serum albumin. J. Biol. Chem.,
£23355.

F‘— -

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

79

Danielli, J. F., and Davson, H.
1934. A contribution to the theory of permeability
of thin films. J. Cellular Comp. Physiol.,

5:495.

Davies, W. L.
1933. The mode of combination and distribution of
traces of heavy metals in dairy products.
J. Dairy Res., 2:255.

DeRobertis, E. D. P., Nowinski, W. W., and Saez, F. A.
1965. Cell Biology. W. B. Saunders Co., Phila.,
Penn.

 

Dixon, F. J., Weigle, W. 0., and Vazquez, J. J.
1961. Metabolism and mammary secretion of serum
proteins in the cow. Lab. Invest. 22:216.

Dowben, R. M., Brunner, J. R., and Philpott, D. E.
1967. Studies on milk fat globule membranes.
Biochim. Biophys. Acta,l35:1.

Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers,
P. A., and Smith, F.
1956. Colorimetric method for determination of
sugars and related substances. Anal.
Chem., 22:350.

Dunkley, W. L., and Sommer, H. H.
1944. The creaming of milk. Wis. Agr. Expt.
Sta. Res. Bull., 151.

Eley, D. D., and Hedge, D. G.
1956. Protein interactions with lecithin and
cephalin monolayers. J. Colloid Sci.,
22:445.

Emmelot, P., Bos, C. J., Benedetti, E. L., and

Rfimke, P.

1964. Studies on plasma membranes. 1. Chemical
composition and enzyme content of plasma
membranes isolated from rat liver. Biochim.
Biophys. Acta, 22:126.

Erickson, D. E., Dunkley, W. L., and Smith, L. M.
1964. Tocopherol distribution in milk fractions and
its relation to antioxidant activity.
J. Food Sci., 22:269.

Folch, J., Lees, M., and Sloane-Stanley, G. H.
1957. A simple method for the isolation and
purification of total lipids from animal
tissues. J. Biol. Chem., 222:497.

40.

“lo

42.

43.

an.

45.

46.

47.

48.

80

Folley, S. J., and White, P.

1936. The effects of thyroxine on milk secretion
and on the phosphatase of blood and milk
of the lactating cow. Proc. Roy. Soc.
(London), Ser. B, 222:346.

Glascock, R. F., Welch, V. A., Bishop, C., and Davies, T.
1964. The contribution of serum beta-lipoprotein
triglycerides to milk fat in the dairy cow.
Proc. Symp. Radioisotopes Amin. Nutr.
Physiol., 253.

Glascock, R. F., Welch, V. A., BishOp, C., Davies, T.,
Wright, E. W., and Noble, R. C.
1966. An investigation of serum lipoproteins and of
their contribution to milk fat in the dairy
cow. Biochem. J.2§:149.

Graham, W. R. Jr., Jones, T. S. G., and Kay, H. D.
1936. The precursors in cow's blood of milk fat
and other milk constituents. Proc. Roy.
Soc. (London), Ser. B, 222:330.

Graham, W. R. Jr., and Kay, H. D.

1933. Phosphorus compounds of milk. V. The phos-
phorus partition in milk with preliminary
observations on milk phOSphatase. J. Dairy
Res., 2:54.

Green, D. E., and Tzagoloff, A.
1966. Role of lipids in the structure and function
of biological membranes. J. Lipid Res.,
1:587.

Greenbank, G. R., and Pallansch, M. J.
1961. Migration of phosphatides in processing
dairy products. J. Dairy Sci., 22:1597.

Hare, J. H., Schwartz, D. P., and Weese, S. J.
1952. The amino acid composition of the fat
globule membrane protein of milk. J. Dairy
Sci., 22:615.

Hattori, K.

1925. Membrane of the fat globule in milk. J.
Pharm. Soc., Japan, No. 516:123. (Original
not seen. Cited by C. T. Herald, Ph.D.
thesis. Michigan State University. 1956).

 

49.

50.

51.

52.

53.

54.

55.

56.

57.

81

Hayashi, S., Erickson, D. R., and Smith, L. M.
1965. Membranous material of bovine milk fat
globules. II. Some physical and enzymic
properties of the deoxycholate-released
lipoproteins. Biochemistry, 2:2557.

Hayashi, S., and Smith, L. M.

1965. Membranous material of bovine milk fat
globules. I. Comparison of membranous
fractions released by deoxycholate and by
churning. Biochemistry, 2:2550.

Herald, C. T., and Brunner, J. R.
1957. The fat- -globu1e membrane of normal cow's
milk. I. The isolation and characteristics
of two membrane-protein fractions. J.
Dairy Sci., 22:948.

Herald, C. T., Brunner, J. R., and Bass, S. T.
1957. A Spectrographic analysis of two protein
fractions of the fat-globule membrane of
normal cow's milk. J. Dairy Sci., 22:446.

Hirs, C. H. W., Stein, W. H., and Moore, 8.
1954. The amino acid composition of ribonuclease.
J. Biol. Chem., 2112941.

Ibbott, F. A., and Abrams, A.
1964. The phospholipids in membrane ghosts from
Streptococcus faecalis protOplasts. Bio-
chemistry, 2:2008.

Jack, E. L., and Dahle, c. D.
1937. The electrokinetic potential of milk fat.
III. Relation to the fat globule membrane.
J. Dairy Sci., 22:637.

Jackson, R. H., Coulson, E. J., and Clark, W. R.
1962. The mucoprotein of the fat/plasma interface

of cow's milk. I. Chemical and physical
characterization. Arch. Biochem. BiOphys.,

£3373.

Jenness, R., and Palmer, L. S.
1945a. Substances adsorbed on the fat globules in
cream and their relationship to churning.
V. Composition of the "membrane" and dis-
tribution of the adsorbed substances in
churning. J. Dairy Sci., 22:611.

Ia '* ‘oh. ' c 'I
I.
Q

58.

59.

60.

61.

62.

63.

64.

65.

66.

57.

82

Jenness, R., and Palmer, L. S.
1945b. Substances adsorbed on the fat globules in
cream and their relationship to churning.
VI. Relation of the high-melting triglyceride
fraction to butterfat and the "membrane."
J. Dairy Sci., 22:653.

Johansen, P. G., Marshall, R. D., and Neuberger, A.

1960. Carbohydrates in proteins. 2. The hexose,
hexosamine, acetyl and amide-nitrogen content

of hen's—egg albumin. Biochem. J. 11:239.

King, N. '
1955. The Milk Fat Globule Membrane and Some
Associated Phenomena. Commonwealth Agricul-
tural Bureaux, Farnham Royal, Bucks, England.

 

King, R. L., and Dunkley, W. L.
1959. Relation of natural COpper in milk to inci-
dence of spontaneous oxidized flavor. J.
Dairy Sci., 22:420.

King, R. L., Luick, J. R., Litman, I. 1., Jennings,
W. G., and Dunkley, W. L.
1959. Distribution of natural and added copper and
iron in milk. J. Dairy Sci., 22:780.

Knoop E., Wortman, A., and Knoop, A. M.
1958. Die Hfillen der Milchfettkfigelchen im Elek-
tronenmikroskopischen Bild. Naturwissen-
schaften, 22:418.

Kon, S. K., Mawson, E. H., and Thompson, S. Y.

1944. Relation of the concentration of vitamin A,
carotenoids and cholesterol in milk fat to
the size of the fat globules. Nature
222282.

Kono, T., and Colowick, S. P.
1961. Isolation of skeletal muscle cell membrane
and some of its properties. Arch. Biochem.
Biophys., 22:520. '

Koops, J., and Tarassuk, N. P.

1959. The effect.of various processing treatments
on the partition of phosphatides between the
fat phase and the milk plasma. Neth. Milk
Dairy J., 22:180.

Koppejan, C. A., and Mulder, H.
1953. The distribution of copper and iron over the
various phases of milk. Intern. Dairy Congr.
XIII, The Hague, 2, 1402.

 

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

83

Korn, E. D.
1966. Structure of biological membranes.
Science, 153:1491.

Larson, B. L.
1965. Biosynthesis of the milk proteins.
J. Dairy Sci., 22:133.

Larson, B. L., and Gillespie, D. C.
1957. Origin of the major Specific proteins in
milk. J. Biol. Chem., 227:565.

Long, C.
1961. In Biochemists' Handbook. (C. Long, ed.)
D. VanNostrand’Co., Inc., Princeton, N. J.

Lucy, J. A., and Glauert, A. M.
1964. Structure and assembly of macromolecular
lipid complexes of globular micelles.
J. Mol. Biol., 2:727.

Luick, J. R.
1961. Synthesis of milk fat in the bovine mammary
gland. J. Dairy Sci., 22: 652.

Marier, J. R., Tessier, H., and Rose, D.
1963. Sialic acid as an index of the k-casein of
bovine milk. J. Dairy Sci., 22:373.

Matsushita, 8., Ibuki, F., Mori, T., and Hata, T.
1965. Phosphodiesterase in microsomes from bovine
milk. Agr. Biol. Chem., 22:436.

McCarthy, R. D., Ghiardi, F. L. A., and Patton, S.
1965. The conversion of stearic acid to oleic in
freshly secreted milk. Biochem. Biophys.
Acta, 22:216.

McMeekin, T. L., and Polis, B. D
1949. Milk proteins. In Advances in Protein
Chemist_y. Vol. 5. (M. L. Anson, John T.
Edsall, and Kenneth Bailey, eds.), Acad.
Press, Inc., New York.

 

Moore 8., and Spackman, D. H., and Stein, W. H.
1958. Chromatography of amino acids on sulfonated
polystyrene resins. Anal. Chem., 2221185.

Morton, R. K.
1953a. Microsomal particles of normal cow's milk.
Nature, 171:734.

80.

81.

82;

83.

84.

85.

86.

870

88.

89.

90.

84

Morton, R. K.
1953b. Alkaline phosphatase of milk. 1. Association
of the enzyme with a particulate lipoprotein
complex. Biochem. J., 22:786.

Morton, R. K. '
1953c. Alkaline phosphatase of milk. 11. Purifi-
cation of the enzyme.. Biochem. J., 22:795.

Morton, R“ K;
1954.. The lipoprotein particles in cow's milk.
Biochem. J. 57: 231.

Mulder, H.
1957. The surface layers of milk fat globules. I.
Neth. Milk Dairy J., 22:197.

Mulder, H., and Menger, J. W.
1958. The surface layers of milk fat globules.
III. The protein content of the surface
layers. Neth. Milk Dairy J., 22gl.

Mulder, H., Menger, J. W., and Koops, J.
1957. The surface layers of milk fat globules.
II. The phospholipid content of the surface
layers. Neth. Milk Dairy J., 22:263.

Mulder, H., and Zuidhof, T. A.
1958. The surface layers of milk fat globules.
IV. The cholesterol content (free and ester-
ified cholesterol) of the surface layers.
Neth. Milk Dairy J. 22:173.

Muralt, G. von, Gugler, E., and Roulet, D. L. A.
1964. Immuno-electrophoretic studies of the
proteins of human milk and colostrum.
Immuno-Electrophoretic Analysis (P. Grabar
and P. Burtin, eds.), Elsevier Publishing
Co., New York.

Novikoff, A. B., Podber, E., Ryan, J., and Noe, E.
1953. Biochemical heterogeneity of the cytoplasmic
particles isolated from rat liver homogenate.
J. Histochem. Cytochem. 2:27.

Owen, E. C., and Hart, L. I.

1962. Xanthine oxidase in human milk. The nutrition
society's meeting for original communications.
Abstracts of communications, Proc. Nutr.

Soc., 22:15.

Paape, M. J., and Tucker, H. A.
1966. Somatic cell content variation in fraction-
collected milk. J. Dairy Sci., 22:265.

 

91.

92.

93.

94.

95.

96.

970

98.

99.

85

Palmer, L. S.
1944. The structure and properties of the natural
fat globule "membrane." J. Dairy Sci.,
22:471.

Palmer, L. S., and Samuelsson, E.
1924. The nature of the substances adsorbed on the
surface of the fat globules in cow's milk.
Proc. Soc. Exptl. Biol. Med., 22:537.

Palmer, L. S., and Tarassuk, N. P.
1936. The effect of the adsorption "membrane"
around the fat globules on curd tension of
cow's milk. J. Dairy Sci., 22:323.

Palmer, L. S., and Tarassuk, N. P.

1940. Physical and chemical properties of the fat
globule adsorption "membrane." II. Nature and
origin of surface active materials involved
in curd tension reduction and prevention of
rennet clot of cow's milk by "membranes"
from natural and synthetic creams. J. Dairy
Sci., 22:861.

Palmer, L. S., and Wiese, H. F.
1933. Substances adsorbed on the fat globules in
cream and their relation to churning. II.
The isolation and identification of adsorbed
substances. J. Dairy Sci., 22:41.

Parpart, A. K., and Ballentine, R.
1952. Molecular anatomy of the red cell plasma
membrane. In Modern Trends in Physiolo
and Biochemistry (E. S. G. Barron, ed. ,
Acad. Press., New York.

 

Patton, S., Durdan, A., and McCarthy, R. D.
1964. Structure and synthesis of milk fat. VI.
Unity of the phospholipids in milk. J.
Dairy Sci., 22:489.

Patton, S., and Fowkes, F. M.
1967. The role of the plasma membrane in the
secretion of milk fat. J. Theoret. Biol.,
22:274.

Patton, S., and Keeney, P. G.
1958. The high-melting glyceride fraction from
milk fat. J. Dairy Sci., 22:1288.

100.

101.

102.

103.

104.

1050

106.

107.

108.

109.

86

Patton, S., McCarthy, R. D., and Dimick, P. S.
1965. Structure and synthesis of milk fat.
IX. Site of lipid synthesis in freshly
secreted milk. J. Dairy Sci., 22:1389.

Payens, T. A. J.
1959. Model experiments on lipoprotein inter-
action in the fat globule membrane of
milk. Neth. Milk Dairy J. 22:237.

Payens, T. A. J., Koops, J., and Mogot, M. F. K.
1965. Adsorption of euglobulin on agglutinating
milk fat globules. Biochim. Biophys. Acta,

22:576.

Peereboom, J. W. C.

1966. Studies on alkaline milk phosphatase. I.
Non-identity of raw and reactivated
alkaline phosphatases from cream. Neth.
Milk Dairy J., 22:9.

Perlman, J. L.
1935. The distribution of phospholipids in cream.
J. Dairy Sci., 22:113.

Polis, B. D., and Shmukler, H. W.
1950. Aldolase in milk. J. Dairy Sci., 22:619.

Polis, B. D., Shmukler, I{.1N., and Custer, J. H.
1950. Isolation of a crystalline albumin from
milk. J. Biol. Chem., 187:349.

Rhodes, D. N., and Lea, C. H.
1958. On the composition of the phospholipids of
cow's milk. J. Dairy Res., 22:60.

Rimpila, C. E., and Palmer, L. S.
1935. Substances adsorbed on the fat globules in
cream and their relation to churning. IV.
Factors influencing the composition of the
adsorption "membrane." J. Dairy Sci.,
22:827.

Rodkey, F. L., and Ball, E. G.
1946. A rapid test for distinguishing human from
cow's milk based upon the difference in
their xanthine oxidase content. J. Lab.

Clin. Med. , 32:354.

110.

111.

112.

113.

114.

115.

116.

1170

118 O

119.

120 O

87

Roland, F.
1956. Die Quantitative Bistimmung der Hfilleneiweiss-
Substanz der Milchfettkfigelchen.

Milchwissenschaft, 22:85.
Rondle, C. J. M., and Morgan, W. T. J.
1955. The determination of glucosamine and galac-
tosamine. Biochem. J. 22:586.

Samuelsson, E. G.

1966. The copper content in milk and the distri-
bution of copper to some various phases of
milk. Milchwissenschaft, 22:335.
Sasaki, R., and Koyama, S.

1956. On the fat globule membrane materials of
cow's milk. 1. Isolation of membrane protein
and its electrophoretic prOperties. Intern.

Dairy Congr. XIV, Rome, 2:381.

Sasaki, R., and Koyama, S.

1959. On the fat-globule membrane materials of
cow's milk. III. ElectrOphoretic and
enzymatic properties of membrane lipoprotein
and its relation to other lipoproteins in
creag. Intern. Dairy Congr. XV, London,
1:30 .

Schwarz, G.
1947. Die Hfillen der Fettkfigelchen.
Molkereiztg. Hildesheim, 2:33.

Neue

Sharp, P. F., and Hand, D. B.
1940. Amount of f1avoprotein on the surface of
milk fat globules. N. Y. (Cornell) Agr.
Expt. Sta. Rpt., 121.

Smith, G. J., and Dunkley, W. L.
1961. The mechanism of spontaneous peroxidation
in milk. J. Dairy Sci., 22:1152.

Sommer, H. H.
1951. The fat emulsion in milk from a chemical
standpoint. The Milk Dealer 22 (l):58.

Spies, J. R.
1967. Determination of tryptophan in proteins.
Anal. Chem., 22:1412.

Spies J. R., and Chambers, D. C.
1948. Chemical determination of tryptophan.
Anal. Chem., 22:30.

 

88

'.: .,
4.2—5-

121. Storch, V. j
1897. On the structure of "fat globules" in cow's
milk. (Translated and communicated by H.
Faber). Analyst, 22:197.

122. Sumner, J. B.
1944. A method for the colorimetric determination
of phosphorus. Science, 100:413.

123. Swanson, A. M., and Sommer, H. H.

1940. Oxidized flavor in milk. I. Effect of the
development of oxidized flavor on the iodine
number of the phospholipid fraction of milk.
J. Dairy Sci., 22:201.

124. Swope, F. C., and Brunner, J. R.
1965. Identification of ribonucleic acid in the
fat-globule membrane. J. Dairy Sci.,
22:1705.

125. Swope, F. C., and Brunner, J. R.
The fat globule membrane of cow's milk: A
reassessment of isolation procedures and
mineral composition. In review.

126. Sw0pe, F. C., Brunner, J. R., and Vadehra, D. V.
1965. Riboflavin and its natural derivatives in
the fat-globule membrane. J. Dairy Sci.,
48: 1707.

127. Tarassuk, N. P., and Frankel, E. N.
1957. The specificity of milk lipase. IV. Parti-
tion.of the lipase system in milk. J. Dairy
Sci., 40:418.

128. Tarassuk, N. P., and Koops, J.

1960. Inhibition of oxidized flavor in homogenized
milk as related to the concentration of
copper and phospholipids per unit of fat
globule surface. J. Dairy Sci., 22:93.

129. Tarassuk, N. P., Koops, J., and Pette, J. W.
1959. The origin and development of trainy (fishy)
flavor in washed cream and butter. 1.
Factors affecting the develOpment of trainy
flavor in washed cream. Neth. Milk Dairy
J.,.22z258.
13o. Tayama,.I. ‘
1932. A simple method for the isolation of xanthine
oxidase from milk. J. Biochem. (Japan),
131.

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

89

Thompson, M. P., and Brunner, J. R.
1959. The carbohydrates of some glycoproteins of
bovine milk. J. Dairy Sci., 42:369.

Thompson, M. P., Brunner, J. R., and Stine, C. M.
1959. Characteristics of high-melting triglyceride
fractions from the fat-globule membrane and
butter oil of bovine milk. J. Dairy Sci.,
42:1651.

Thompson, M. P., Brunner, J. R., Stine, C. M., and
Lindquist, K.
1961. Lipid components of the fat-globule membrane.
J. Dairy Sci., 33:1589.

Titus, R. W., Sommer, H. H., and Hart, E. B.
1928. The nature of the protein surrounding the
fat globules in milk. J. Biol. Chem.,

16:237.

Tjepkema, R., and Richardson, T.

1966. Electrophoretic mobility of milk fat globules.
III. Effect of diluent on mobility of fat
globules and fat globule membrane. J. Dairy
Sci., 49:1120.

Trautman, R.

1963. Computational methods of ultracentrifugation.
In Ultracentrifugal Analysis in Theory and
Experiments (J. w. Williams, ed.). Acad.
Press., New York.

 

Vasié, J., and Man, J. M. de
1966a. The composition of the liquid phase of milk
fat obtained by a pressure filteration method.
Intern. Dairy Congr. XVII, Munich, C, 149.

Vasié, J., and Man, J. M. de
1966b. High melting glycerides and the milk fat
globule membrane. Intern. Dairy Congr.
XVII, Munich, c, 167.

VBltz w.
1904. Untersuchungen fiber die Serumhfillen der
Milchkfigelchen. Pflug. Arch. Ges. Physiol.,

102:373.

Wake, R. G., and Baldwin, R. L.
1961. Analysis of casein fractions by zone elec-
trophoresis in concentrated urea. Biochim.

Biophys. Acta. 41:225.

 

141.

142.

143.

144.

145.

146.

147.

148.

90

Warren, L.
1959. The thiobarbituric acid assay of sialic
acids. J. Biol. Chem. 234:1971.

Wathen, J. D., and Levy, R. S.
1966. Composition of the protein from dog lymph
chylomicrons and dog serum high density
lipOprotein. Biochemistry, 5:1099.

Weiss, L.
1963. The mammalian tissue cell surface. In
The Structure and Function of the Membrane
and Surfaces of Cells, Biochemical Society
Symposium No 22‘(D. J. Bell and J. K.
Grant, eds.), Cambridge Univ. Press,
Cambridge, England.

 

White R. F., Eaton, H. D., and Patton, S.

1954. Relationship between fat globule surface
area and carotenoid and vitamin A content of
milk in successive portions of a milking.

J. Dairy Sci., 31:147.

Whitney, R. McL.
1958. The minor proteins of bovine milk.
J. Dairy Sci., 41:1303.

Wolf, D. P., and Dugan, L. R. Jr.
1964. Structure of high melting glycerides from
the milk fat-globule membrane. J. Am.
Oil Chemists' Soc., 41:139.

Wood, H. G., Gillespie, R., Seymour, J., Hansen,
R. G., and Hardenbrook, H. 14
1958. Lactose synthesis. V. C in lactose, glycerol,
and serine as indicators of the triose
phosphate isomerase reaction and pentose
cycle. J. Biol. Chem. 233:1271.

Zittle, C. A., DellaMonica, E. S., Custer, J. H.,
and Rudd, R. K.
1956. The fat globule membrane of milk: Alkaline
phosphatase and xanthine oxidase in skim-
milk and cream. J. Dairy Sci., 32:528.

‘K’m

 

APPENDIX

ABSTRACTS OF PUBLICATIONS

91

92
IDENTIFICATION OF RIBONUCLEIC ACID IN
THE FAT GLOBULE MEMBRANE
F. C. Swope and J. R. Brunner
Department of Food Science

Michigan State University
East Lansing, Michigan 48823

SUMMARY

Isolation and identification of a nucelic acid com-
ponent, presumably of ribosomal origin, was obtained from
fat globule membrane preparations. Membrane material was
prepared by churning thrice-washed, raw cream separated
from uncooled milk. Nucleic acid-rich fractions, obtained
in yields of 5-10 mg per 40 liters of milk, were isolated
from washed cream buttermilk using a sodium dodecyl
sulfate-liquid phenol extraction procedure. Specific
identification of nucleic acid was achieved by assay with
crystalline pancreatic RNase. The TCA-soluble nucleo-
tides were followed spectrophotometrically at 260 mu.

It is suggested that the fat globule membrane con-

tains small amounts of microsomal material.

J. Dairy Sci., 48:1705 (1965).

 

W

93

'EZT

RIBOFLAVIN AND ITS NATURAL DERIVATIVES IN
THE FAT GLOBULE MEMBRANE

F. C. Swope, J. R. Brunner, and D. V. Vadehra
Department of Food Science
Michigan State University
East Lansing, Michigan 48823

SUMMARY

A greenish-yellow color was observed in aqueous
suspensions<fi‘fat globule membrane preparations. Liber-
ation of this material from the membrane particulate is
enhanced by addition of acid to pH 4.0 or the anionic
dissociating agent sodium dodecyl sulfate. Freezing or
lyophilization, or both, also dissociates this substance
from the membrane.

Fluorometric assays made on the membrane prepar—
ations showed that the substance released possesses
fluorescent characteristics similar to that observed in
whey which, presumably, emanates from riboflavin. Of
the total riboflavin found in the membrane, 92 to 95%
exists as FAD, probably as the co-enzyme of xanthine
oxidase. This compound is easily degraded by acid,

dissociating agents and freeze-drying to riboflavin.

J. Dairy Sci., 48:1707 (1965).

 

94

 

APPARENT HOMOGENEITY OF LACTOPEROXIDASE
IN GEL ELECTROPHORETOGRAMS

F. C. Swope, C. W. Kolar, Jr., and J. R. Brunner
Department of Food Science
Michigan State University
East Lansing, Michigan 48823

SUMMARY

Lactoperoxidase (doner: H2O2 oxido-reductase, EC
1.11.1.7), a heme-containing protein which catalyzes the
transfer of oxygen from peroxides to other substances, is
one of the principal enzymes in cow's milk. Most of the
available information on lactoperoxidase indicates that
the enzyme exists in nature as isozymes or that process-
induced alterations in the native protein might possibly
account for the observation of other forms upon purifi-
cation.

A lactoperoxidase-rich whey fraction was isolated
from freshly drawn, uncooled cow's milk and submitted to
zonal electrophoresis in a formic acid buffer, pH 3.8.

The enzyme band was developed with a solution of guaiacol.
Results of these studies suggest that lactOperoxidase
exists as a single molecular species when isolated rapidly
and that other forms.may be caused by manipulations in

isolation procedures.

J. Dairy Sci., 49:1279 (1966).

95

THE FAT GLOBULE MEMBRANE OF COW'S MILK:
A REASSESSMENT OF ISOLATION PROCEDURES
AND MINERAL COMPOSITION

F. C. Sw0pe and J. R. Brunner
Department of Food Science
Michigan State University

East Lansing, Michigan 48823

SUMMARY

Cream washed three times with three volumes of
deionized water yielded fat globule membrane preparations
essentially free of plasma components. Loss of loosely
bound membrane material resulting from the washing process
amounted to approximately 4% per washing. The loss was
minimized when an isotonic sucrose-saline solution was
employed as the washing medium. Membrane preparations
obtained from three separate lots of Jersey, Brown Swiss
milk by carefully controlled procedures amounted to 1.4
to 1.6 g/lOO gr of fat. Whereas about 70% of the membrane
was recovered from the buttermilk (aqueous membrane
fraction) the remaining portion was closely associated
with the butter granules (serum membrane fraction). The
serum membrane contained considerably more lipids than the
aqueous membrane fraction.

Molybdenum, iron and copper were the principal metal-
lic components detected in membrane preparations. It is
suggested that the Mo content is attributable exclusively
to the presence of xanthine oxidase.

IN REVIEW

 

96

ISOLATION AND PARTIAL CHARACTERIZATION OF A
GLYCOPROTEIN FROM THE FAT GLOBULE MEMBRANE

F. C. Swope, K. C. Rhee, and J. R. Brunner
Department of Food Science
Michigan State University
East Lansing, Michigan 48823

SUMMARY

Fragmented fat globule membranes from thrice-
washed cream were separated into 7,5008, 2308, and 358
pellets by ultracentrifugation. After lipid extraction
the pellets were partially resolubilized with 1M KCl. A
high carbohydrate-containing protein was found to reside
primarily in the 35S fraction which accounted for approxi-
mately 9% of the pellet. The hexose of the protein was
7.8% and the total carbohydrate content was estimated to
range between 15-l9%. Some physical parameters of this
glycoprotein were S20,w (11.6), MW (306,500), D20,w (2.0)
and u (-3.6). The amino acid content showed relatively
high amounts of lysine, threonine and glutamic acid as
compared to the insoluble pellet proteins.

It is theorized that this protein may have a func-

tional role in the creaming phenomenon of milk.

IN REVIEW

 

 

11711111171111“

“mm

3 03174 82

NI!!!
U"
E”
III]

3 129

«WWW