DAIRY FOODS
On the Stability of Casein Micelles1
ABSTRACT
A view of the structure of the casein
micelle is given. It is built of submicelles,
roughly spherical aggregates of several
casein molecules held together by
hydrophobic bonds and salt bridges.
Regions of amorphous calcium phosphate
link the submicelles to each other; the
ester phosphate groups form part of this
colloidal phosphate. In this way, almost
all regions of the casein molecules are
severely restricted in mobility. The C-
terminal part of the lC-casein is however
predominantly present as flexible "hairs"
located at the outside of the micelles.
There are essentially two types of sub-
micelles, with and without (much) lC-
casein. The casein micelles greatly
change in properties upon lowering the
pH, mostly due to dissolution of colloidal
phosphate; at still lower pH, increased
fonnation of salt bridges predominates.
Temperature also has pronounced effects:
upon lowering it, the micelles become
more voluminous, presumably due to pro-
truding hairs of (mainly) ~-casein. Also
at high temperature (>70"C), parts of the
casein molecules become more flexible.
Casein micelles are very stable. If con-
ditions are changed they may disintegrate
or aggregate. Aggregation mostly leads to
fonnation of a gel. The application of the
theory of fractal floc formation to gela-
tion is briefly discussed: it serves to ex-
plain the very strong dependence of gela-
tion times on volume fraction of
aggregating particles.
The stability against aggregation is
primarily due to sterle repulsion, caused
by the hairs of lC-casein, and at low tem-
Received September 30,1989.
Accepted ll111lWY 29. 1990.
lMiles-MarschalI Award Invited Lecture.
1990 1 Dairy Sci 73:1965-1979
PIETER WALSTRA
Department of Food Science
Wageningen Agricultural University
Wageningen, The Netherlands
perature presumably also ~-casein. The
hairs on different casein micelles may
however touch, and this may lead to last-
ing contact of the micelles, Le., aggrega-
tion. The bonds formed can be salt
bridges or, at high temperature, covalent
bonds (chemical crosslinks). Hydropho-
bic bonds are probably not involved. The
probability that casein molecules in dif-
ferent micelles may touch each other for
a sufficient time for bonds to be formed
appears to depend on electrostatic as well
as steric repulsion, which thereby affect
aggregation rate.
(Key words: casein micelle structure, sta-
bility, heat coagulation)
INTRODUCTION
In 1981, the late Theo Payens was the first
recipient of the Miles-Marschall Award, and on
that occasion he reviewed the subject of stable
and unstable casein micelles (44). Since then,
new work has been published and ideas have
matured. Some of these developments and their
technological significance will be briefly re-
viewed in this article.
Before the discussion on stability, the struc-
ture and some properties of the casein micelles
will be considered, but detailed information is
elsewhere (35, 52). Micelles are roughly spheri-
cal aggregates, mostly between 40 and 300 om
in diameter, and they are fairly voluminous,
containing l1!uch water or, more precisely, a
solution similar to milk serum. They contain
essentially four kinds of casein molecules, at a
molar ratio of about €Xsl:€Xs2:(~ + y):lC = 4:1:4:
1.3, and about 7% of the DM of micelles
consists of inorganic material, predominantly
calcium and phosphate. The micelles show con-
siderable variation in composition, structure,
and size distribution; milk serum also varies,
especially in salt composition. The effects of
variation will not be systematically considered
here.
1965
1966
Figure 1. Section through a casein micelle; highly schematic. Prom (79), by courtesy of Wiley.
A MODEL OF THE CASEIN MICELLE
Figure 1 illustrates the structure of the
casein micelle (79). The model has evolved
over the years, and several authors have contri-
buted to it (37, 56, 58), especially Schmidt (52)
and Schmidt and Payens (54).
The reality of submicelles is now fairly
clear. All electron microscopic methods show
them, and the not-quite-spherical shape of the
micelles fits well with the assembly from dis-
crete units. Such an assembly is also strongly
suggested by electron micrographs of lactating
cells (3, 17), which show casein micelles in
various stages of development. Neutron diffrac-
tion (60) and X-ray diffraction (46) also show
the existence of discrete subunits in the micelle.
Full agreement has not been reached about their
size, but the diameter is probably between 10
and 15 nm, and they probably contain between
15 and 25 casein molecules. The bonds be-
tween the molecules in a submicelle are both
hydrophobic and electrostatic (salt bridges)
(52).
The submicelles are not all the same. Essen-
tially, there are two major types, with and
Journal of Daily Science Vol. 73, No.8, 1990
WALSlRA
o submicelle
protruding
'r chain
calcium
phosphate
100 nm
without K-<:asein (40, 41, 79). This is not sur-
prising, since K-<:asein exists in milk as an
oligomer, on average consisting of six
molecules (61). K-Casein is predominantly at
the outside of the micelles, as follows for in-
stance from the electron microscopy of micelles
with labeled K-<:asein (53) and from the propor-
tionality between K-<:asein content and specific
surface area of casein micelles (12, 15). The
very hydrophilic C-terminal part of most K-
casein molecules is sticking out from the
micelle core into the solvent as flexible ''hairs.''
Evidence for this comes from hydrodynamic
studies (24, 28, 29, 77) and from proton nuclear
magnetic resonance (NMR) in ~O (20, 49),
which shows that part of the K-<:asein has con-
siderable freedom of motion. This concerns the
C-terminal end, starting at a point between
residues 86 and 96 (49); the phenylalanine-
methionine (Phe-Met) bond cleaved by chymo-
sin is at residues 105-106. 1bese hairs are
essential in providing stability against floccula-
tion of the micelles. The hydrodynamic thick-
ness of the hairy layer is about 7 nm.
The submicelles are linked together by col-
loidal calcium phosphate (CCP) (52). Much
 MILES-MARSCHALL AWARD INVITED LECfURE
TABLE 1. Approximate relaxation times l for changes occurring with casein micelles when changing the temperature or
some other change at constant temperature.
Phenomenon observed
Dissolution of ~casein
Exchange of ~ein
Exchange of Ca
Ca goes to micelles
Specific viscosity
Gel permeability'2
Firming of rennet curd2
Dynamic shear modulus2
Dynamic shear modulus2
lTime needed for a change to proceed to lie of its final value.
20f renneted skim milk.
debate has focused on the composition and
lattice structure of the CCP. It has no clear
crystalline structure - the regions of CCP are
too small for that - but Holt (23) and Holt et al.
(25) have shown that CCP probably has a
structure somewhat resembling brushite. This
fact may seem surprising, since brushite has a
Ca:P ratio of 1, whereas in the micelles the
ratio of calcium to inorganic phosphate (Ca:PiJ
is much higher, but is explained by the ester
phosphate groups of the casein being part of the
CCP (23). This fits with the observation that
the very small regions of CCP nevertheless
show no Ostwald ripening: these regions are
thus stabilized by bonds with the casein. The
CCP does not necessarily represent a state of
thermodynamic equilibrium: that state would
probably be precipitated hydroxyapatite, sepa-
rate from the micelles. Additional CCP formed
in the casein micelles by heat or concentration
also has a Ca:P ratio of about 1 (39), but this
does not necessarily have the same properties
as the "natural" CCP.
The CCP does more than link the sub-
micelles together. Proton NMR in D:zO on
loose submicelles shows that considerable por-
tions of the casein molecules have great free-
dom of motion; when CCP is introduced, this
flexibility is almost fully lost (49). The casein
molecules, despite their relative scarcity of sec-
ondary structure, have a fairly rigid configura-
tion in the casein micelles at physiological con-
ditions with only part of the K-casein molecules
being flexible.
1967
Temperature Relax. time Reference
CC) (h)
37~5 .5 (6)
4 o=(j (6)
'" 18 "'1.5 (81)
78~20 18 (79)
3~20 .5 - 1 (68)
3~20 .6 (68)
30 2 (34, 82)
3~20 1.5 (83)
2~30 .1 (83)
The latter observation leads to the question:
how dynamic are casein micelles? Exchange of
radiolabeled casein (6) and Ca (80, 81) has
been observed. However, most dynamic
equilibria between substances in the serum and
the micelles are far to the micelle side. Several
kinds of small changes in the serum cause
changes in the micelles, but these usually take
some time. Table 1 summarizes some of the
relaxation times observed, which are very long
compared with those of most changes occurring
at the molecular level (say, nanoseconds). Even
after acid is added to milk. several minutes may
be required before the pH becomes more or less
constant (Geurts, unpublished). Presumably,
most changes in equilibrium are more compli-
cated, more than one relaxation phenomenon
occurring. Moreover, the dissociation of a pro-
tein molecule requires that several bonds are
broken simultaneously.
Casein micelles change considerably when
pH is lowered (Figure 2). The CCP goes into
solution, and although micelle-like particles re-
main, they have very different properties (48).
Figure 2 suggests that the loss of CCP is pri-
marily responsible for the changes observed;
this is especially clear for the electrokinetic
potential. The variation in voluminosity in the
pH range 6 to 6.6 is not quite certain; moreo-
ver, some slight change in the size distribution
of the micelles may occur when lowering the
pH (76). The change in voluminosity parallels
that of the spin-spin relaxation time (f2) of the
water protons in the solution; in the case of a
caseinate solution, T2 is similar at low pH, but
Journal of Dairy Science Vol. 73, No.8, 1990
1968 WALSlRA

CASEI N MICELLES PARACASEINATE GEL

40 .....-----,-----,r----::;;;--.
% in micelles proteolysis of
20 u s,- casein

0 f----r-+--.---.---....-+O
- ~ (mV) modulus
500
(Po)
10
---..L.250

0 tan 5
g water/
g casein 2 .4

- .2
o -+--r-I---,.--,--~ f----r---r--.---,----rl- a
400 rate of
Tz (m s) syneresIs
200

O-+--.,.---'---r--,---r-'
4.5 5.5 6.5 4.5 5.5 6.5
pH pH
Figure 2. Properties of casein micelles and paracaseinate gels as a function of pH: percentage of calcium and inorganic
phosphorus in the micelles (69); electrokinetic or zeta potential (55); vOluminosity, expressed as amount of water per
amount of casein in a pellet (69); spin-spin relaxation time of water protons (47); rate of proteolysis by rennet of CXs1-
casein arbitrary scale; dynamic shear storage modulus at a frequency of I s-1 (84); loss tangent at a frequency of .01 s-1
(84); and rate of syneresis [mainly from (47 and 67)].

it keeps increasing when the pH is raised. The
very sharp transitions near pH 5.2 are also
manifest in some properties of rennet skim milk
gels (Figure 2). The loss tangent of the gel is a
measure of the extent to which viscous relative
to elastic behavior is present. (Note that the
peak in viscous-like behavior near pH 5.2 cor-
responds to the optimum for "meltability" or
"stretchability" of curd) We may tentatively
conclude that the bonds keeping the casein
micelles together are weakest or fewest at pH
5.2 or 5.3. At lower pH, increasing electrostatic
attraction between casein molecules keeps the
"micelles" more tightly together; at higher pH
an increasing quantity of CCP does the same
(74).
Some changes with temperature are depicted
in Figure 3. The increase in the electrokinetic
potential with temperature seems to conflict
with the increase in bound Ca, since more
bound Ca should correspond to a smaller nega-
tive charge, hence, a smaller potential. The
voluminosity of the micelles markedly in-
creases at low temperature, although the data
below 20·C are not quite certain because of the
partial dissolution of l3-casein. Because 13-
casein is primarily held in the submicelles by
hydrophobic bonds, its dissolution at low tem-
perature, where hydrophobic bonds are weak, is
no surprise (52). In addition to some l3-casein
molecules going into solution, others may only

Journal of Dairy Science Vol. 73, No.8, 1990

 MILES-MARSCHALL AWARD INVITED LECTIJRE 1969
TYPES OF INSTABILITY
20 Casein micelles can be disintegrated by
- ~ (mV) removing (dissolving) CCP, thereby leaving
10 free submicelles, or by adding agents that break.
hydrogen bonds or hydrophobic bonds, thereby
o+--,.-.-,--..,----r--rl disintegrating the submicelles (62). This aspect
"soluble" 40 ~ will not be considered herein.
(3 -casein Casein micelles, or rather the particles der-
ived from them by changing their environment,
1%) 20 can aggregate. They are generally much more
prone to aggregation than is dissolved casein
o "- under comparable conditions. This is presum-
vot~~~~~sitY 4 '_'~~_~ ably due to the far greater loss of conforma-
tional entropy on aggregation of casein
2 molecules, which compensates for the decrease
in enthalpy when the bonds causing aggrega-
2-
o+-----.---,---r------.----,-j tion are formed.
Ca bound Some types of aggregation can be distin-
6 guished. 1) Simple flocculation, as occurs with
(mol/mol)
4 hard lyophobic particles; undisturbed, eventu-
2 ally leads to formation of a gel (2). 2) Floccu-
lated micelles may fuse into bigger micelles of
0 roughly spherical shape. If this type
0 20 40
predominates, visible particles appear and pos-
temperature (OC) sibly a sediment. 3) Micelles may aggregate by
Figure 3. Properties of casein micelles at physiological
means of adsorbing (macromolecular) material
pH as a function of temperature. Electrokinetic or zeta connecting them, also leading to a gel.
potential (various sources). Percentage of lkasein found in The first is the most common type of aggre-
the supernatant after high-speed centrifugation (6). Volumi- gation, although casein micelles are certainly
nosity of the casein micelles, calculated from specific
not hard lyophobic particles. Fusion, as noted
viscosity data (79). Amount of Ca bound to lXsl-casein
(13). under the second type, may occur after rennet-
ing, albeit slowly, and at high temperature if
the pH is not too low. In these situations, the
micelles are more or less hairless.
Table 2 lists various causes for coagulation.
Cause 3, 5, and 6 will be considered further on
be loosened, thereby constituting another cate- in some detail; 1 will not be discussed, since its
gory of flexible hairs at low temperature. If the explanation is unclear (21); the gelation may be
increase in voluminosity (Figure 3) resulted comparable with type 3. Cause 2, i.e., the irre-
only from this effect, the hairy layer would versible aggregation of casein adsorbed onto air
become quite thick. Presumably, the core of the bubbles, even after the air itself has dissolved,
micelle would also swell to some extent and a has received little attention, but is quite clear
limited disintegration of micelles into smaller (38). Cause 4 is rather trivial: the casein be-
ones might occur, which would also cause the comes insoluble near its isoelectric pH. At low
voluminosity to increase (77). temperature, where it still is soluble, the
At high temperatures, starting above about molecules form micelle-like particles that ag-
70·C, another change occurs. Proton NMR in gregate on increasing the temperature. Cause 7
D20 shows that considerable parts of the casein is comparable to salting out, but not quite: a
molecules become much more flexible (49), as high concentration of CCP leads to a kind of
if part of the submicelle structure melts. This growing or fusion of micelles. Apparently, the
may be important in the changes that occur stabilizing factors are overcome if much CCP
upon heating. can be deposited. In the rare ''Utrecht milk

Journal of Dairy Science Vol. 73, No.8, 1990

1970 WALSTRA

TABLE 2. Various causes for the aggregation of casein micelles.

Micelles Aggregation Aggregation
Cause changed? reversible? at low temperatures?
1. Long time
(age gelation) Presumably No No
2. At air-water interface Spreading No No
3. High temperature
(heat coagulation) Chemically No
4. Acid to pH = 4.6 No CCP left Yes No
5. Ethanol Presumably No ?
6. Renneting lC-casein split No No
7. Excess ea2+ etc. More CCP Yes Yes
8. Freezing ? Partly
9. Some polymers No Mostly Yes

abnonnality," Le., milk with a low citrate con-
tent and a relatively high ca2+ activity, this
seems to occur spontaneously (51). Cause 8 is
comparable, but here there also is true salting
out because of the very high ionic strength
resulting from freezing most of the water. An
irreversible change appears to occur in the
CCP, causing the aggregated micelles to be not
fully redispersible after thawing (5). The mate-
rial causing aggregation meant in Cause 9 may
be various polysaccharides (59, 63).
It may be noted from Table 2 that in nearly
all cases, the aggregating particles are no longer
the native casein micelles. The latter are pre-
sumably perfectly stable, and they have to un-
dergo some essential change before aggregation
can occur. This fits with the irreversibility of
the aggregation mostly observed. Also, even
the altered micelles are in many cases stable at
very low temperature, but this may be a kinetic
and not necessarily a thennodynamic stability.
Once the "micelles" are aggregated, they do not
disperse again on lowering the temperature.
Renneted micelles do aggregate at 5·C at pH
4.6 {47).
KINETIC ASPECTS
For particles to aggregate, they must en-
counter each other. The encounter frequency
can in principle be calculated from
Smoluchowski's theory (42). We will first con-
sider aggregation type 2, where aggregating
particles fuse, assuming the fusion to be rapid
compared with the time between encounters.
The coagulation time then is the time needed
for visible particles to appear, say .2 mm large.
For diffusion controlled encounters it can be
derived that
teoag = (1ta31'\/8kT) (q3/cl» W [1]
where k = Boltzmann's constant, T = absolute
temperature, a = original particle radius, 11 =
viscosity, qa = radius of the particle when
visible (or q3 the number of original particles
making up a visible particle), cl> = the volume
fraction of the particles, and W = the stability
factor, i.e. the ratio of the number of encounters
to the number leading to lasting contact.
For simple flocculation, it is now well estab-
lished that the flocs fonned are of a fractal
nature (16, 28, 36), i.e., their structure is scale
invariant at scales larger than a. Considering
the radius R of a growing floc containing Np
particles, we have:
[2]
where D is the fractal dimensionality, which is
always <3. Equation [2] is found to hold
remarkably well in a wide variety of situations
and over a very wide range of floc sizes. The
magnitude of D depends on some conditions,
such as the colloidal interactions between the
particles (16). The number of particles a floc of
the same radius could contain if they were
close packed N a = (RIa)3. Consequently, the
volume fraction of particles in the floc is given
by:
lPnoc = NplNa = (RIa)D-3 [3]
Thus, lPnoc becomes ever smaller as the floc

Journal of Dairy Science Vol. 73. No.8, 1990

 MILES-MARSCHALL AWAIm INVITED LECfURE
coagulation time/W (s)
4 micelles or fragments thereof to become at-
10
1 stability, especially of cream (33).
"fractal flocs"
0.01 .03 .1 .3 0.63
volume fraction
Figure 4. Coagulation times for casein micelles (aver-
age radius 60 nm) at room temperature as a function of the
volume fraction of micelles, calculated according to Equa-
= =
tions [1] with q 200 and [5] with D 2.3, assuming the
stability factor W to be unity.
grows in size, and as soon as it becomes equal
to the total volume fraction of particles in the
liquid cjl, the flocs fill the total volume and a gel
fonns (2). At that moment the flocs have a
radius given by:
Rait = acjll/(D-3) [4]
Applying again Smoluchowski's theory, the ge-
lation time turns out to be:
[5]
For various types of casein gels (2) and for heat
coagulation (Nieuwenhuijse et aL, to be pub-
lished), D =
2.3 ± .1.
Results calculated according to Equations
[1] and [5] are in Figure 4. The theory outlined
here is simplified, not taking into account poly-
dispersity (which may cause somewhat quicker
aggregation), any effects of velocity gradients
(quicker), or increased viscous resistance when
particles approach closely (slower). Neverthe-
less, Figure 4 serves to illustrate important
points, such as the enormous effect of cjl on
gelation time. For example, homogenization of
milk and especially cream may greatly speed
1971
up any aggregation process of the casein
micelles. Homogenization causes casein
tached to the fat globules, which now react as if
they were casein micelles (38). Thus, homogen-
ization greatly increases the effective volume
fraction of casein, the more so if homogeniza-
tion clusters have fonned (38). This, in combi-
nation with Figure 4, serves to explain the great
detrimental effect of homogenization on heat
What is the magnitude of W or what magni-
tude is needed to provide stability? Figure 4
shows that milk at room temperature (cjl == .09)
would have a tgel of about lOs if W = 1.
Because the milk may be stable (if sterile and
without proteolytic enzymes) for say 3 yr, this
implies that W is at least about 107 . With
concentrated (evaporated) milk at 120·C, W = 1
would lead to a coagulation time of .01 s,
taking into account that the encounter fre-
quencyat l20·C is about 10 times that at 20·C.
Since the coagulation time mostly is at least
1Q3 s, W must be about lOS or more under
these conditions.
The essential question is whether the stabil-
ity factor can be predicted: W can be >1 be-
cause only part of the surface of the particle is
reactive, i.e., lasting contact between two parti-
cles depends on their orientation when they
meet. An example may be casein micelles dur-
ing renneting. The other cause for slowing
down aggregation is that the particles do not
always come close enough to make lasting
contact, because of a net repulsion between
them caused by colloidal interaction forces.
COLLOIDAL INTERACTION
In the classical Deryagin Landau Verwey
Overbeek (DLVO) theory (22) electrostatic
repulsion and Van der Waals attraction are
taken into account to calculate the free energy
needed to bring two particles from infinite to
some close distance from each other. If this
interaction free energy is negative at all dis-
tances, the particles can come to touch each
other, ie., flocculate. If the interaction free
energy is positive at some (small) distance, its
maximum value may roughly be taken as an
activation free energy for aggregation, thus per-
mitting the calculation of W. The DLVO theory
has often been applied to casein micelles, but
Journal of Dairy Science Vol. 73, No.8, 1990
1972 WALSTRA

cannot explain their stability (43): the maxi- (QJ

mum is too low. Moreover, the presuppositions
of the theory are not fully met; the particles are
not perfect homogeneous spheres, and so-called
hydration repulsion, which is manifest up to
distances from the particle surface of about I
om (32), is not taken into account. Neverthe-
less, some variables that affect micelle stability
seem to correlate at least qualitatively with
predictions from the DLVO theory. Some im-
portant variables in this theory are:
1. For large particles, the maximum interac-
tion free energy is higher, and the particles
would thus be more stable: This does not fit
with the observations (10), but such a size (bl
dependence has rarely been found at all. segment density probability
2. A higher surface potential, hence, a
higher charge of the particles, causes better
stability: This may well fit, and the negative
t 1
potential of the micelles is particularly
decreased b~ decreasing the pH and by increas-
ing the Ca + activity (71).
3. A higher ionic strength causes less stabil-
ity, because the electrostatic repulsion works
over a shorter distance: This does not always
fit, but addition of salt also dissolves some o+-_ _ -=:-_~.L.----====-- _ __'_ 0
micellar Ca and somewhat increases micelle o .------ distance from surface
voluminosity (see variable 5).
4. A lower dielectric constant gives less sta-
Figure 5. Hypothetical picture of interactions between
bility, because it decreases electrostatic repul- two casein micelles, (a) illustrating the configuration of
sion: This would fit with the destabilizing ac- hairs and charges (oversimplified). In (b), the average
tion of alcohols. segment density of the hairs of one micelle as a function of
5. A higher Hamaker constant would cause the distance from the core of that micelle is illustrated, as
well as the relative probability of finding a segment of a
lower stability. The Hamaker constant is a ma- hair of the other micelle at the same place, for a high and a
terial property and is a measure of the intensity low negative charge of the micelles.
of the attractive London-van der Waals forces
acting between the particles. In the DLVO the-
ory, it is the difference in properties between
the particles and the continuous phase (solvent) freedom of motion of the flexible hairs on the
that matters. If casein micelles have a higher surface of a particle, this always causes repul-
voluminosity, i.e., contain more solvent, the sion. This is called the volume restriction term,
difference between them and the solvent be- and its magnitude is proportional to the hair
comes less, hence the Hamaker constant lower. density. If the hairy layers of two particles
This appears to fit with the general observation overlap (interpenetrate), the solvent quality of
that a higher voluminosity of the micelles goes
the hairs determines what will happen: if it is
along with a higher stability (79).
good, the so-called mixing term is repulsive
All the same, an additional mechanism for
(osmotic repulsion); if it is poor, the term may
providing colloidal stability must be present,
be attractive. Often, the mixing term is predom-
and it is now fairly clear that sterlc repulsion
inant; its magnitude is proportional to the hair
caused by the micellar hairs is responsible (14,
24, 26, 27, 71, 73, 77, 78). In sterlc repulsion, density squared. Theorles for the calculation of
two mechanisms can be distinguished (75). If sterle repulsion are now available (50, 75), but
the presence of a second particle restricts the too many uncertainties remain to apply them to

Journal of Dairy Science Vol. 73, No.8, 1990

 MILES-MARSCHALL AWARD INVITED LECfURE
the casein micelles. Nevertheless, the repulsion
must be very considerable, because of the
hydrophillic nature of the macropeptide part of
K-casein. This is borne out by the observation
that casein micelles in a pellet obtained by high
speed centrifugation, where they are so closely
packed as to materially deform one another, can
nevertheless be resuspended How then can
casein micelles be made to aggregate in some
conditions, and how can the effects of the
variables listed above be explained?
An attempt at an explanation is illustrated in
Figure 5. Figure Sa is a greatly oversimplified
picture of the micellar surface. The hairs them-
selves are charged, in addition to the charge on
the core surface, and this has two effects. First,
the electrostatic repulsion extends over a dis-
tance much farther from the core surface than
would be the case if the hairs were uncharged;
this is because the thickness of the electrical
double layer (the distance over which an elec-
trostatic potential decreases - due to shielding
by counterions in the solution - to lie of its
value at the surface) is in milk only about 1.1
nm (79), whereas the hairy layer is much thick-
er. A first attempt at deriving the electric poten-
tial of the micelles as a function of distance
from the core has been made (24) and appears
to fit with this idea. Second, the charge affects
the conformation of the hairs themselves. At
high pH the charges on the hairs are predomi-
nantly negative, and this causes them to
"stretch" somewhat, thus extending farther
from the particle surface. At lower pH, fewer
negative and more positive charges are on the
hairs, which would cause them to more or less
"curl up". At the isoelectric pH, no hairy layer
is probably left, except at low temperature.
Figure 5b shows a presumption about the
probability that a segment of a hair at one
micelle is found within the hairy layer of an-
other micelle. This probability must be deter-
mined by the steric repulsion exerted by the
hairy layers themselves, but also by the elec-
trostatic repulsion. At present, these probabili-
ties cannot be calculated, but in a qualitative
sense the picture may be correct. TIle impor-
tance is that the hairs, which show continuous
Brownian motion, may touch one another. If
that occurs at positions of reactive side groups,
crosslinking may occur, thereby linking the
micelles together. The stability factor W may
now be interpreted as the inverse of the proba-
1973
bility of touching of reactive sites of hairs from
two micelles encountering each other, multi-
plied by the inverse of the rate constant of the
crosslinking reaction. It appears very unlikely
that the cores of the micelles can come to
touch, unless the hairy layer is somehow re-
moved.
Figure 4 reveals that at low temperature the
voluminosity of the micelles is higher, and this
is presumably due to a (further) extension of
hairs of J3-easein (77), thereby providing addi-
tional steric repulsion. The higher volurninosity
may also go along with a lower Hamaker con-
stant, as mentioned. The absolute value of the
electrokinetic potential is lower at a lower tem-
perature, which fits neither with the increased
stability (although variation in the potential is
of questionable importance in these conditions)
nor with the decreased adsorption of Ca ions,
which would have an opposite effect on the
potential. The explanation probably lies in the
fact that a thicker hairy layer causes the slip-
ping plane, at which the potential is sensed, to
be farther away from the micelle core, thus at a
distance where the potential has further
decayed.
The considerable effect of temperature on
the aggregation rate of micelles under various
conditions (Table 2) may not be taken as an
indication that hydrophobic bonds are the cause
for keeping the micelles together after they
have touched. Undoubtedly, any stronger steric
repulsion at low temperature due to hairs of 13-
casein is ultimately caused by the hydrophobic
bonds responsible for keeping lJ-casein in the
micelles to be much weaker at a lower tempera-
ture. But this causes only a higher activation
free energy for aggregation, not a lower bond
energy between micelles. This is borne out by
the observation that the shear moduli of acid
and rennet milk gels considerably increase after
lowering the temperature (47,83). If hydropho-
bic bonds were to keep the micelles aggregated,
the modulus should become lower or the gel
should even dissolve again on lowering the
temperature.
What then may be the kind of bonds causing
lasting contact between closely approached
micelles? The first type of bond is salt bridges,
either between negatively and positively
charged groups on either peptide chain, or me-
diated by Ca ions, or even by a CCP-lik.e
linkage. Presumably, several bonds must be
Journal of Dairy Science Vol. 73, No.8, 1990
1974
fonned between two micelles for the contact to
be lasting, considering the short lifetime of
many ion pair bonds. The tendency to fonn salt
bridges presumably increases with increasing
supersaturation of milk salts (predominantly
phosphates), thus with increasing concentration
and increasing pH, and for the same pH at
increasing Ca2+ activity and temperature. A
parallel may be seen with the effect of pH in
the range 5.3 to 6.7 on micelle properties
(Figure 2). A second type of bond is the cova~
lent linkages between groups on the peptide
chains, of which various types can be possibly
fonned (79), but only at high temperatures at a
reasonable rate. Presumably, one bond would
suffice to ensure lasting contact.
After the micelles have been linked, further
bonds, also of another nature, can be fonned,
the micelles more or less fusing. Complete or
partial fusion has been observed after renneting
(19), at high temperature at high pH (7), and at
the air/water interface at room temperature
(38).
ETHANOL STABILITY
This aspect has especially been studied by
Horne et al. (26, 27, 29, 30, 31). Their work.
confirms that steric stabilization by a hairy
layer is essential. But electrostatic effects are
also present. Ethanol stability decreases with
decreasing pH, hence decreasing charge; Ca2+
activity and ionic strength also play a part.
Addition of ethanol lowers the dielectric con-
stant, and according to the DLVO theory mar-
kedly reduces electrostatic repulsion. Ethanol
also lowers the solvent quality for the macro-
peptide part of K-easein. This factor and the
decreased electrostatic repulsion between nega-
tive groups on one hair may cause the hairs to
curl up and finally collapse. Correlation was
good between the hydrodynamic thickness of
the hairy layer and the ethanol concentration
needed to cause aggregation.
RENNETING
Renneting involves two reactions, an enzy-
matic one in which chymosin (BC 3.4.23.4) or
another proteolytic enzyme splits a maeropep-
tide off K-easein, followed by aggregation of
the now fonned paracasein micelles if tempera-
ture and Ca2+ activity are high enough. Payens
Journal of Dairy Science Vol. 73, No.8, 1990
WALSTRA
In W
15
10
300
1 [
5 .39
149
.60
O-L.-----.-----r-------l
80 90 100
split x-casein (%)
Figure 6. The relative aggregation rate of partly ren-
neted casein micelles, expressed as the logarithm of the
stability factor W, as a function of the degree of conversion
of the K-easein. Results (73) derived from the viscosity
change in undiluted skim milk (o) and (9) derived from tOO
turbidity change of a diluted micellar suspension (e).
Temperature of experiment indicated.
(45) pioneered the quantitative study of the
kinetics of both reactions combined Others
have contributed substantially (8, 14, 73). Dal-
gleish has especially studied the stability of
paracasein micelles (9, 11).
A look at Figure 1 makes clear that chymo-
sin thus acts as a "depilatory" (77, 78). Because
the position where the enzyme splits the hair is
near the micelle core, to realize the splitting by
means of immobilized enzyme is virtually im-
possible, which agrees with the conclusion
reached by careful consideration of the experi-
mental evidence (4). Because Brownian motion
of the casein micelles is negligible compared
with that of the enzyme molecules, it is to be
expected that the enzymatic reaction is first
order, as is indeed observed (72). This implies
and also agrees with other evidence, that the
enzyme randomly attacks the K-casein: the en-
zyme molecule moves by Brownian motion
through the senun and through the hairy layer
around a micelle encountered, until it happens
to reach the vulnerable site at the K-casein
chain, which it will attack if it is in the correct
orientation, subsequently to diffuse away. This
 MILES-MARSCHALL AWARD INVITED LECI'URE
is in agreement with the observation that
negligible adsorption of chymosin on
paracasein occurs, at least at physiological pH;
at lower pH, the enzyme becomes somewhat
adsorbed (73; Geurts, unpublished data). The
rate of the enzyme reaction and the effect of
some variables on it could be fairly precisely
calculated from first principles (73).
The stability factor of aggregating micelles
for various degree of lC-casein conversion is
illustrated in Figure 6. Measurable aggregation
only occurs after most hairs have been re-
moved, in agreement with the concept of sterlc
repulsion. There may be two alternative expla-
nations for the linear relation between log W
and degree of conversion. One is that a few
remaining hairs cause a weak. repulsion. At a
low hair density, the volume restriction term of
sterlc repulsion appears likely to remain, the
mixing term becoming negligible (73, 75). In
that case kTlnW may be considered to repre-
sent an activation free energy for aggregation.
Because of the uncertainty about the conforma-
tional freedom of the hairs, the effect cannot be
calculated, although the slope observed in Fig-
ure 6 is not unreasonable. The other explana-
tion is that even one remaining hair causes too
much repulsion and that the possibility of last-
ing contact depends on the orlentation that two
partly depleted micelles have with respect to
each other when meeting. Only if both micelles
are "bare" at the area of meeting, would aggre-
gation occur. This yields a similar relation be-
tween log W and degree of conversion (11),
and the predicted slope is close to that ob-
served. Possibly, the real situation involves
both mechanisms. [The line in Figure 6 cannot
be extrapolated to lower degrees of conversion,
since at, say, 65% conversion, a very slow
aggregation could be observed (73); partially
converted micelles may be subject to another,
very slow, aggregation reaction.]
Another repulsion also exists. It has been
observed (9) that only at high temperature
(about 60°C) and a not too low ea2+ activity
does the aggregation rate approximate the
Smoluchowski limit; it is then slower by a
factor of only two or three. The temperature
effect is presumably explained by additional
repulsion due to protruding f3-casein hairs at
lower temperature. A clear effect of calcium
has been fOWld: at a Ca2+ concentration below
2 mM and 30°C, fully converted micelles do
1975
,
6.86 7.2
4J
I
J' __ 0_~_·6 0 \
j
. ............
~0 20 30
heating tlrre (mr n)
Figure 7. The relative rate of aggregation of casein
micelles, expressed as the stability factor W, in whey
protein-free skim milk at 140·C as a function of time,
derived from the turbidity change at 1150 nm. Mer (65).
not or hardly aggregate (9). One may expect
this to act via the higher charge on the micelles,
because Ca ions lower the charge. However,
lowering the pH at constant Ca2+ activity (and
thus lowering the charge) hardly affects rennet
coagulation rate. Although no convincing
results have been published on the effect of pH
on the aggregation rate of fully renneted
micelles, the results on firming rate of rennet
gels at a stage where the enzymatic reaction is
virtually complete leave little doubt (18, 84).
Moreover, at high temperature, variation in the
Ca2+ activity has far less effect on the aggrega-
tion rate (9). Thus, variations in the electric
charge are unlikely to have much effect under
these conditions. This is in agreement with
calculations done according to the DLVO the-
ory that show little effect of variation in elec-
trostatic repulsion for paracasein micelles (18).
Another explanation of the effect of calcium
may be that deflocculation of paracasein
micelles occurs if Ca2+ activity is low com-
pared with its activity at saturation. In other
words, ea may be needed in providing lasting
bonds, and if insufficient Ca is present, once
aggregated micelles may separate again. This
Journal of Dairy Science Vol. 73, No.8, 1990
1976
works out as if the stability factor is increased.
In practice, renneting occurs much faster at
lower pH. This is partly due to the enzymic
reaction proceeding faster (70). Moreover, at
lower pH chymosin tends to absorb onto
paracasein micelles, and once adsorbed, it may
thus remove K-casein hairs over a certain area
of one micelle before desorbing and diffusing
away. This implies that at an earlier stage bare
patches are formed, permitting aggregation of
micelles at a lower degree of conversion, as is
indeed observed (70).
HEAT COAGULATION
The heat stability of milk is an intricate
subject. The strong and strange dependence of
heat coagulation time (HCf) on pH, with an
optimum of about 6.6 and a pessimum (i.e., the
pH at which the HCf shows a local minimum)
of about 6.85, has long defied explanation.
Heating milk profoundly affects the serum
composition, the most striking change being a
gradual lowering of the pH, but the micelles
may also change considerably. In recent work
by van Boekel et al. (64, 65, 66), much of the
literature is also reviewed, and we will largely
follow those studies.
Figure 7 gives information on the rate of
aggregation reactions of the casein micelles at
high temperature, in this case, without whey
proteins being present. Figure 7 shows that two
aggregations reactions occur, and these were
studied in more detail. Reaction 1 starts imme~
diately after a high temperature is reached,
mostly at a slow, constant rate. It does not
significantly depend on temperature in the
range 120 to 140°C. It is very dependent on the
Ca2+ activity, and it most probably leads to
formation of salt bridges between the micelles.
If only this reaction has occurred, the ag-
gregated micelles can be dispersed again by
CCP-dissolving agents. The reaction is depen-
dent on pH, proceeding fast at very low initial
pH; but it is fairly independent of the rate of
lowering of pH during heating, which may not
be surprising, because such a lowering does not
significantly affect the Ca2+ activity. The reac-
tion is second order with respect to time and
casein concentration (other conditions being
equal). Reaction 2 starts only after the pH has
reached a low value, and then proceeds at a
rapidly increasing rate (Figure 7). The HCf (if
Journal of Dairy Science Vol. 73, No.8, 1990
WALSTRA
type of
coseln Jortrcle
o
o
I
I cr T
high
000 non-dl'pll'tE,d
I
nDn - dep,eted ncn-deplete.: d ~p[l'ted :Jl'plet'-"G
I~;:mlrrn~'_",_,o_ca+(_~+ + -I
~eactlon "chemlcJI-
+ +
H~T
6.2 6.4 6.6 6.8 7.0 7.2
Inltlol r:H
Figure 8. Model for the effect of initial pH heat coagu-
lation time (HC11. on the type of casein micelle emerging
at high temperature and on the nature of the rate determin-
ing reaction. After (64).
due to reaction 2) is therefore dependent on the
initial pH and on the rate of pH decrease, but
not on protein concentration, since that hardly
affects the rate of acidification. The tempera-
ture dependence is high, QlO being about 3,
which is somewhat higher than the QlO for acid
production. The aggregates formed by this reac-
tion cannot be redispersed any more, neither by
CCP dissolving agents, nor by urea, nor by
both. The aggregation thus appears to be caused
by chemical cross-linking.
At high temperature K-casein becomes disso-
ciated form the micelles at high pH (1). In the
absence of whey proteins, the dissociation in-
creases from about 0 to over 50% in the pH
range 6.2 to 7.6 (57). In normal milk, the
depletion of K-casein is less at low pH and
higher at high pH, a comparable transition oc-
curring in the range 6.5 to 6.9. This is presum-
ably caused by the association of 13-lactoglobu-
lin with K-casein, which occurs at all pH: at
low pH denatured 13-1actoglobulin is at the
 MILES-MARSCHALL AWARD INVITED LECTURE
micelles, at high pH in the serum. The deple-
tion of x:-casein from the micelles also depends
on the Ca2+ activity.
From these and some other observations, a
model for the heat coagulation was proposed,
which is schematically given in Figure 8.
Depleted micelles are supposed to be far less
stable than nondepleted ones because of the
scarcity of x:-casein hairs. In either case, the
stability will increase with pH because of in-
creased charge, hence, increased electrostatic
repulsion. Near pessimum pH, salt-induced ag-
gregation predominates if the HCf is indeed
short; this also depends on the ratio of x:-casein
to 13-lactoglobulin and the Ca2+ activity. As the
pH decreases to values below the pessimum,
the x:-casein-I3-lactoglobulin complex associates
again with the micelles, rendering them far
more stable. When starting at a high pH, this
may already happen before the micelles have
aggregated to a considerable extent via reaction
1; this is reasonable, because during this
acidification, the Ca2+ activity does not materi-
ally increase and it is thus much lower at, say,
pH 6.8 than when starting to heat at that pH.
Which cross-linking reaction is involved is
unknown. Several reactions have been reported
(79), and they all proceed faster at a higher
temperature but also at a higher pH. This seems
to disagree with the effect of pH on HCf.
However, at lower pH there will be less elec-
trostatic repulsion and as illustrated in Figure 5,
this will provide opportunities for the hairy
layers to interpenetrate more deeply and more
often, thereby providing more possibilities for
reactive side groups on the peptide chains to
make contact.
The heat stability of concentrated milk. is at
present under study in the author's laboratory
(Nieuwenhuijse et al., unpublished). The con-
clusions reached for nonconcentrated milk. ap-
pear to hold The theory of fractal floc forma-
tion can be usefully applied at low pH. At
higher pH, where the micelles are depleted of
x:-casein, rapid fusion of flocculated micelles
occurs, making the heat coagulation still more
intricate.
ACKNOWLEDGMENTS
The author gratefully acknowledges the con-
tributions of several co-workers: Leon Bremer,
Harrie van den Bijgaart, Tiny van Boekel, Hein
1977
van Dijt, Tom Geurts, Toon van Hooydonk,
Hans Nieuwenhuijse, Ton van Vliet, and Nel
Zoon. Useful comments on a frrst draft of this
article were given by some of these co-workers.
(H. v.D, H. N., T. v.V.) and by D. G. Dalgleish
and D. S. Home of the Hannah Research Insti-
tute, Ayr, Scotland
REFERENCES
I Aoki, T., H. Suzuki, and T. Imamura. 1974. Formation of
soluble casein in whey protein-free milk heated at high
temperature. Milchwissenscbaft 29:589.
2 Bremer, L.G.B., T. van Vliet, and P. Walstra. 1989.
Theoretical and experimental study of the structure of
casein gels. J. Chem. Soc., Faraday Trans. I. 85:3359.
3 Buchheim, W., and U. Welsch. 1973. Evidence for the
submicellar composition of casein micelles on the basis
of electron microscopical studies. Neth. Milk Dairy J. 27:
163.
4 Carlson, A., G. C. Hill, and N. F. Olson. 1986. The
coagulation of milk with immobilized enzymes: a critical
review. Enzyme Microbiol. Technol. 8:642.
5 Chen, C., and K. Yamauchi. 1969. Changes of salt
distribution in milk during frozen storage and its partial
reversion after thawing. Agric. BioI. Chem. 33:1333.
6 Creamer, L. K., G. P. Berry, and O. E. Mills. 1977. A
study of the dissociation of ~casein of the bovine casein
micelle at low temperature. NZ. J. Dairy Sci. Technol.
12:58.
7 Creamer, L. K., and A. R. Matheson. 1980. Effect of beat
treatment on the proteins of pasteurized skim milk. N Z.
J. Dairy Sci. Technol. 15:37.
8 Dalgleish, D. G. 1980. A mechanism for the cbymosin-
induced flocculation of casein micelles. Biophys. Chem.
11:147.
9 Dalgleish, D. G. 1983. Coagulation of renneted bovine
casein micelles: dependence on temperature, calcium ion
concentration and ionic strength. J. Dairy Res. 50:331.
10 Dalgleish, D. G., J. Brinkhuis, and T.AJ. Payens. 1981.
The coagulation of differently sized casein micelles by
rennet. Eur. J. Biochem. 119:257.
11 Dalgleish, D. G., and C. Holt. 1988. A geometrical model
to describe the initial aggregation of partly renneted
casein micelles. J. Co1k>id Interface Sci. 123:80.
12 Dalgleish, D. G., D. S. Home, and AJ.R. Law. 1989.
Size related differences in bovine casein micelles. Bi-
ochim. Biophys. Acta 991:383.
13 Dalgleish, D. G., and T. G. Parker. 1980. Binding of
calcium ions to bovine a~)l-casein and precipitability of
theprotein-ca1cium ion complexes. J. Dairy Res. 47:113.
14 Darling, D. F., and A.C.M van Hooydonk. 1981.
Derivation of a mathematical model for the mechanism
of casein micelle coagulation by rennet. J. Dairy Res. 48:
189.
15 Donnelly, W. J., G. P. McNeill, W. Buchheim, and
T.CA McGann. 1984. Comprehensive study of the
relationship between size and prorein composition in
natural bovine casein micelles. Biochim. Biophys. Acta
789:136.
16 Family, F., and D. P. Landau, ed. 1984. Kinetics of
aggregation and gelation. North Holland. Publ., Amster-
Journal of Dairy Science Vol. 73, No.8, 1990
1978 WALSTRA
dam, Neth.
17 Farrell, H. M, and M P. Thompson. 1971. Biological
significance ofmilk protein polymorphism. I. Daily Sci.
54:1219.
18 Green, M L., and G. Crutchfield. 1971. Density gradient
electrophoresis of native and rennet-treated casein
micelles. I. Daily. Res. 38:151.
19 Green. M L., D. G. Hobbs, S. V. Morant, and V. A. Hill.
1978. Inter-micellar relationships in rennet-treated milk.
I. Preparation of representative electron micrographs. I.
Dairy Res. 45:413.
20 Griffin, M.C.A., and G.ClC. Roberts. 1985. A
IH-n.m.r. study of casein micelles. Biochem. I. 228:273.
21 Harwalker, V. R. 1982. Age gelation of sterilized milk.
Page 229 in Developments in dairy chemistry. Vol. 1.
Proteins. P. F. Fox, ed. Applied Sci. PubL New Yon,
NY.
22 Hiemenz, P. C. 1986. Principles of colloid and surface
chemistry. 2nd ed. Dekker, New York, NY.
23 Holt, C. 1985. The milk salts; their secretion, concentra-
tions and physical chemistry. Page 143 in Developments
in dairy chemistry-3. Lactose and minor constilUents. P.
F. Fox, ed. Elsevier Appl SCi. Pub!. London, UK.
24 Holt, C., and D. G. Dalgleish. 1986. Electrophoretic and
hydrodynamic properties ofbovine casein micelles inter-
preted in terms of particles with an outer haiIy layer. I.
Colloid Interface Sci. 114:513.
25 Holt, C., MU-M. van Kemenade, L. S. Nelson, L.
Sawyer, J. E. Harries, R. T. Baily, and D.W L. Hukins.
1989. I. Daily Res. 56:411.
26 Horne, D. S. 1984. Sterlc effects in the coagulation of
casein micelles by ethanol. Biopolymers 23:989.
27 Home, D. S. 1986. Sterlc stabilization and casein micelle
stability. I. Colloid Interface Sci. 111:250.
28 Home, D. S. 1989. Application of fractal concepts to the
study of casein aggregation phenomena. I. Daily Res. 56:
535.
29 Horne, D. S., and C. M Davidson. 1986. The effect of
environmental conditions on the sterlc stabilization of
casein micelles. Colloid. Pol. Sci. 264:727.
30 Home, D. S., and T. G. Parker. 1981. Factors affecting
the ethanol stability of bovine milk. I. Effect of serum
phase components. J. Daily Res. 48:273.
31 Home, D. S., and T. G. Parker. 1981. Factors affecting
the ethanol stability of bovine milk. n. The origin of the
pH transition. J. Daily Res. 48:285.
32IsraeIachvili, 1. N. 1985. Intermolecular and surface
forces. Academic Press, London, UK.
33 Koops, I. 1967. Preparation and properties of sterilized
coffee cream. I. In-bottle sterilization. Neth. Milk Daily
I. 21:29.
34 Kowalchyk, A. W., and N. F. Olson. 1977. Effects of pH
and tempenllUre on the secondary phase of milk clotting
by rennet. I. Daily Sci. 60:1256.
35 McMahon, D. J., and R. J. Brown. 1984. Composition,
structure and integrity of casein micelles: a review. I.
Daily Sci. 67:499.
36 Meakin, P. 1983. Formation of fractal clusters and
netwom by irreversible diffusion limited cluster-cluster
aggregation. Phys. Rev. Lett. 51:1119.
37 Morr, C. V. 1967. Effect of oxalate and urea upon
ultracentrifugation properties of raw and heated skim-
milk casein micelles. I. Dairy Sci. 50:1744.
38 Mulder, H., and P. Walstra. 1974. The milk fat globule.
IournaI of Daily SCience Vol. 73, No.8, 1990
Pudoc. Wageningen, Neth.
39 Nienwenhuijse, 1. A., W. Timmermans, and P. Walstra.
1988. Calcium and phosphate partitions during the
manufacture of sterilized concentrated milk and their
relations to heat stability. Neth. Milk Dairy 1. 42:387.
40 Ono, T., and T. Obata.1989. A model for the assembly of
bovine casein micelles from F2 and F3 subunits. I. Dairy
Res. 56:453.
41 Ono, T., S.Odagi, and T. Takagi. 1983. Separation of the
submicelles from micellar casein by high performance
gel chromatography. I. Dairy Res. 50:37.
42 Overbeek, I.T.G. 1952. Kinetics of flocculation. Page
278 in Colloid science. VoI.1. Irreversible systems. H. R.
Kruyt, ed. Elsevier, Amsterdam, Neth.
43 Payens, T.AJ. 1979. Casein micelles: the coloid-chemi-
cal approach. I. Daily Res. 46:291.
44 Payens, T.AJ.1982. Stable and unstable casein micelles.
1. Dairy Sci. 65:1863.
45 Payens, T.A.J. 1989. The enzyme-triggered coagulation
of casein micelles. Adv. Colloid Interface Sci. 30:31.
46 Pessen, H., T. F. Kumosinski, and H. M. Farrell. 1989.
Small-angle X-ray scattering investigation of the micel-
lar and submicellar forms of bovine casein. I. Daily Res.
56:443.
47 ROOs, S.P.F.M. 1986. Structure of acid casein gels.
Ph.D. Diss., Wageningen Agric. Univ., Wageningen,
Neth.
48 ROOs, S.P.F.M, P. Walstra, D. G. Dalgleish, and D. S.
Home. 1985. Preliminary note on the change in casein
micelles caused by acidification. Neth Milk Daily I. 39:
119.
49 Rollema, H. S., I. A. Brinkhuis, and H. I. Vreeman. 1988.
lH-NMR studies of bovine lC-casein and casein micelles.
Neth. Milk Daily 1. 42:233.
50 Scheutjens, I.M.H.M., and G. J. Fleer. 1985. Interaction
between two adsorbed polymer layers. Macromolecules
18:1882.
51 Schipper, C.l. 1961. (The caseinate phosphate complex
in milk). Ph.D. Diss., WageningenAgric. Univ., Wagen-
ingen, Neth.
52 Schmidt, D. G. 1982. Association of caseins and casein
micelle structure. Page 61 in Developments in daily
chemistry -1. Proteins. P. F. Fox, ed. AppI. Sci. PubI.,
London, UK.
53 Schmidt, D. G., and P. Both. 1982. Location of asl-' l'l-
and lC-casein in artificial casein micelles. Milchwissen-
schaft 37:336.
54 Schmidt, D. G., and T.AJ. Payens. 1976. Colloidal
aspects of casein. Surface Colloid Sci. 9:165.
55 Schmidt, D. G., and I. K. Pol. 1986. Electrokinetic
measurements on heated and unheated casein micelle
systems. Neth. Milk Daily J. 40:269.
56 Shimmin. P. D., and R. D. Hill. 1964. En electron
microscope study of the internal structure of casein
micelles. I. Daily Res. 31:121.
57 Singh, H., and P. F. Fox. 1987. Heatstability ofmilk: role
of l'l-Iactoglobulin in the pH-dependent dissociation of
micellar kappa-casein. J. Daily Res. 54:509.
58 Slattery, C. W. 1976. Review: casein micelle structure;
an examination of models. 1. Dairy Sci. 59:1547.
59 Snoeren, T.H.M., T .A.J. Payens, 1.1eunink, and P. Both.
1975. Electrostatic interaction between kappa-carragee-
nan and kappa-casein. Milchwissenschaft 30:393.
60 SlOthart, P. H., and D. 1. Cebula. 1982. Small-angIe
scattering study of bovine casein micelles and sub-
 MILES-MARSCHALL AWARD INVITED LECTURE
micelles. J. Mol. BioI. l60:39l.
61 Swaisgood,H. E., andJ.R. Brunner. 1962. Characteriza-
tion of kappa-casein. J. Dairy Sci. 45:1
62 Thompson, M. P., and H. M. Farrell. 1973. The casein
micelle - the forces contributing to its integrity. Neth.
Milk Dairy J. 27:220.
63 Tolstoguzov, V. B. 1986. Functional properties of pro-
tein-polysaccbaride mixtures. Page 385 in J. R. Mitchell
and D. A. Ledward, ed. Functional properties of food
macromolecules. msevier Appl. Sci. PubI., London, UK.
64 van Boeke!, M.AJ.S., J. A. Nieuwenhuijse, and P.
Walstra. 1989. The heat coagulation of milk. 1. Mecha-
nisms. Neth. Milk Dairy J. 43:97.
65 van Boeke!, M.AJ.S., J. A. Nieuwenhuijse, and P.
Walstra. 1989. The heat coagulation of milk. 2. Kinetic
stndies. Neth. Milk Dairy J. 43:129.
66 van Boekel, M.AJ.S., J. A. Nieuwenhuijse, and P.
Walstra.1989. Theheatcoagulationofmilk. 3. Compar-
ison of theory and experiment. Neth. Milk Dairy J. 43:
147.
67 van den Bijgaart, HJ.C.M. 1988. Syneresis of rennet-
induced milk gels as influenced by cheesemaldng
parameters. PhD. Diss., Wageningen Agric. Uuiv.,
Wageningen, Neth.
68 van den Bijgaart, HJ.C.M., T. van Vliet, P. Walstra, and
P. Zoon. 1989. On the rate of swelling of (para)casein
micelles after a drop in temperature. Neth. Milk Dairy J.
43:85.
69 van Hooydonk, A.C.M., I. J. Boerrigter, and H. G.
Hagedoom 1986. pH-induced physico-chemical
changes of casein micelles in milk and their effect on
renneting. 1. Effect of acidification on physiclrChemical
properties. Neth. Milk Dairy 1. 40:281.
70 van Hooydonk, A.C.M., I. J. Boerrigter, and H. G.
Hagedoom 1986. pH-induced physico-chemical
changes of casein micelles in milk and their effect on
renneting. 2. Effect of pH on remlC:ting of milk. Neth.
Milk Dairy J. 40:297.
71 van Hooydonk, A.C.M., H. G. Hagedoom, and I. J.
Boemgter. 1986. The effects of various cations on the
1979
renneting of milk. Neth. Milk Dairy J. 40:369.
72 van Hooydonk, A.C.M., C. Olieman, and H. G.
Hagedoom 1984. Kinetics of the chymosin-catalysed
proteolysis of lC-casein in milk. Neth. Milk Dairy J. 38:
207.
73 van Hooydonk, A.C.M., and P. Walstra. 1987. interpre-
tation of the kinetics of the renneting reaction in milk.
Neth. Milk Dairy J. 41:19.
74 van Vliet, T., S'p.F.M. Roofs, P. Zoon, and P. Walstra.
1989. Rheological properties of casein gels. J. Dairy Res.
56:529.
75 Vincent, B. 1974. The effect of adsorbed polymers on
dispersion stability. Adv. Colloid Interface Sci. 4:193.
76 Vreeman, H. J., B. W. van Markwijk, and P. Both. 1989.
The structure of casein micelles between pH 55 and 6.7
as determined by light scattering, electron microscopy
and voluminosity experiments. J. Dairy Res. 56:463.
77 Walstra, P. 1979. The voluminosity of casein micelles
and some of its implications. J. Dairy Res. 46:317.
78 Walstra, P., V. A. Bloomfield, G. J. Wei, and R. Jenness.
1981. Effect of chymosin on the hydrodynamic diameter
of casein micelles. Bioch. Biophys. Acta 669:258.
79 Walstra, P., and R. Jenness. 1984. Dairy chemistry and
physics. John Wiley, New York:, NY.
80 Wiechen, V. A., and A. -M. Knoop. 1978. Investigations
on ea distribution between serum and casein by means of
the radioisotope ea 45 in low cooled and pasteurized
milks. Milchwissenschaft 33:213.
81 Yamauchi, K., Y. Yoneda, Y. Koga, and T. Tsugo. 1969.
Exchangeability of colloidal caleium in milk with solu-
ble calcium. Agric. BioI. Chem. 33:907.
82 Zoon, P., T. van Vliet, and P. Walstra. 1988. Rheological
properties of rennet-induced skim-milk gels. 1. Introduc-
tion. Neth. Milk Dairy J. 42:249.
83 Zoon, P., T. van Vliet, and P. Walstra. 1988. Rheological
properties of rennet-induced skim-milk gels. 2. The
effect of temperature. Neth. Milk Dairy J. 42:271.
84 Zoon, P., T. van Vliet, andP. Walstra. 1989. Rheological
properties of rennet-induced skim-milk gels. 4. The
effect of pH and NaCI. Neth. Milk Dairy J. 43:17.
Journal of Dairy Science Vol. 73, No.8, 1990