CHAPTER 6

Enzymatic Coagulation of Milk

As discussed in Chapter 2, the milk for most tached to the casein micelle, whereas the C-ter-
cheese varieties is coagulated through the action minal part, referred to as the (caseino) macro-
of selected proteinases, called rennets. The ren- peptide (CMP; or glycomacropeptide, since it
net-induced coagulation of milk is in fact a two- contains the carbohydrate moieties of K-casein)
stage process (Figure 6-1). The primary phase is lost in the surrounding aqueous medium. It has
involves the specific enzymatic modification of been recognized since the end of the 19th cen-
the casein micelles to produce paracasein mi- tury that small peptides are produced upon ren-
celles that aggregate in the presence of Ca2+ at neting. As discussed in Chapter 3, there are
temperatures above about 2O0C. Aggregation of about 10 forms of K-casein that differ in sugar
the rennet-altered micelles is referred to as the content; hence, 10 CMPs are produced. All the
secondary phase of coagulation. The primary CMPs are soluble in 2% trichloroacetic acid
phase of rennet action is well characterized, but (TCA) but only the glycosylated forms are sol-
the secondary phase is less clear. The subject has uble at higher concentrations of TCA. Thus,
been reviewed by Dalgleish (1992, 1993); Fox TCA-soluble N, or more specifically TCA-
(1984); Fox and McSweeney (1997); Fox and soluble sugars (e.g., N-acetyl neuramic acid),
Mulvihill (1990); and Fox, O'Connor, can be used to monitor the primary phase of ren-
McSweeney, Guinee, and O'Brien (1996). net coagulation (Figure 6-3).
The unique sensitivity of the Phe-Met bond of
6.1 THE PRIMARY PHASE OF RENNET K-casein has aroused interest. The dipeptide
COAGULATION H.Phe-Met.OH is not hydrolyzed, nor are tri- or
tetrapeptides containing a Phe-Met bond. How-
As discussed in Chapter 3, the caseins exist as ever, this bond is hydrolyzed in the pentapeptide
micelles stabilized by a surface layer of K- H.Ser-Leu-Phe-Met-AIa-OMe, and reversing
casein. Following the isolation of K-casein in the positions of serine and leucine, to give the
1956, it was shown that this protein is the mi- correct sequence of K-casein, increases the sus-
celle-stabilizing protein and that its stabilizing ceptibility of the Phe-Met bond to chymosin.
properties are destroyed on renneting. Shortly Both the length of the peptide and the sequence
afterwards, it was shown that K-casein is the around the Phe-Met bond are important determi-
only protein hydrolyzed during the rennet co- nants of enzyme-substrate interaction. Serinei04
agulation of milk and that it is hydrolyzed spe- appears to be particularly important, and its re-
cifically at the bond Phei05-Meti06 (Figure 6-2). placement by Ala or even L-Ser in the above
The N-terminal part of the molecule, K-CN fl- pentapeptide renders the Phe-Met bond very re-
105, referred to as para-K-casein, remains at- sistant to hydrolysis by chymosin. Extension of
 rennet
Casein Para-casein + small peptides
(primary, enzymatic phase)

(secondary,
non-enzymatic phase)
Coagulum (gel)
Figure 6-1 Summary of the rennet coagulation of milk. The primary phase involves enzymatic hydrolysis of K-
casein, while the secondary stage involves aggregation of the rennet-altered (paracasein) micelle into a three-
dimensional gel network or coagulum.

the pentapeptide H.Ser.Phe.Met.Ala.Ile.OH tible to hydrolysis by chymosin at pH 4.7 as it is

(i.e., K-CN f!04-108) from the N- and/or C-ter- in intact K-casein; it is hydrolyzed around
minal to reproduce the sequence of K-casein 66,000 times faster than the parent pentapeptide
around the chymosin-susceptible bond increases (K-CN f 104-108), with a kcat/KM of about 2 M-1
the efficiency with which the Phe-Met bond is S"1, which is similar to that for intact K-casein. K-
hydrolyzed by chymosin (Table 6-1). The se- Casein and the peptide K-CN f98-lll are also
quence K-CN f98—111 includes all the residues readily hydrolyzed at pH 6.6, but smaller pep-
necessary to render the Phe-Met bond as suscep- tides are not.

i
Pyro Glu-Glu-Gln-Asn-Gln-Glu-Gln-Pro-Ile-Arg-Cys-Glu-Lys-Asp-Glu-Arg-Phe-Phe-Ser-Asp-

21
Lys-Ile-Ala-Lys-Tyr-Ile-Pro-Ile-Gln-Tyr-Val-Leu-Ser-Arg-Tyr-Pro-Ser-Tyr-Gly-Leu-

41
Asn-Tyr-Tyr-Gln-Gln-Lys-Pro-Val-Ala-Leu-Ile-Asn-Asn-Gln-Phe-Leu-Pro-Tyr-Pro-Tyr-

61
Tyr-Ala-Lys-Pro-Ala-Ala-Val-Arg-Ser-Pro-Ala-Gln-Ile-Leu-Gln-Trp-Gln-Val-Leu-Ser-

81
Asn-Thr-Val-Pro-Ala-Lys-Ser-Cys-Gln-Ala-Gln-Pro-Thr-Thr-Met-Ala-Arg-His-Pro-His-

101 1051 106

Pro-His-Leu-Ser-PhetMet-Ala-Ile-Pro-Pro-Lys-Lys-Asn-Gln-Asp-Lys-Thr-Glu-Ile-Pro-

121 He (Variant B)
Thr-Ile-Asn-Thr-Ile-Ala-Ser-Gly-Glu-Pro-77zr- Ser-Thr -Pro-Thr - -GIu-Ala-Val-Glu-
Thr (Variant A)
141 Ala (Variant B)
Ser-Tfer-Val-Ala-Thr-Leu-Glu- -SerP - Pro-Glu-Val-Ile-Glu-Ser-Pro-Pro-Glu-Ile-Asn-
Asp (Variant A)
161 169
Thr-Val-Gln-Val-Thr-Ser-Thr-Ala-Val.OH

Figure 6-2 Amino acid sequence of K-casein, showing the principal chymosin cleavage site (i); oligosaccha-
rides are attached at some or all of the threonine residues shown in italics.
 TCA-solubleN

Time from rennet addition

Figure 6-3 Release of nitrogen soluble in 2% (O) or 12% (•) TCA by rennet from casein in milk.

The Phe and Met residues in the chymosin- the hydrophobic residues Leuioa and Ileios* at
sensitive bond of K-casein are not intrinsically least one of the three histidines (residues 98,
essential for chymosin action. There are numer- 100, or 102, as indicated by the inhibitory effect
ous Phe and a substantial number of Met resi- of photo-oxidation), and LySm. Studies on
dues in all milk proteins. In porcine and human chemically or enzymatically modified peptide
K-caseins, the chymosin-sensitive bond is Phe- analogues of K-CN f98-112 indicated the rela-
Ue, while in rat and mouse K-caseins, it is Phe- tive importance of residues in the sequences of
Leu; yet, these proteins are readily hydrolyzed 98-102 and 111-112. It has been suggested that
by calf chymosin, although more slowly than the sequence Leuios to Ilcios of K-casein, which
bovine K-casein. In contrast, porcine milk is co- probably exists as an extended (3-structure, fits
agulated more effectively than bovine milk by into the active site cleft of acid proteinases. The
porcine chymosin, indicating that unidentified hydrophobic residues Leuios, Prices, Metioe, and
subtle structural features influence chymosin ac- He iog are directed toward hydrophobic pockets
tion. Peptides in which Phe is replaced by Phe along the active site cleft, while the hydroxyl
(NO2) or cyclohexylamine are also hydrolyzed group of Serio4 forms part of a hydrogen bond
by chymosin, although less effectively than with some counterpart in the enzyme. It has
those with a Phe-Met bond. Oxidation of Met106 been proposed that the sequences 98-102 and
reduces kcat/KM roughly tenfold, but substitution 109-111 form (3-turns around the edges of the
of He for Met increases it about threefold. active site cleft of the enzyme. This conforma-
A genetically engineered mutant of K-casein tion is stabilized by Pro residues at positions 99,
in which Met106 was replaced by Phei06 (i«e., the 101, 109, and 110. The three His residues at po-
chymosin-sensitive bond was changed from sitions 98, 100, 102, and LySm are probably in-
Pheio5-Met106 to Pheios-Pheioe) was hydrolyzed volved in electrostatic bonding between enzyme
1.8 times faster by chymosin than natural and substrate; none appears to have a predomi-
K-casein. These findings suggest that the se- nant role. Lys 112 appears not to be important in
quence around the Phe-Met bond, rather than the enzyme-substrate binding as long as Lysm is
residues in the bond itself, contains the impor- present.
tant determinants of hydrolysis by chymosin. The significance of electrostatic interactions
The particularly important residues are Seri04, in chymosin-substrate complex formation is in-
Table 6-1 Kinetic Parameters for Hydrolysis of K-Casein Peptides by Chymosin at pH 4.7

Peptide Sequence kcat (S-1) MmM) kcat/KM(s-1mM-1)

S.F.M.A.I. 104-108 0.33 8.50 0.038

S.F.M.A.I.P. 104-109 1.05 9.20 0.114
S.F.M.A.I. P.P. 104-110 1.57 6.80 0.231
S.F.M.A.I. P.P.K. 104-111 0.75 3.20 0.239
LS.F.MAI. 103-108 18.3 0.85 21.6
LS.F.M.A.I.P. 103-109 38.1 0.69 55.1
LS.F.M.A.I.P.P. 103-110 43.3 0.41 105.1
LS.F.MAI. P.P.K. 103-111 33.6 0.43 78.3
LS.F.M.A.I.P.P.K.K. 103-112 30.2 0.46 65.3
H.LS.F.M.A.I. 102-108 16.0 0.52 30.8
P.H.LS.F.M.A.I 101-108 33.5 0.34 100.2
H.P.H.P.H.LS.F.M.A.I.P.P.K. 98-1 1 1 66.2 0.026 2509
98-1 11a 46.2a 0.029a 1621a
K-Caseinb 2-20 0.001-0.005 200-2,000
LS.F.(NO2)Nle A.L.OMe 12.0 0.95 12.7
a
pH 6.6.
b
pH 4.6.

dicated by the effect of added NaCl on the rennet portions of acid proteinases in commercial ren-
coagulation time (RCT) of milk. Addition of nets have been proposed.
NaCl up to 3 mM reduces RCT but higher con-
centrations have an inhibitory effect. It is claimed 6.2 RENNET
that the effect of NaCl is on the primary, enzy-
matic phase rather than on the aggregation of ren- Several proteinases will coagulate milk under
net-altered micelles. Increasing ionic strength suitable conditions but most are too proteolytic
(0.01-0.11) reduced the rate of hydrolysis of K- relative to their milk-clotting activity (MCA).
CN f HiS98-LySi imu in a model system; the effect Consequently, they hydrolyze the caseins in the
became more marked as the reaction pH was in- coagulum too quickly, causing a reduced cheese
creased, but it was independent of ion type. yield. (MCA is the inverse of RCT; i.e., MCA =
As well as serving to elucidate the importance 1/RCT.) Excessive proteolysis or incorrect
of certain residues in the hydrolysis of K-casein specificity may also lead to defects in the flavor
by chymosin, small peptides that mimic or are (especially bitterness) and texture of the cheese.
identical to the sequence of K-casein around the Although plant proteinases appear to have been
Phe-Met bond are very useful substrates for de- used as rennets since prehistoric times, gastric
termining the activity of rennets in absolute proteinases from calves, kids, or lambs have
units, that is, independent of variations in the been traditionally used as rennets, with very few
nonenzymatic phase of coagulation of different exceptions.
milks. Standard methods for such quantification Animal rennets are prepared by extracting the
have been developed and chromogenic peptide dried (usually) or salted gastric tissue (referred
substrates are available commercially. Since the to as veils) with 10% NaCl and activating and
specific activity of different rennets on these standardizing the extract. Standard calf rennet
peptides varies, methods for quantifying the pro- contains about 60-70 RU/ml and is preserved by
making the extract up to 20% NaCl and adding 6.3 FACTORS THAT AFFECT THE
sodium benzoate or sodium propionate. A rennet HYDROLYSIS OF K-CASEIN AND THE
unit (RU) is the amount of rennet activity that PRIMARY PHASE OF RENNET
will coagulate 10 ml of milk (usually low-heat COAGULATION
skim milk powder reconstituted in 0.01% CaCl2
and perhaps adjusted to pH 6.5) in 100 s. The hydrolysis of K-casein is influenced by
Chymosin (an aspartyl acid proteinase, i.e., a many factors, some of which are discussed be-
proteinase with two aspartic acid residues at the low. While many factors influence both the pri-
active site and with a pH optimum of 2-4) repre- mary and secondary stages, the effects on each
sents more than 90% of the MCA of good-qual- are discussed separately.
ity veal rennet, the remaining activity being due
to pepsin. As the animal ages, especially when • The pH optimum for chymosin and bovine
fed solid food, the secretion of chymosin de- pepsin on small synthetic peptides is about
clines and that of pepsin increases. 4.7 but is 5.3-5.5 on K-CN f HiS98-LySm71I2.
Like many other animal proteinases, chy- Chymosin hydrolyzes insulin, acid-dena-
mosin is secreted as its zymogen, prochymosin, tured hemoglobin and Na-caseinate opti-
which is autocatalytically activated on acidifica- mally at pH 4.0, 3.5, and 3.5, respectively.
tion to pH 2-^ by removal of a 44-residue pep- The pH optimum for the first stage of rennet
tide from the N-terminal of the zymogen (see action in milk is about 6.0 with 40C or
Foltmann, 1993). 3O0C.
Chymosin is well characterized at the molecu- • The influence of ionic strength on the pri-
lar level (see Chitpinityol & Crabbe, 1998; mary phase of rennet coagulation is dis-
Foltmann, 1993). The enzyme, which was crys- cussed in Section 6.1.
tallized in the 1960s, is a single-chain polypep- • The optimum temperature for the coagula-
tide that contains about 323 amino acid residues tion of milk by calf rennet at pH 6.6 is
and has a molecular mass of 35,600 Da. Its pri- around 450C. Presumably, the optimum for
mary structure has been established, and a con- the hydrolysis of K-casein is around this
siderable amount of information is available on value. The temperature coefficient (Q10) for
its secondary and tertiary structures. The mol- the hydrolysis of K-casein in solutions of
ecule exists as two domains separated by an ac- Na-caseinate is about 1.8; the activation en-
tive site cleft in which the two catalytically ergy, Ea, is about 40,000 J-mol'1; and the
active aspartyl groups (Asp32 and Asp215) are lo- activation entropy, AS, is about -90
cated. !•Kr^moH. Generally similar values have
Calf rennet contains three chymosin isoen- been reported for the hydrolysis of isolated
zymes, principally A and B, with lesser amounts K-casein by chymosin.
of C. Chymosins A and B are produced from the • Heat treatment of milk at temperatures
corresponding zymogens, prochymosins A and above 650C adversely affects its rennet co-
B, but chymosin C appears to be a degradation agulability. If the heat treatment is very
product of chymosin A that lacks three residues, severe (> 9O0C for 10 min), the milk fails
ASp244-PhC246. The specific activity of chymosin to coagulate upon renneting. Although
A, B, and C is 120, 100, and 50 RU/mg, respec- changes in salts equilibria are contributory
tively. Chymosins A and B differ by a single factors, the principal causative factor is in-
amino acid substitution, Asp and GIy, respec- termolecular disulphide bond formation be-
tively, at position 244 and have an optimum pH tween K-casein and (3-lactoglobulin and/or
at 4.2 and 3.7, respectively. a-lactalbumin. Both the primary and espe-
The properties of different rennets are dis- cially the secondary phases of rennet action
cussed in Section 6.8. are inhibited in heated milk, as reflected by
 the marked decrease in the curd-firming ures 6-4 and 6-5). It has been suggested that the
rate and in the strength of the resulting gel. decrease in viscosity is due to a decrease in the
The adverse effects of heating can be re- voluminosity of the casein micelles following
versed by acidification to pH values in the release of the macropeptides, which form a
region 6.6-6.0 before or after heating or by "hairy layer" about 10 nm thick (Figure 6-4).
the addition of CaCl2 (which causes a re- The decrease in micelle size has been confirmed
duction in pH). The secondary, rather than by quasi-elastic light scattering.
the primary, phase of rennet action prob- It is generally agreed that following the initial
ably benefits from these treatments. lag period, the viscosity increases exponentially
up to the onset of visual coagulation or gelation,
that is, 100% RCT (Figure 6-5). The gelation
6.4 THE SECONDARY process, generally referred to as the secondary
(NONENZYMATIC) PHASE OF phase of rennet coagulation, involves initially
COAGULATION AND GEL ASSEMBLY the formation of chains and clumps of micelles
and leads eventually to the formation of a net-
Hydrolysis of K-casein by chymosin or simi- work of partly fused micelles. During the first
lar enzymes during the primary phase of rennet 60% of the visually observed RCT, the micelles
action releases the highly charged, hydrophilic exist as individual particles; the primary enzy-
C-terminal segment of K-casein (macropeptide), matic reaction is about 85% complete at 60% of
as a result of which the zeta potential of the the visual RCT. Between 60% and 80% of the
casein micelles is reduced from -10/-20 to -5/ RCT, the rennet-altered micelles begin to aggre-
-7 mV and the protruding peptides (hairs) are gate steadily, with no sudden change in the type
removed from their surfaces, thus destroying the or extent of aggregation. Small chain-like aggre-
principal micelle-stabilizing factors (electro- gates, rather than clumps, form initially (Figure
static and steric) and their colloidal stability. 6-6). At 100% of the RCT, most of the micelles
When roughly 85% of the total K-casein has have aggregated into short chains, which then
been hydrolyzed, the stability of the micelles is begin to aggregate to form a network. Initially,
reduced to such an extent that when they collide, most micelles are linked by bridges (655 nm
they remain in contact and eventually build into long and 40 nm wide) and do not touch. An un-
a three-dimensional network, referred to as a co- expectedly large proportion of the surface of the
agulum or gel (Figure 6—4). Gel formation is ac- participating micelles is involved in the bridg-
companied by sharp increases in viscosity and ing, indicating a large amount of material, the
elastic shear modulus, G', which is a measure of origin of which is unknown. Although the mi-
gel firmness (Figure 6-5; see Section 6.6.2). Re- celles themselves appear the most probable
ducing the pH or increasing the temperature source, if they are the source, micellar rearrange-
from the normal values (~ 6.6 and ~ 310C, re- ment would be necessary. No change is ob-
spectively) permits coagulation at a lower level served in the size, surface structure, or general
of K-casein hydrolysis. Although the precise re- appearance of the micelles up to 60% of the RCT
actions involved in aggregation are not known, (i.e., lag phase). Thus, if a micellar rearrange-
the kinetics of aggregation have been described. ment is a prerequisite for aggregation, it must
The assembly of rennet-altered micelles into a occur during the latter half of the RCT. Linkage
gel has been studied using various forms of vis- of the rennet-altered micelles probably occurs at
cometry, electron microscopy, and light scat- definite surface sites.
tering. Viscosity measurements show that the Aggregation of the rennet-altered micelles
viscosity of renneted milk remains constant or can be described by the von Smoluchowski
decreases slightly during a period equivalent to theory for diffusion-controlled aggregation of
roughly 60% of the visually observed RCT (Fig- hydrophobic colloids when allowance is made
 a b c

Rfilfiasft nf mar.Yrmfimiilfis
n xei

% of visually observed clotting time

d
Figure 6—4 Schematic representation of the rennet coagulation of milk, (a) Casein micelles with intact K-casein
layer being attacked by the chymosin (C); (b) micelles partially denuded of K-casein; (c) extensively denuded
micelles in the process of aggregation; (d) release of macropeptides (^) and changes in relative viscosity (Q)
during the course of rennet coagulation.

for the need to produce, by enzymatic hydroly- iting and is determined by the random fruitful
sis, a sufficient concentration of particles ca- collision of particles (rennet-altered micelles).
pable of aggregating (i.e., casein micelles in The rate of aggregation is not consistent with a
which > 97% of the K-casein has been hydro- branching process model, since the micellar
Iyzed). The diffusion of the particles is rate lim- functionality is 1.8 whereas an average function-
 (A)
Phase angle, 6, °

Viscosity, Pas
(B)
Gel moduli, Pa

Time from rennet addition, min

Figure 6-5 Rheological changes in milk during rennet coagulation under quiescent conditions. (A) Phase angle
(•) and viscosity (O); (B) elastic modulus (•) and loss modulus (Q).

ality greater than 2 is required for network for- 3. aggregation of paracasein micelles via a
mation. von Smoluchowski process
According to Dalgleish (1980), the overall
rennet coagulation of milk can be described by The overall clotting time tc is the sum of the
combining three factors: enzymatic phase and the aggregation phase:

1. proteolysis of K-casein, which may be de-

scribed by Michaelis-Menten kinetics
.=W"^4^}
^max V 1 ac J + ^S
%iax ZK
0+ ^(^-l)
s^Q \ 1V10 )
2. the requirement that around 97% of the K-
casein on a micelle be hydrolyzed before
it can participate in aggregation where: Km and Vmax are the Michaelis-Menten
 Degree of coagulation

% of visual rennet coagulation time

Gel suitable for

cutting

Figure 6-6 Schematic representation of the progress of micelle aggregation during the rennet coagulation of
milk. Aggregation of rennet-altered micelles results in the formation aggregates that fuse to form chain-like
structures, and these eventually overlap and cross-link to form a three-dimension casein network or gel after a
time greater than the visual rennet coagulation time.

parameters, occ is the extent of K-casein hydroly-

sis, S0 is the initial concentration of K-casein, ks ,.^^^.^^^±-±]
is the rate constant for aggregation, C0 is the con-
centration of aggregating material, Mcrit is the where V is the velocity of enzymatic hydrolysis
weight average molecular weight at tc («10 mi- of K-casein, S0 is the initial concentration of K-
cellar units), and M0 is the weight average mo- casein, Cm is a constant relating the stability of
lecular weight at t0. the casein micelle to K-casein concentration, W0
Darling and van Hooydonk (1981) proposed is the initial stability factor for casein micelles,
an alternative model for rennet coagulation, W0 is the initial concentration of casein micelles,
again by combining Michaelis-Menten enzyme and nc is the concentration of casein aggregates
kinetics with von Smoluchowski aggregation at the observed clotting time tc. It is claimed that
kinetics. The stability factor in von this theoretical model explains the experimen-
Smoluchowski's theory is considered as a vari- tally observed influence of protein concentra-
able determined by the concentration of un- tion, enzyme concentration, and temperature on
hydrolyzed surface K-casein. The coagulation RCT and the occurrence of a lag phase equal to
time, tc, is given by 60% of RCT.
 With milk of normal concentration, about 90% where E is the enzyme, S is the substrate, P is the
of the micelles are incorporated into the curd at reaction product, P* is the paracasein micelle
100% of the visual RCT, but only about 50% are with transformed quaternary structure, and Pn is
incorporated in a fourfold concentrate when the the gelled micelle aggregate. The first two steps
same level of rennet is used. The micelles that are are the Michaelis-Menten model for the primary,
"free" at or after the RCT may react differently enzymatic phase and are essentially as proposed
from those free prior to RCT. That is, before vi- by Payens, Wiersma, and Brinkhuis, 1977):
sual coagulation, all micelles are freely dispersed
in the serum and can aggregate randomly, but
once a gel matrix has started to form, free micelles k, k2
may react either with the gel matrix or with other E + S " •- ES >• E + P1 + M
free micelles. Therefore, a gel assembly may be k_,
regarded as a two-stage process, and the proper-
ties of the final curd may be affected considerably where P\ is para-K-casein and M is a
by the amount of casein free during the RCT. macropeptide. Payens et al. (1977) suggested
Since this amount is particularly high in concen- that the second, nonenzymatic phase may be
trated milks, it may explain the coarser structure represented by
of curd made from these milks.
Measurements made using the Instron Uni-
versal Testing machine showed that the rate of
firming of a renneted milk gel as a function of
Jl_
time has two maxima (Stony & Ford, 1982): un-
der the experimental conditions used, firming where i is any number of the aggregating particle
was first observed about 2.5 min after the visual Pi.
assessment of clotting, and the firming rate in- Surkov et al. (1982) suggested that the en-
creased to a maximum 12-15 min after clotting. zyme-altered micelles (paracasein micelle P)
The rate decreased over the next 10-15 min to undergo a cooperative transition in quaternary
about 80% of the maximum value, after which it structure to yield clot-forming particles (P*)
either remained constant or increased slightly with a rate constant kc and with E0 = 191 kJ
for a further 15 min and thereafter decreased moh1 and QIQ-C = 16. These values are close to
steadily. However, this two-stage gel firming those reported by Tuszynski (1971).
process has not been observed when dynamic The sites involved in the aggregation process
rheometry, a very sensitive technique, is used. are not known. Following reduction of the mi-
Based on viscometric data, Tuszynski (1971) cellar zeta potential by proteolysis of K-casein,
suggested that gel assembly is a two-stage pro- linkage of particles is facilitated. Interparticle
cess consisting of what he called "flocculation" linkage could be via calcium bridges and/or hy-
and "gelation," but the meaning of these terms is drophobic interactions (which the marked tem-
not clear. Turbidity experiments also suggest a perature-dependence of the secondary phase
two-stage gel assembly process (Surkov, would indicate). Changes in the surface hydro-
Klimovskii, & Krayushkin, 1982), although phobicity of casein micelles during renneting
again the terminology is not very clear: has been demonstrated through changes in the
binding of the fluorescent marker 8-anilino
naphthalene-1-sulfonate (Iametti, Giangiacomo,
Messina, & Bonomi, 1993; Peri, Pagliarini,
Iametti, & Bonomi, 1990). The hydrophobic
amino terminal segment (residues 14-24) of
(Xsi-casein appears to be important in the estab- where x is time required for the coagulation of
lishment of a rennet curd structure. It has been the enzymatically altered casein micelles and
suggested that the matrix of young cheese curd tc-x is the time required for the enzymatic stage.
consists of a network of ocsi-casein molecules Rearrangement of this equation results in a more
linked together via hydrophobic patches that ex- convenient form:
tends throughout the cheese structure. The soft-
ening of texture during the early stages of ripen-
ing is considered to be due to breaking of the
tc = k ( l / E ) + x
network upon hydrolysis of the Phe23-Phe24 bond
of ocsi-casein. Modification of histidyl, lysyl, and which is valid within a certain range of rennet
arginyl residues in K-casein inhibits the second- concentrations and under certain conditions of
ary phase of rennet coagulation, suggesting that temperature and pH. A very good linear relation-
a positively-charged cluster on para-K-casein in- ship exists between clotting time and the recip-
teracts electrostatically with unidentified nega- rocal of enzyme concentration (Figure 6-7).
tive sites. In native micelles, this positive site The coagulation equations developed by
may be masked or covered by the macropeptide Dalgleish (1980) and by Darling and van
segment of K-casein but becomes exposed and Hooydonk (1981) might be regarded as greatly
reactive when this peptide is released. refined versions of these simpler equations and
Following the RCT, the network appearance reduce to them on first approximations. Rennet
becomes gradually more apparent. The strands clotting time (tc) has also been expressed
of the network, which are more or less in paral- (Payens and Wiersma, 1977) by the equation:
lel, are roughly 5 micelles thick and 10 micelle
diameters apart. The bridges between the mi-
, ^ 1 2
celles contract slowly, forcing the micelles into c
\\kV
V 7 S'max
contact and eventually causing them to fuse par-
tially. The fate of the bridging material upon mi-
celle fusion has not been explained, but it may where ks, the diffusion-controlled flocculation
be that its disappearance is responsible for the rate constant according to von Smoluchowski's
second maximum in the firming rate time curve theory (nonenzymatic phase), is proportional to
and for the reported flocculation and gelation the concentration of reactive (coagulable) par-
stages in the gel assembly process. ticles (proteolyzed micelles) and hence to en-
Normally, the rate of an enzymatic reaction zyme concentration; Fmax is the maximum veloc-
increases linearly with enzyme concentration, ity in Michaelis-Menten kinetics (enzymatic
within certain limits. In the case of rennet coagu- phase) and is proportional to enzyme concentra-
lation, RCT is inversely related to enzyme con- tion; and c is a constant that depends on the ex-
centration, as expressed by the formula perimental conditions.

Et0 = k
6.5 FACTORS THAT AFFECT THE
where E = enzyme concentration and tc is the NONENZYMATIC PHASE OF RENNET
RCT. COAGULATION
This equation, which assumes that visually
The coagulation of renneted micelles is very
observed coagulation is dependent only on the
temperature dependent (Q10 ~ 16), and bovine
enzymatic process, has been modified to take
milk does not coagulate at less than around 180C
account of the duration of the secondary, nonen-
unless the Ca2+ concentration is increased. The
zymatic phase:
marked difference between the temperature de-
E(tc-x) = k pendence of the enzymatic and nonenzymatic
 the secondary phase of coagulation. Reduction
in pH in the range 6.6-6.0 is accompanied by in-
Rennet coagulation time

creases in the rates of the enzymic and coagula-

tion reactions, reductions in gelation time (time
for gel onset) and the degree of K-casein hy-
drolysis necessary for the onset of gelation (e.g.,
from « 97% to « 80% of total K-casein), and in-
creases in the curd-firming rate and firmness af-
ter a given renneting time. Although it is claimed
that pH has essentially no effect on the coagula-
tion process, the rate of firming of the resultant
gel is significantly increased upon reducing the
1/E pH (Figure 6-8).
The rate of firming of renneted milk gels is
Figure 6-7 Relationship between enzyme con- influenced by the type of rennet, especially un-
centration (E) and rennet coagulation time. der unfavorable conditions, such as high pH or
low Ca2+. Perhaps differences in the firming rate
reflect the effect of pH on rennet activity or per-
phases of rennet coagulation has been exploited haps some general proteolysis by rennet substi-
in studies on the effects of various factors on the tutes.
rennet coagulation of milk—studies done as part Heat treatment of milk under conditions that
of attempts to develop a system for the continu- denature p-lactoglobulin and promote its inter-
ous coagulation of milk for cheese or rennet action with K-casein via sulfydryl-disulfide in-
casein manufacture and to discover possible ap- teraction adversely affects all aspects of rennet
plications of immobilized rennets. The very high coagulation but especially the buildup of a gel
temperature dependence of rennet coagulation network (Figure 6-9) (van Hooydonk, deKoster,
suggests that hydrophobic interactions are im- & Boerrigter, 1987). Presumably, the attach-
portant. ment of denatured p-lactoglobulin to the surface
Coagulation of rennet-altered micelles de- of the casein micelles (as is evident from elec-
pends on a critical concentration of Ca2+, which tron micrographs of casein micelles) prevents
may act by cross-linking rennet-altered micelles, their aggregation in a form capable of building
possibly via serine phosphate residues or simply up a gel network.
by charge neutralization. Colloidal calcium
phosphate is also essential for coagulation but 6.6 MEASUREMENT OF RENNET
can be replaced by increased Ca2+. Partial enzy- COAGULATION PROPERTIES
matic dephosphorylation of casein, which re-
duces micellar charge, reduces coagulability; in- The rennet gelation of milk under quiescent
teraction of casein micelles with various cationic conditions involves the conversion of milk from
species predisposes them to coagulation by ren- a colloidal dispersion of stable micelles to a net-
net and may even coagulate unrenneted mi- work (gel) of aggregated paracasein micelles,
celles. Chemical modification of histidine, which forms a continuous phase, entrapping
lysine, or arginine residues inhibits coagulation, moisture and fat globules in its pores. The gel
presumably by reducing micellar positive becomes more elastic and firm with time (i.e.,
charge. upon aging). The transformation is accompanied
The apparent importance of micellar charge in by a number of physicochemical changes, in-
the coagulation of rennet-altered micelles sug- cluding hydrolysis of K-casein, with a concomi-
gests that pH should have a major influence on tant increase in the concentration of the gly-
 Elastic shear modulus, G1, kPa

Time after rennet addition, min

Figure 6-8 Development of elastic shear modulus (G', equivalent to curd firmness) in rennet-treated, high-
protein (-180 g/kg) milk retentate obtained from skim milk heated to 10O0C for 12Os and renneted at pH 6.67
(O), 6.55 (•), 6.45 (Q), 6.3 (•), 6.15 (A), and 6.0 (A).

comacropeptide; aggregation of the sensitized Aggregation. The joining of particles, such

paracasein micelles; increases in viscosity and as micelles, by various types of electrostatic
elasticity; and a decrease in the ratio of the vis- or hydrophobic bonds. The aggregates are
cous to elastic character of the milk. Such visible by electron microscopy.
changes may also alter some of the physical Coagulation or flocculation. The collision
properties of the milk, such as light reflectance and joining of aggregates, especially under
and thermal conductivity. nonquiescent conditions, to form floes vis-
Numerous methods, the principles of which ible to the naked eye.
are based on detection of one or more of the Gelation. The aggregation of particles (e.g.,
above changes, have been developed to measure micelles or aggregates of micelles) to form
the rennet coagulation characteristics of milk or particulate strands whose particles undergo
the activity of rennets. Owing to the commercial limited touching and that eventually form a
importance of gel formation in milk as a means gel network.
of recovering milk fat and casein in the form of Elasticity. The ability of the gel to recover,
cheese curd, most methods measure gel forma- instantaneously, its original shape and di-
tion (also referred to as curd formation or rennet mensions after removal of an applied stress.
coagulability)—that is, the combined first and Viscoelastic materials, such as a renneted
second stages—but some specifically monitor milk gel, are elastic at relatively small
the hydrolysis of K-casein. Various terms or de- strains (e.g., 0.025, which is much less than
scriptors, some of which are used interchange- fracture strain). In this region of strain,
ably, are employed to describe the rennet coagu- known as the linear viscoelastic stress-
lation of milk. These are defined below: strain region, the strain is directly propor-
 Arbitrary units

Time after rennet addition, min

Figure 6-9 Release of casein macropeptides (•, Q) and changes in the viscosity (A, A) and curd firmness (•,
O) in skim milk unheated (closed symbols) or heated at 950C for 15 s (open symbols). Arrows indicate the time
at which cutting is initiated during cheese manufacture.

tional to the applied stress, and the material dynamic measurement of the viscous drag
(e.g., a section of a gel strand that bears the of gelling milk by suspending a pendulum
applied stress and is strained) recovers its in the milk and determining the tilt of the
original dimensions immediately upon re- pendulum over time using instruments such
moval of the stress. as the Thromboelastrograph, Formagraph,
• Viscosity. The physical property of a gel and Gelometer
given by the ratio between the stress and dynamic measurement of the ability of gel-
strain rate. ling milk to transmit a pressure by using, for
• Curd firmness, curd strength, or curd ten- example, a hydraulically operated oscillat-
sion. The stress required to cause a given ing diaphragm apparatus
strain or deformation. (Curd tension is a measurement of the apparent viscosity of
term frequently used to express the firm- gelled milk after a given time at a fixed
ness of formed gels.) shear rate (e.g., by using various types of
rotational viscometers) or, alternatively,
Types of measurement used to evaluate gel- measurement of the firmness of the gel us-
forming characteristics include ing various types of penetrometer
dynamic measurement of parameters such
• measurement of flocculation time under as viscosity, elastic shear modulus (G % loss
nonquiescent conditions (e.g., rennet co- modulus (G"), and phase angle (8) by ap-
agulation test) plying a low-amplitude oscillating strain or
 stress to the milk sample using, for ex- for example, 3O0C and the onset of visual coagu-
ample, controlled strain or controlled stress lation. If the coagulating activity of a rennet
rheometers preparation is to be determined, a "reference"
• dynamic measurement of some physical milk, such as low-heat milk powder reconsti-
properties of the gelling milk in the cheese tuted in 0.01% CaCl2 and perhaps adjusted to a
vat using special probes, including thermal certain pH (e.g., 6.5), should be used. A standard
conductivity (using a hot wire probe) and method has been published (International Dairy
reflectance of near infrared light (using a Federation [IDF], 1992), and a reference milk
near infrared diffuse reflectance probe) powder may be obtained from Instirut National
de Ia Recherche Agronomique, Poligny, France.
Some of the more commonly used laboratory If the coagulability of a particular milk is to be
and online methods for monitoring the rennet- determined, the pH may or may not be adjusted
induced gelation of milk are described below. to a standard value (e.g., 6.55) to reflect that
which is typical at setting (rennet addition) dur-
6.6.1 Measurement of the Primary Phase of ing cheese manufacture.
Rennet Coagulation The coagulation point may be determined by
placing the milk sample in a bottle or tube that is
The primary phase of rennet action may be rotated in a water-bath (Figure 6-10); the fluid
monitored by measuring the formation of either milk forms a film on the inside of the rotating
product, that is, para-K-casein or the CMP. Para- bottle or tube, but floes of protein form in the
K-casein may be measured by SDS-polyacryla- film upon coagulation. The rennet coagulation
mide gel electrophoresis (PAGE), which is slow time (RCT) provides a very good index of the
and cumbersome, or by ion-exchange high-per- gelation potential of milk; a low RCT usually
formance liquid chromatography (HPLC). The indicates potentially good gel formation and
CMP is soluble in TCA (2-12% depending on high gel strength after a given renneting time.
its carbohydrate content) and may be quantified The method is simple and enables the accurate
by the Kjeldahl method or more specifically by measurement of several samples simulta-
determining the concentration of N-acetylneura- neously. It has been used to accumulate much of
minic acid (Figure 6-3) or by RP-HPLC. The the extensive information reported in the scien-
activity of rennets can be determined easily us- tific literature on the effects of various process-
ing chromogenic peptide substrates, a number of ing parameters on the rennet coagulability of
which are available. The latter method is gener- milk. However, in contrast to cheese manufac-
ally used as a research tool to study rennet char- ture, where milk is renneted under quiescent
acteristics and/or the kinetics of the primary conditions to ensure gel formation, this method
phase of rennet coagulation. determines the time for coagulation (i.e., aggre-
gation and flocculation) of the paracasein under
6.6.2 Methods for Assessing Coagulation, agitation.
Gel Formation, and/or Curd Tension
Nondynamic Assessment of Viscosity
Measurement of Rennet Coagulation Curd Firmness Tests
Time The apparent viscosity and firmness of the co-
The simplest laboratory method for measur- agulum may be measured using various types of
ing the overall rennet coagulation process is to viscometers and penetrometers, respectively.
monitor the time between the addition of a mea- However, use of these instruments permits mea-
sured amount of diluted rennet to a sample of surement at only a single point in time, which is
milk in a temperature-controlled water-bath at, a serious limitation in kinetic studies, and it also
 Milk sample

Water-bath at 3O0C

Figure 6-10 Schematic of apparatus for visual determination of the rennet coagulation time of milk.

requires meticulous test conditions, since curd Samples of milk to be analyzed are placed in
strength increases with time after renneting. the cuvettes and tempered to the desired tem-
perature (typically 310C) in the heating block.
Dynamic Curd Firmness Tests Rennet is then added, and the cuvettes are re-
placed in the instrument so that a loop-shaped
Various instruments, involving different prin- pendulum is suspended in each sample. The
ciples, have been developed to monitor changes metal block is moved back and forth, creating a
continuously throughout the gelation process. "drag" on the pendulum in the milk. A flashing
These are discussed below. light is directed onto the mirror on the arm of
Formagraph. The most popular of the dy- each pendulum and reflected onto photosensi-
namic measuring instruments, although it is not tive paper, creating a mark. While the milk is
widely used, is the Formagraph (e.g., Type fluid, the viscosity is low and the drag on the
1170, Foss Electric, Denmark), a diagram of pendulum is slight. It scarcely moves from its
which is shown in Figure 6-11. The apparatus vertically suspended "zero-time" position, and
consists of hence a single straight line appears on the paper.
As the milk coagulates, its viscosity increases
• an electrically heated metal block
and the pendulum is dragged from its zero-time
• a sample rack with cavities (usually 10) into
position, resulting in bifurcation of the trace.
which sample cuvettes fit
The rate at and extent to which the arms of the
• a set of pendulums, each having an arm
trace diverge are indicators of the gel-forming
with an attached mirror
characteristics of the milk. A typical trace (see
• an optical source that beams a light ray onto
Figure 6-11) may be used to calculate the fol-
the mirror attached to each pendulum
lowing parameters:
• a chart recorder with photosensitive paper
onto which the reflected light from each • The rennet coagulation time (r) is the time
mirror is directed (in minutes) from rennet addition to the on-
 Photographic
chart paper
Lighl Hash

Mirror

Damper

MILK

(a) Oscillating heating block

(b)

Figure 6-11 (a) Schematic representation of the Formagraph apparatus for determining the rennet coagulation
of milk, (b) Typical formagram. The asterisk represents the point of rennet addition, r is the rennet coagulation
time, k2Q is the time required from coagulation for the arms of the formagram to bifurcate by 20 mm, and ^30 is the
extent of bifurcation 30 min after rennet addition (the approximate time at which the coagulum is cut in
cheesemaking).

set of gelation (i.e., point where the trace • ah which represents the curd firmness at
begins to fork). time t after rennet addition, is the trace
• &2o, essentially the inverse of the curd-firm- width (in millimeters) at time t.
ing rate, is the time from the onset of gela-
tion until a firmness corresponding to a Good gel-forming properties are character-
trace width of 20 mm is obtained. ized by a relatively rapid coagulation time (low r
value), high-curd firming rate (low k20 value), tion on the changes in curd strength over
and a high-curd firmness or strength after a time.
given renneting time (high ^30 value). Typical
values for these parameters for a pasteurized However, production of the instrument has
midlactation milk (3.3%, w/w, protein) renneted been discontinued.
under normal conditions (rennet dosage, « 1 6 Hydraulic ally oscillating diaphragm. In a
RU/L; pH, 6.55; temperature, 310C) are r = 5.5 hydraulically operated oscillating diaphragm ap-
min, A2O = 11 min, and ^30 = 48.5 mm. paratus, a sample of milk is placed between two
While the latter parameters have no precise diaphragms (Figure 6-12) and rennet is added.
rheological significance, the Formagraph One diaphragm (the transmitting diaphragm) is
method offers many advantages over the RCT made to vibrate through the cyclical application
test: of hydraulic pressure. When the milk is liquid,
the effect of the vibration is dissipated rapidly
• The method simulates gel formation during and does not affect the receiving diaphragm.
cheesemaking. When a gel is formed, the vibrations emitted by
• The results of the assay are less subjective, the transmitting diaphragm reach the receiving
being independent of operator judgment. diaphragm, causing it to vibrate. These vibra-
• The test parameters provide more informa- tions are detected and quantified by a suitable

(a)
Oscillating diaphragm

Sensor Pulsed pressure

Receiver Transmitter

(b) firmness suitable

for cutting
Amplitude of pulse

Coagulation time

Time from rennet addition

Figure 6-12 Schematic representation of a pressure transmission apparatus for measuring the rennet coagulation
time of milk and the strength of the resulting gel.
sensing device. An output generally similar to sures the torque, digitizes it, and relays it to
that of the Formagraph is obtained, from which the software of the interfacing computer
the coagulation time and a measure of gel
strength can be determined. Dynamic measurements are performed by ap-
Low-Amplitude Stress or Strain Rheometry. plying a low-amplitude oscillating shear stress
Since the 1980s, several controlled-strain rhe- (a) or shear strain (y), depending on the type of
ometers (e.g., Bohlin VOR, Bohlin Rheologi, rheometer, to the milk sample via oscillations of
Sweden; Rotovisco RV 100/CV 100, Haake the outer cylinder. The value of G or y is main-
Bucchler Instruments, United States) and con- tained sufficiently low in order to stay within the
trolled-stress rheometers (e.g., Bohlin CS, linear viscoelastic limits of the sample (i.e., the
Bohlin Rheologi, Sweden; Cari-med CSL2, TA region where G and y are directly proportional).
Instruments, United States; Rheometric Scien- Hence, the terms low-amplitude stress oscilla-
tific SR5, Rheometric Scientific Inc, United tion and low-amplitude strain oscillation. (Am-
States) have been used increasingly as research plitude refers to the maximum displacement of
tools for the continuous measurement of the vis- any point on the oscillating cup [and hence in the
coelastic properties of renneted milks as a func- milk sample or on the inner bob] from its mean
tion of time from rennet addition. A rheometer [or "zero"] position.) Under these conditions,
(Figure 6-13) essentially consists of: the gel strands of the gelling milk are strained to
a fixed displacement (within their elastic limit)
• a DC motor and recover instantaneously when the stress is
• a gear box removed. The stress required to achieve a fixed
• an electromagnetic clutch strain (e.g., displacement of a gel strand at a
• a direct drive position servo actuator, given position) increases as the gel strands be-
which, combined with the clutch, enables come more elastic and firm; hence, measure-
low-amplitude angular deflection (and ment of stress energy provides a measure of the
hence strain) to be transferred from the mo- gel strength.
tor via the gear box to the stalk of the outer When a controlled-strain rheometer is used,
cylinder of the sample cell the sample of renneted milk is subjected to an
• a sample cell, which can be of different ge- harmonic, low-amplitude shear strain, (y) of an-
ometries but typically consists of two co- gular frequency co:
axial cylinders—an outer cup (e.g., internal
diameter, 27.5 mm), into which the milk y =y 0 cos cot
sample is placed, and an inner bob (e.g., in-
ternal diameter, 25 mm); the strain on the
sample results in a stress that, when re- where: y0 is the shear strain amplitude, co is the
leased, creates a strain on or angular dis- angular frequency (i.e., 27U)), and cos cot is a
placement of the bob) term of the simple harmonic function. The shear
• a temperature sensor, which assists in accu- strain results in an oscillating shear stress (a) on
rate temperature control within the heating/ the milk that is of the same angular frequency
cooling chamber surrounding the outer cyl- but is out of phase by the angle 8:
inder of the sample cell and hence within
the sample G = G0 cos (cot + 8)
• a frictionless air bearing, which ensures that
the strain or torque created on the bob by where: G0 is the stress amplitude and G is the
the sample is transferred to a torsion bar phase angle between the shear stress and shear
• a torque-measuring transducer, which mea- strain oscillations, the magnitude of which de-
 water

Figure 6-13 Schematic representation of the main parts of a controlled-strain rheometer: DC motor (1), gear box
(2), electromagnetic clutch (3), direct drive position servo actuator (4), sample cell (5), temperature sensor (6),
frictionless air bearing (7), torsion bar (8), and torque-measuring transducer (9).

pends on the viscoelasticity (ratio of viscous to

elastic properties) of the gelling system. The
phase angle ranges from 0° for an elastic solid to Typical changes in the above rheological pa-
90° for a Newtonian liquid and has intermediate rameters from the time of rennet addition are
values for viscoelastic materials (Figure 6-6). presented in Figure 6-5. The onset of gelation is
The rheological parameters a, G', and G" are marked by sharp increases in G' and G" and a
computed continually from the measurement of decrease in 5, whose abrupt decline from about
stress energy over time. The first, a, is of course 80° in milk to about 10° marks the transition
the phase angle. G' is the storage or elastic shear from a viscoelastic material that is largely vis-
modulus, which represents elastically stored cous (i.e., milk) to a gel that is largely elastic in
stress energy and thus gel elasticity or firmness. character. Structural elements that impart elas-
It is given by the equation: ticity include the relatively weak, continuous
paracasein gel (formed from overlapping strands
of aggregated paracasein micelles), while those
that contribute to viscosity include aggregates
G", the viscous or loss modulus, which repre- and/or short dangling strands not yet connected
sents energy dissipated in flow, is given by the to the main network. Following the onset of ge-
equation: lation, the gel undergoes a progressive increase
in firmness over time. However, the rate of curd skill required for accurate measurement and the
firming increases initially to a maximum and fact that only one sample can be analyzed at any
then decreases (as reflected by the changes in one time. Moreover, instruments currently avail-
slope of the G'/time curve). The increase in curd able cannot be used for online measurement of
firmness ensues from the following changes: gel formation in the cheese vat.
Online Sensors for Predicting Curd Firmness
• the inclusion of free aggregates or dangling
and Cutting Time during Cheese Manufac-
strands into the network
ture. Following the onset of gelation, there is a
• the increase in contact surface area and
progressive increase in curd firmness, and the gel
number of attractions between neighboring
eventually attains an optimum firmness (e.g., 40
micelles in the gel (Walstra & van Vliet,
Pa after 40-50 min, depending on milk composi-
1986), which result in thicker strands with
tion and renneting conditions), which, for a given
higher stress-bearing capacity and thus
vat design, allows it to withstand the mechanical
more elasticity
action of the cutting knives in the cheese vat with-
G" and 8 are useful parameters for monitor- out shattering. Curd shattering corresponds to the
ing the viscoelastic changes in the gel during ag- fracturing of the individual curd particles (e.g., by
ing but are not directly related to gel strength and the cheese knife and by impact with other curd
are thus not discussed further. In contrast, G' is a particles and/or the knife and walls of the cheese
direct measure of curd firmness and is thus of vat), especially if the coagulum is soft at cutting.
significance in cheese manufacture. Various ob- Shattering results in an excessively large curd
jective rennet coagulation parameters pertinent particle surface area, through which fat is lost into
in cheesemaking may be derived from the G' the cheese whey, and it is conducive to the forma-
time curve, upon modelling, as described below tion of very small curd particles (i.e., curd fines
(Guinee, O'Callaghan, Mulholland, Pudja, & < 1 mm), which are also lost in the whey. Thus,
O'Brien, 1996): cutting at the optimum firmness and the rate of
curd firming are crucial for obtaining the correct
• gel time, defined as the time at which G'
particle size, minimizing the losses of fat and
reaches a threshold value, Gg, arbritarily set
fines in the whey, and maximizing cheese yield
at 0.2 Pa
(Figure 6-14).
• the firmness after a fixed renneting time
The firmness of the coagulum after a given
(e.g., 30 or 60 min, G'30 or G «,)
set-to-cut time is influenced by many factors,
• maximum curd-firming rate, defined as the
such as the concentrations of fat and casein in
maximum slope, *Smax, of the GY graph
the milk, the stage of lactation and the diet of the
• set-to-cut time (SCT), which is the time be-
cow, the milk pH, the starter activity (which in-
tween rennet addition and curd cutting at a
fluences pH at set and at cut time), and the
suitable firmness (e.g., 40Pa, SCT40Pa)
rennetcasein ratio. Variation in the firmness at
Low-strain or -stress rheometry gives param- cutting can, in turn, lead to variations in cheese
eters that are Theologically precise and accu- composition (especially moisture), yield, and
rately quantify the dynamic rheological changes quality. Hence, standardizing and optimizing the
that occur during renneting without altering the firmness at cutting is essential for consistently
process of gel formation. Hence, it accurately ensuring high cheese yields and good-quality
reflects the changes in curd firmness that occur cheese.
upon renneting milk in the cheese vat under Formerly, the time at which the coagulum was
quiescent conditions. However, limitations cut was usually determined by the cheese maker,
compared to other instruments, such as the who subjectively assessed firmness by various
Formagraph, include the high level of operator means, such as by cutting a small portion with a
 • data from laboratory analysis from a previ-
ous day's production showing the level of
fat and fines in cheese whey
Moisture-adjusted cheese yield,

• recovery of fat and casein in cheese

The above methods are not sufficiently pre-
kg/100 kg milk

cise to allow cutting at a constant gel firmness in

every vat. The limitations of these methods have
led to the development of in-vat gel-firmness
sensors, which dynamically monitor milk co-
agulation and, when incorporated into an inte-
grated system, activate the curd-knives to cut the
gel when it has attained the desired firmness
(strength). The mechanisms employed to date in
designing in-vat sensors to monitor the develop-
ment of curd firmness over time include moni-
toring related changes in
Set-to-cut time, % of control
• convective heat transfer from a probe (a
Figure 6-14 Effect of set-to-cut time (and hence "hot wire") to the surrounding milk, as in
firmness at cutting) and healing time on moisture-ad- hot wire probe sensors (Bellon, Quiblier,
justed Cheddar cheese yield. Healing times were O Durier, & Noel, 1988; Hori, 1985; LeFevre
min (O), 5 min (•), and 10 min (Q). & Richardson, 1990)
• turbidity (McMahon, Brown, & Ernstrom,
1984)
• diffuse reflectance of visible (e.g., A,, 660
hand knife and observing the "cleanness" of the nm), near infrared (e.g., K9 820 nm), or in-
cut and the clarity of the exuding whey. How- frared light (e.g., X, 950 nm), as in various
ever, in large modern factories, conditions are fiber-optic probes such as the CoAguLite
not conducive to testing gel firmness in cheese fiber-optic sensor and the Omron E3XA
vats from separate milk silos or from an indi- (Payne, 1995)
vidual day's milk, because of the large scale of • near-infrared light transmissions as in vari-
operations (frequently more than 106 L of milk ous infrared probes such as TxPro and
are processed per day) and the use of pre- GelographNT (O'Callaghan, O'Donnell, &
programmed vats with limited operator access. Payne, in press)
Hence, most of the cheesemaking operations are • absorption and attenuation of ultrasound
performed on the basis of a preset time schedule waves, or pulses, of different frequencies
rather than on the basis of objective criteria, such (e.g., > 0.5 MHz) passed through the milk
as gel firmness at cutting or pH at whey drain- (Benguigui, Emerery, Durand, & Busnel,
age. The criteria for determining gel cut times 1994; Gunasekaran & Ay, 1994).
are probably based on one or more of the fol-
The hot wire and fiber-optic probes are manu-
lowing:
factured commercially and are being used in
• data from previous production years that cheese factories as online curd-firmness sensors;
suggest that milk at different periods of the these are discussed briefly below. Sensors based
year requires different set-to-cut times (e.g., on measuring changes in tubidity or ultrasound
due to differences in milk composition) attenuation have not yet, to the authors' knowl-
 Next Page

edge, been successfully developed for use in the light-scattering properties of milk. Infrared
cheese plants. light is emitted from an LED and transmitted
The Hot Wire Probe. The hot wire probe has through one branch of a bifurcated cable contain-
achieved the furthest development toward com- ing optical fibers to the tip of a probe in contact
mercial application of all curd firmness sensors, with the renneted milk (Figures 6-16 through
with most of the research occurring at the Snow 6-18). Light reflected by both the fat globules and
Brand Dairy Products Company (Japan) and the casein micelles is detected by the optical fibers in
National Institute for Agronomic Research the other branch of the bifurcated cable and trans-
(France). Stoelting Inc. (Kiel, Wisconsin) has mitted to a photodetector. As the milk coagulates,
commercialized the Optiset II hot wire sensor more light is reflected (due to the aggregation of
(from Snow Brand), which is now operating in the paracasein micelles) and transmitted to the
several cheesemaking plants in the United photodetector, the output signal from which is
States. directly proportional to the amount of light re-
The principle of measurement is based on ceived. As in the case of the hot wire probe, the
changes in heat transfer from a hot wire to the peak in the first derivative of the output signal
milk. A thin platinum wire probe is immersed in corresponds to the onset of gelation, which is then
the milk. A constant current is passed through related to the cut time at a given firmness as deter-
the wire, generating heat, which is dissipated mined by laboratory instruments.
readily, by convection currents near the wire,
while the milk is liquid. As the milk coagulates, 6.7 FACTORS THAT AFFECT RENNET
its viscosity increases, and generated heat is no COAGULATION
longer readily dissipated. The temperature of the
wire increases, causing an increase in its resis- The strength of the resulting gel (curd tension)
tance. The resistance and temperature of the is as important as the coagulation time, if not
wire are dynamically measured by monitoring more so, especially from the point of view of
changes in voltage across the wire, giving a con- cheese yield. The gel assembly process is quite
tinuous output signal. slow (see Figure 6-6), and in the case of most
A typical output signal is shown in Figure cheese varieties a period roughly equal to the
6-15. The peak in the first derivative of the out- RCT is allowed from the onset of visual coagu-
put signal corresponds to the onset of gelation. lation for the gel to become sufficiently firm
The instrument does not detect a gel cutting prior to cutting. If the gel is too soft when cut, fat
time; the increase in gel firmness beyond the on- and casein losses in the whey will be high (see
set of gelation (i.e., the gel point) is not readily Bynum & Olson, 1982, for a description of the
detected by the hot wire, as it has a relatively influence of curd firmness on cheese yield and
small effect on heat dissipation (compared with for references on this subject). In general, there
the transition from a liquid to a gel). However, is an inverse relationship between RCT and curd
empirical equations have been developed to re- tension, which means any factor that reduces
late the gel point, as detected by the hot wire, to RCT increases curd tension and vice versa. The
cut times (at a particular firmness, e.g., 40 Pa) as effects of various compositional and environ-
determined by low-amplitude strain oscillation mental factors on the primary and secondary
rheometry, the Formagraph, or other laboratory phases of rennet coagulation are summarized in
methods. Table 6-2.
Diffuse Reflectance Fiber-Optic Probe. An
infrared diffuse reflectance probe, designed at the 6.7.1 Milk Protein Level
University of Kentucky, was installed in two
cheese plants in the United States in 1993. The The coagulation time of milk decreases mark-
principle of measurement is based on changes in edly with protein (and thus casein) content, in