The effect of heating processes on milk whey protein denaturation ...

Master Thesis

The effect of heating processes on milk whey protein

denaturation and rennet coagulation properties

Effekt af varmebehandling af mælk på valleprotein denaturering og koaguleringsegenskaber med chymosin

_______________Marije Akkerman__________

Department of Food Science, Aarhus University

Student number 20092385

Preface

The present master thesis, “The effect of heating processes on milk whey protein denaturation and rennet

coagulation properties” of 60 ECTS was part of the education “Molecular Nutrition and Food Technology”

at Aarhus University and was performed in the period October 2013 to October2014. The thesis was carried

out at Arla Strategic Innovation Centre, Brabrand, and department of Food science, Aarhus University. The

project was done with supervision from Lotte Bach Larsen from Aarhus University, and Mette Christensen

from Arla Foods.

Acknowledgement

First, I would like to thank my supervisors Lotte Bach Larsen and Mette Christensen for good and help-

ful supervision. Furthermore, I would like to thank Dairy technician Kent Matzen for assistance in the

pilot plant and especially Lene Buhelt Johansen, Per N. Andersen, Betina Hansen and Hanne Sønder-

gaard for experimental guidance in the laboratories at Arla Strategic Innovation Centre, Brabrand, and

Department of Food Science, Aarhus University, and for help with data analysis.

l would also like to thank Valentin Rauh for help with data analysis and good discussions of results

throughout the project and Eva Hansen for professional revision of the final report.

At last, special thanks to my family and friends for moral support.

Aarhus University, Department of Food Science, October 22. 2014

Abstract Whey protein denaturation as a cause of heat treatment has been investigated by various authors with the

study by Dannenberg and Kessler (1988) being the most acknowledged. Skim milk with various contents of

whey proteins and caseins were heat treated using three different heating processes, namely Plate Heat

Exchanger (PHE), Tubular Heat Exchanger (THE) and Direct Steam Injection (DSI) to provide further insight

into how heat treatment affects whey protein denaturation and rennet coagulation. The samples were

subjected to heating from 80 °C to 145 °C and holding times from 2s to300s. The milk samples were ana-

lysed on liquid chromatography (LC) to analyse the degree of whey protein denaturation, while rennet co-

agulation analysis were made with ReoRox rheometry. Heat induced aggregates were analysed at 1D- and

2D gel electrophoresis and size exclusion chromatography.

Heat treatment increased the degree of whey protein denaturation as the holding time increases for all

temperatures. Heat treatment using DSI gave the smallest increase in denaturation at all temperatures,

with a degree of denaturation of 40 % of β-Lactoglobulin B (β-Lg B) heating at 145 °C, while heating using

THE and PHE showed denaturation degrees above 95 % heating at 130 °C. Temperatures below 100 °C re-

sulted in higher degree of denaturation for heat treatment using THE compared to PHE, while at 130 °C

more than 90 % of β-Lg B was denatured for both methods. A reaction order of 1.5 was found for β-Lg and

1 for α-Lactalbumin, with reaction kinetics showing similar pattern compared to previous findings. The

variations are caused by different heating systems and heating profiles which have great impact on the

whey protein denaturation.

Rennet coagulation properties were impaired as the holding time increases for all temperatures. Heating

using DSI resulted in the least impairments. Heat treatment using PHE gave better rennet coagulation

properties when heating at temperatures below 100 °C, compared to THE, while THE reached Rennet co-

agulation time(RCT) within two hours at temperatures of 130 °C which was not observed for heat treat-

ment using PHE. The differences in heating systems using PHE and THE were primarily caused by variations

in formation of heat induced aggregates. THE was found to result in a higher degree of large whey protein

and κ-casein complexes, while a higher proportion of smaller whey protein aggregates were found when

heating using PHE.

Milk with reduced whey protein content showed improved rennet coagulation properties for all heating

methods compared to skim milk. Increasing the casein level and reducing the whey protein content gave

further improvements of rennet coagulation properties. Heating these at high temperatures, however,

resulted in impairments of rennet coagulation properties which can be caused by a change in mineral solu-

bility, casein dissociation and degradation of lactose which decrease the ability for rennet cleavage.

Resume Denaturering af valleproteiner pga. varmebehandling er blevet undersøgt af forskellige forskere, hvor de

mest anerkendte undersøgelser er foretaget af Dannenberg og Kessler (1988). For at få indsigt i forskellige

varmebehandlingers påvirkning af denaturering af valleprotein og koaguleringsegenskaberne med chymo-

sin, blev skummetmælk med forskellige indhold at valleproteiner og kaseiner udsat for tre typer varmebe-

handling: plade varmeveksler (Plate Heat Exchanger, PHE), rør varmeveksler (Tubular Heat Exchanger, THE)

og injektion af damp (Direct steam injection, DSI). Temperaturer fra 80 °C til145 °C og holdetider fra 2 til

300 sekunder blev anvendt. Mælkeprøverne blev analyseret på væskekromatografi (liquid chroma-

tographt,LC), for at undersøge denatureringen af valleproteiner, mens koagulerings-egenskaberne blev

analyseret med ReoRox rheometer. Proteinaggregater dannet ved varmebehandling blev analyseret ved 1D

og 2D gel-elektroforese og størrelseskromatografi. Varmebehandling øgede denatureringsgraden af valle-

protein når holdetiderne øgedes for alle temperaturer. Varmebehandling med DSI gav mindst denaturering

for alle undersøgte temperaturer, hvor 40 % af β-lactoglobulin B (β-Lg B) var denatureret ved en varmebe-

handling på 145 °C. Varmebehandling med THE og PHE resulterede i over 90 % denaturering af β-Lg B ved

varmebehandlinger på 130 °C. Temperaturer under 100 °C gav højere denaturering af valleproteiner ved

brug af THE i forhold til PHE, hvor der ved de højere temperaturer var denatureringsgrader over 90 % for

begge metoder. En reaktionsorden på 1,5 blev fundet for β-Lg og 1 for

α-Lactalbumin (α-La). Variationer i aktiveringsenergi, sammenlignet med tidligere studier skyldes forskelle i

varmebehandlingssystemer og varmeprofilen, hvilket har stor indflydelse på denatureringsgraden af valle-

proteinerne. Koaguleringsegenskaberne for koagulering med chymosin blev forværret når holdertiden blev

øget for alle undersøgte temperaturer, hvor varmebehandling med DSI gav færrest ændringer. Varmebe-

handling med PHE viste bedre koagulering i forhold til THE for temperaturer under 100 °C, hvorimod dette

var modsat ved varmebehandlinger ved 130 °C. Koaguleringspunktet blev opnået indenfor to timer ved

THE, men ikke ved PHE. Forskellene i denaturering og koagulering med chymosin mellem THE og PHE skyl-

les primært dannelsen af proteinaggregater som konsekvens af varmebehandlingen. THE havde en større

andel af store aggregater bestående af κ-kasein og valleproteiner, hvorimod PHE havde en større andel

mindre aggregater bestående af κ-kasein og valleproteiner.

Varmebehandling af mælk med reduceret indhold af valleproteiner gav forbedringer af koagulerings-

egenskaberne ved alle varmebehandlingsmetoder, i forhold til skummetmælk. Mælk med øget kasein ind-

hold og reduceret valleprotein indhold gav yderligere forbedringer. Varmebehandling af disse med ved høje

temperaturer gav fald i koaguleringsegenskaberne i forhold til kontrolmælken, hvilket kan skyldes en lavere

opløselighed af mineraler, at kaseiner forlader kaseinmicellen, og degradering af laktose pga. varme-

behandlingern og disse har alle negativ indflydelse på chymosins kløvning af κ-kasein.

Abbreviations 1DGE 1-dimensional gel electrophoresis

2DGE 2-dimensional gel electrophoresis

BSA Bovine Serum Albumin

CFR Curd firming rate

Cys Cysteine

DSI Direct Steam Injection

DTE Dithiothreitol

Ea Activation energy

FTIC Flouroscein-5-isothiocyanate

Gly Glycine

HCl Hydrochloric acid

HTST High Temperature, short-time heating

IG Immunoglobulin

LC Liquid chromatography

LTLT Low temperature, long time

LVR Linear viscoelastic region

MCI Micellar Casein Isolate milk, total protein adjusted

MCIc Micellar Casein Isolate milk, casein adjusted

MF Microfiltration

MS Mass spectrometry

OHTC Overall heat transfer coefficient

PHE Plate Heat Exchanger

pI Isoelectric point

RCT Rennet coagulation time

RP-HPLC Reverse Phase High Performance liquid chromatography

RT Retention time

SDS Sodium dodecyl sulphate

SDS-PAGE Sodium dodecyl sulphate polyacrylamide gel electrophoresis

SEC Size exclusion chromatography

TFA Triflouroacetic acid

THE Tubular Heat Exchanger

UHT Ultra High Temperature

α-La α-Lactalbumin

β-Lg β-Lactoglobulin

Contents 1 Objective ................................................................................................................................................ - 1 -

2 Background ............................................................................................................................................ - 2 -

2.1 Milk proteins .................................................................................................................................. - 2 -

2.1.1 Casein..................................................................................................................................... - 3 -

2.1.2 Whey proteins ....................................................................................................................... - 4 -

2.1.2.1 β-Lactoglobulin .................................................................................................................. - 5 -

2.1.2.2 α-Lactalbumin .................................................................................................................... - 6 -

2.2 Heat treatment of milk .................................................................................................................. - 7 -

2.2.1 Heat treatment methods ....................................................................................................... - 8 -

2.2.1.1 Plate Heat Exchanger ....................................................................................................... - 10 -

2.2.1.2 Tubular Heat Exchanger .................................................................................................. - 11 -

2.2.1.3 Direct Steam Injection ..................................................................................................... - 12 -

2.3 Heat induced denaturation of milk proteins ............................................................................... - 13 -

2.4 Protein separation methods ........................................................................................................ - 17 -

2.4.1 Liquid Chromatography ....................................................................................................... - 17 -

2.4.2 Gel electrophoresis .............................................................................................................. - 19 -

2.5 Protein identification ................................................................................................................... - 20 -

2.6 Rennet coagulation of milk proteins ........................................................................................... - 20 -

2.6.1 Measurement of coagulation properties ............................................................................ - 21 -

3 Materials and methods ....................................................................................................................... - 23 -

3.1 Milk types .................................................................................................................................... - 23 -

3.2 Heat treatments .......................................................................................................................... - 23 -

3.3 Protein analysis ............................................................................................................................ - 25 -

3.3.1 Analysis of total protein ...................................................................................................... - 25 -

3.3.2 Analysis of pH 4.5 soluble protein ....................................................................................... - 25 -

3.4 Measurement of rennet coagulation .......................................................................................... - 26 -

3.5 Analysis of protein structure ....................................................................................................... - 27 -

3.6 Analysis of protein aggregates .................................................................................................... - 27 -

3.6.1 1DGE .................................................................................................................................... - 28 -

3.6.2 2DGE .................................................................................................................................... - 28 -

3.6.3 Analysis of protein aggregate composition ......................................................................... - 29 -

3.7 Statistical analysis ........................................................................................................................ - 30 -

4 Results ................................................................................................................................................. - 31 -

4.1 Variation in milk protein composition ......................................................................................... - 31 -

4.2 Effect of indirect heating on skim milk ........................................................................................ - 33 -

4.3 Effect of indirect and direct heating systems on heat treatment of skim milk ........................... - 36 -

4.4 Effect of whey protein denaturation on rennet coagulation ...................................................... - 38 -

4.5 Effect of heat treatment of MCI milk samples on rennet coagulation ........................................ - 39 -

4.6 Effect of milk type on rennet coagulation ................................................................................... - 41 -

4.6.1 Formation of protein network ............................................................................................. - 43 -

4.7 Heat induced protein aggregation............................................................................................... - 45 -

4.7.1 1-DGE ................................................................................................................................... - 45 -

4.7.2 2DGE .................................................................................................................................... - 47 -

4.7.3 Size exclusion chromatography ........................................................................................... - 48 -

4.7.3.1 Identification of aggregates ............................................................................................. - 49 -

4.8 Kinetics of denaturation of whey proteins .................................................................................. - 52 -

5 Discussion ............................................................................................................................................ - 56 -

6 Conclusion ........................................................................................................................................... - 64 -

7 Perspectives ......................................................................................................................................... - 66 -

7.1 Future research ........................................................................................................................... - 66 -

8 References ........................................................................................................................................... - 67 -

Appendix 1 ................................................................................................................................................... - 75 -

Appendix 2 ................................................................................................................................................... - 77 -

- 1 -

1 Objective For decades, the whey fraction was considered a waste product from the cheese production but today it is

used as functional ingredient in a wide variety of food products. Some important functions of whey pro-

teins in foods are water binding capacity, emulsification, foaming, whipping and gelation properties

(Singh, 2011). To utilize these functionalities of the whey proteins optimal it is crucial to know the physical

and chemical properties of each of the components in the product, as well as the interactions between

the different components. The physical properties of whey proteins have been investigated for many

years, both in purified form and in the milk matrix. One of the most dominant features is the heat denatu-

ration of the proteins which has great impact on the usage and functionality of the proteins. The most

acknowledged studies on the heat induced whey protein denaturation in skim milk are done by Dannen-

berg and Kessler (1988). Their results have been supported by various authors since, but given that the

industry uses several types of heating protocols and new analysing methods, it is of interest to be able to

verify these results and to be able to use this knowledge in the manufacturing of dairy products.

The objective of this master thesis is to:

Study the effect of three heating processes on milk whey protein denaturation and rennet coagu-

lation properties on milk with various whey protein and casein content.

Study how whey proteins influence the formation of heat induced protein aggregations.

The hypotheses are:

Increase in heating temperature and holding time will lead to an increase in whey protein denatu-

ration which will have a negative effect on the coagulation properties of the milk in cheese pro-

duction.

Direct steam injection is the most gentle heat treatment, leading to least whey protein denatura-

tion, while heat treatment using tubular heat exchanger and plate heat exchanger results in

higher degrees of whey protein denaturation.

Removal of whey proteins from milk will improve coagulation properties compared to skimmed

milk.

Increasing the heating temperature and holding time will lead to formation of large protein ag-

gregates due to unfolding and denaturation of whey proteins.

- 2 -

To test the hypotheses, skimmed milk with various whey protein and casein content were heated treated

using three different heating methods, namely Plate Heat Exchanger, Tubular Heat Exchanger and Direct

Steam Injection, at varying temperatures and holding times and combinations to investigate rennet co-

agulation properties and protein composition.

2 Background

Milk and its individual compounds are of major importance in relation to human nutrition throughout

millennia. The oldest evidence for consumption of milk from domesticated animals dates back to 4000 BC

(Schmid, 2003). Bovine milk is the main type of milk used for human consumption. Various other species

such as goats, sheep and horses also yield milk that is used in the food industry but these play a minor

role (Belitz et al., 2004). Throughout this report, the term milk refers to bovine milk. Milk composition

consists of approximately 87 % water, 3.5 % protein, 4.4 % fat, 4.5 % lactose, 0.7 % minerals with a pH of

6.5-6.75. The composition of milk is not completely constant. There are variations in relation to breed,

genetic variation, lactation stage, health, feed composition, climate and season (Heck et al., 2009).

2.1 Milk proteins

Bovine milk contains between 2.5-3.7 % proteins and variations are mostly caused by breed and genetic

variation. More than 200 types of protein have been found in milk but only few groups of proteins are

present in large quantities. The two major groups of proteins are caseins, representing 80 % of the protein

in milk, and whey proteins which represent 20% of the milk protein. More than 60 different enzymes are

found in milk, but they represent less than 1 % of the total protein content in milk (Farrell Jr. et al., 2004).

The casein and whey protein composition is shown in Table 1 and key points from this table will be used in

the following sections.

- 3 -

2.1.1 Casein

Casein is the protein fraction in milk which is characterized by precipitating at pH 4.6 at a temperature of

20 °C. The caseins are divided into four major components, -, -, β- and κ-casein which are generally

distributed in the proportions 40 %, 10 %, 40 %, 10 %, respectively, of the total casein (Dalgleish, 1993).

They all contain high amounts of ester bound phosphate, proline residues and contain no or very few di-

sulphide bonds, which is shown in Table 1. This results in very little secondary structure and random coil-

ing of the primary chain. The structure makes the casein flexible, and the lack of secondary structure and

thus an open structure exposes the hydrophobic parts of the amino acid chain which gives a higher sur-

face hydrophobicity (Fox and Kelly, 2006). This open structure also makes caseins more vulnerable to pro-

teolysis by various enzymes, especially pepsin, as they have easy access to the protein backbone and al-

lowing them to cleave the protein into various peptides(Swaisgood, 2003).

Caseins belong to the group of phosphoproteins which have phosphoric acid groups bound to the amino

acid backbone. The phosphate groups are located in clusters bound to Serine resides. The phosphate

groups have, because of their negative charge, the ability to bind ions, especially Ca2+. The binding of ions

is essential for the transport of calcium and phosphate to the neonate. Caseins are sensitive to change in

the calcium level in milk, and increasing calcium level can induce precipitation of these (Considine and

Flanagan, 2008). -casein is the most calcium sensitive type of casein while κ-casein with the lowest

amount of phosphate groups is not affected by the calcium concentrations normally present in milk (Gin-

ger and Grigor, 1999).

Table 1. Distribution of the major proteins in milk and their main characteristics. Modified according to (DaLgleish, 2014;

Farrell Jr. et al., 2004).

Protein Amino acid residues

Molecular mass (kDa)

Proline residues

Phos-phoseryl residues

Di-sulphid- residues

S-S linkages

Approx. amount in milk (g/L)

% of protein

Total casein 26 79.5 αS1-Casein 199 23.164 17 8-9 0 0 10 30.6 αS2-Casein 207 25.388 10 10-13 2 1 2.6 8.0 β-Casein 209 23.983 35 5 0 0 9.3 28.4 κ-Casein 169 19.038 20 1-2 2 1-2 3.3 10.1 Total whey 6.3 19.3 β-Lactoglobulin 162 18.277 8 0 5 2 3.2 9.8 α-Lactalbumin 123 14.175 2 0 8 4 1.2 3.7 Bovine Serum Albumin

582 66.463 0 35 17 0.4 1.2

Immuno-globulin

103-150 0.7 2.1

- 4 -

Caseins have a tendency to associate with each other in casein micelles through hydrogen bonds to stabi-

lize the structures. Around 95 % of the native casein is bound in casein micelles. 94 % of the casein mi-

celles are protein and the remaining 6 % are referred to as colloidal calcium phosphate consisting of cal-

cium, phosphate and small amounts of magnesium, citrate and other trace elements (Walstra, 1990).

Casein micelles vary in size with an average diameter of 200 nm. There are several models trying to ex-

plain the composition of the casein micelle structure, e.g. the nanocluster model by Holt and Horne (Far-

rell Jr. et al., 2006) and the sub-micelle model by Walstra (1999) which are shown in Figure 1. Until now,

none of the models are completely verified. It is surely known that the most hydrophobic and most cal-

cium sensitive caseins, α- and β-casein, are primarily found in the core of the casein micelle while the

more hydrophilic and calcium insensitive κ-casein is situated at the surface with its polar C-terminal out-

side the micelle core, making the casein micelle soluble in solution.

2.1.2 Whey proteins

Whey proteins are the proteins in milk, which are soluble in solution at pH 4.6. Whey can be separated

from the casein faction of milk during coagulation processes of the casein, such as rennet or acid coagu-

lated cheese (Singh and Havea, 2003).

The whey proteins contribute about 20 % of the total protein content in milk. The heterogeneity is large

among the whey proteins and they share only few characteristics such as they are all soluble at pH 4..6

and at a temperature of20 °C, at which the caseins are precipitated (Farrell Jr. et al., 2004). The four major

whey proteins represent 90 % of all whey proteins. These are β-Lactoglobulin (β-Lg), α-Lactalbumin (α-La),

Figure 1. Two models of the casein micelle structure. A: Nanocluster model with tread-like casein monomers and calcium

phosphate nanoclusters (Farrell Jr. et al., 2006). B: casein sub-micelle structure with caseins bound in small sub-micelles in

corporation with calcium phosphate which are bound to each other (Walstra, 1999).

- 5 -

Bovine Blood Serum Albumin (BSA) and Immunoglobulins (Ig’s). The remaining 10 % are different proteins

such as lactoperoxidase, serum transferrin, enzymes and milk fat globular membrane proteins (Fox and

Kelly, 2006).The main characteristics for the major whey protein are shown in Table 1.

The whey proteins are highly structured proteins with stable secondary and tertiary structures. The major

forces responsible for maintaining their globular structure are disulphide bonds, hydrophobic interactions,

hydrogen bonding, ion-pair interactions and van der Waal’s interactions (Singh and Havea, 2003).

The native composition of the whey proteins makes them highly soluble in the milk over a broad range of

pH. This is due to the large proportion of hydrophilic residues on the surface of the globular structure and

the large amount of disulphide bonds (Dissanayake and Vasiljevic, 2009). The globular structure also

makes the proteins resistant to proteolysis. They function as carrier proteins for different molecules,

partly responsible to maintain the osmotic balance and immune responses, but not all functions are yet

fully understood (de Wit, 1998).

Pressure has also been shown to have an impact on whey protein stability. Pressure changes the globular

structure which can lead to denaturation (Mozhaev et al., 1996). β-Lg is the most pressure sensitive and

can withhold pressure up to 200 Mpa, while the other whey proteins are stable for pressures up to 400

MPa (Considine et al., 2007).

In this project, the focus will be on the two main whey proteins, namely β-Lg and α-La, which contributes

to 70 % of the total whey protein content and will be discussed below.

2.1.2.1 β-Lactoglobulin

β-Lg is the major whey protein in milk, representing around 50 % of the whey proteins, and 12 % of the

total protein content. β-Lg consists of 162 residues per monomer with a molecular mass of 18,3 kDa.

These main characteristics are also shown in table 1.The structure of β-Lg is shown in figure 2. Five of the

residues which are highlighted in figure 2, are cysteine (Cys). Cys forms intermolecular disulphide bonds at

Cys66-Cys160 and cys106-Cys119, whereas Cys121, which does not form intermolecular disulphide bonds,

is buried within the native structure. β-Lg is a highly structured protein with anti-parallel β-sheets formed

by nine β-strands and one α-helix (Kontopidis et al., 2004).

- 6 -

13 different genetic variants of β-Lg are identified, but the most common variants in Western dairy cattle

are variant A and B which differ at position 64 (Glycine Aspartic acid) and 118 (Alanine Valine) in the

amino acid composition. Variant A has shown to yield smaller aggregates than variant B which can be

caused by variant B being less heat stable than variant A (Jakob and Puhan, 1992).

β-Lg belongs to the protein family of lipocalins, which act as transport proteins for small hydrophobic

molecules, such as retinols (vitamin A) and lipids. The biological function of β-Lg in milk is not fully under-

stood, but it is believed to have other functions than delivering large quantities of amino acids to the neo-

nate. It is known that β-Lg is involved in the retinol transport from the mother to the neonate as it is resis-

tant to proteolysis and acids in the stomach (Miranda and Pelissier, 1983). Palmitate has also shown to

bind to β-Lg and as vitamin A and it derivates are fat soluble, it can also be reasonable to believe that β-Lg

is related to fatty acid metabolism in milk (Pérez and Calvo, 1995).

Under physiological conditions, β-Lg is found as a dimer of two β-Lg proteins in equilibrium with the

monomer and can rapidly be transformed to native monomers (Croguennec et al., 2004; Papiz et al.,

1986). The native monomer-dimer equilibrium shifts towards the monomer state when the β-Lg concen-

tration is low, the ionic strength is low or pH is increased above 7 (Reithel and Kelly, 1971).

2.1.2.2 α-Lactalbumin

α-La represents 20 % of the whey protein in bovine milk. It is a small protein with 123 residues and a mo-

lecular weight of 14 kDa. These main characteristics are also shown in table 1. 8 of these residues are cys-

teine which makes four intermolecular disulphide bonds at Cys6-Cys-120, Cys28-Cys111, Cys61-Cys77 and

Cys 73-Cys91(Belitz et al., 2004)..There are several different genetic variants of α-La, but only variant B is

present in western dairy cattle.

Figure 2. Structure of β-Lg with the 5 cysteine residues

and disulphide bonds highlighted.

- 7 -

α-La is the regulatory protein of the lactose synthase enzyme system and the concentration of lactose in

milk is directly related to the concentration of α-La (Caffin et al., 1985). α-La enhances the binding of glu-

cose to galactosyl transferase which is the limiting step in the lactose synthesis. Lactose is a very important

component in milk and is responsible for maintaining 50 % of the osmotic pressure in milk which is there-

fore also an indirect physiological role of α-La (Brew, 2003). α-La is a metallo protein and can bind one Ca2+

molecule per molecule. α-La has a strong calcium binding site which is important for the stability of α-La

during heating as calcium increases the stability of α-La.

2.2 Heat treatment of milk

Heat treatment is one of the major processing steps of milk. Regardless of its final use, the majority of

milk is heat treated at least once. The most used heat treatments of milk are pasteurization and steriliza-

tion. Heat treatment of milk is preformed to limit bacterial load and enzyme activity to secure the safety

of the dairy product for human consumption and for extending the shelf life of the final product. Heat

treatment also gives rise to different chemical changes in milk, such as non-enzymatic browning reactions

involving lactose and especially lysine residues in the protein. This gives rise to off-flavours, change in

colour and loss of available lysine which has high nutritive value (Singh and Waungana, 2001). Further-

more, when heating above 100 °C can cause decrease in pH, which is caused by formation of organic acids

from lactose degradation and precipitation of calcium phosphate (Martinez-Castro et al., 1986).

Pasteurization is a heat treatment which aims to reduce the number of pathogens to such an extent that

it is does not constitute any health hazard. There are different pasteurization methods but one of the

most used is high-temperature, short-time heating (HTST) where the milk is heated to 72-80 °C for

15-30 s. Pasteurization at low temperature, long time (LTLT) at 63 °C for 30 min is still used, but not to the

same extent as HTST due to longer processing time and it has been shown that HTST gives less chemical

changes than LTLT heat treatment (Lewis and Deeth, 2008). Ultra high temperature (UHT) heating is a

sterilization process of milk which is used to destroy all microorganisms and spores in the milk and many

enzymes are also inactivated. This is done by heating temperatures of 135-150 °C for 1-10 seconds (Fox

and Kelly, 2006).

The design of heat treatment and combination of temperature and heating time depends on the desired

approach with least undesirable changes. High heat treatment and heating time gives most significant

changes. Heat treatment causes whey protein denaturation which is an irreversible process. The mineral

balance also changes during heat treatment. Calcium and phosphate becomes more insoluble and binds

to the casein micellar structure. This is reversible for temperatures below 100 °C, while severe heating can

- 8 -

cause hydrolysis of phosphorserine of caseins and calcium phosphate can precipitate out of solution

which are irreversible processes (Gaucheron, 2011).

2.2.1 Heat treatment methods

Various systems and technologies associated with different time–temperature profiles are developed to

obtain pasteurized and UHT milk, and the effect on the nutritional and sensory quality of commercial milk

samples may vary substantially depending on the process.

All heating processes of milk is divided in three steps; a preheating period, a heating period and a cooling

period. In all steps heat must be transferred from one material to another. The heat exchange between two

materials can be classified according to various parameters such as construction of the heat exchangers,

flow arrangements, and phase of the process fluid (Thulukkanam, 2013). Here, the heat exchange is classi-

fied according to the heat transferring process. The preheating and cooling steps in all heating processes

are always done indirectly, but the heating to the desired temperature can be done either directly or indi-

rectly.

In the indirect heating system, the heat transfer to milk is done via a medium, separating milk and the heat-

ing fluid which is normally hot water or steam. The temperature difference between the two fluids or

steam facilitates the heat transfer between the fluids. This type of heat exchange requires large surface

areas for heat exchange to get a sufficient heat transfer and an equal heat distribution in milk. In the direct

Figure 3. Heating profiles for indirect and direct heating systems. Modified

from Deeth and Datta (2002).

- 9 -

methods, the heating medium is steam which is in direct contact with the milk and as the steam conden-

sates the heat is transferred to the milk (Lewis and Deeth, 2008).

Indirect continuous heating systems are most commonly used. One of the main advantages of the indirect

systems is the regeneration of heat which is possible by using counter-current heat exchange. The fluids

flow in reverse directions of each other to minimize the temperature gradient between the two fluids and

thereby reducing the fouling and minimize the energy used in the heating and cooling processes. In co-

current heat exchange the fluids flow parallel to each other in the same direction. This gives a large tem-

perature gradient when starting the heating process which can cause thermal stress in the exchanger mate-

rial and gives rise to larger loss of heat as it cannot be utilized to the same extent as for counter current

heat exchange (Visser and Jeurnink, 1997). Fouling can be a major issue in indirect heating systems. Fouling

can be protein denaturation, which is mostly seen at temperatures below 110 °C, while mineral fouling is

seen at higher temperatures. Fouling is seen as deposits layer on the surface of the heat exchanger. Fouling

can cause decrease in milk flow, increase in pressure through the system and reduces the heat rate

transfer(Boxler et al., 2014).

The indirect heating systems can be used in a temperature range from 0 °C up to 150°C while the direct

heating system is only used for high temperature treatments due to milk being mixed with steam under

pressure which forces the temperature to rise above 100°C ( Bansal and Chen, 2006). The direct heating

system has a lower heat load than indirect heating when comparing these at equivalent bacterial effective-

ness due to fast heating and cooling rates. This is shown in figure 3 at which the temperature-time profiles

for UHT treatment using indirect and direct heating systems is displayed (Deeth and Datta, 2002). The dif-

ference in heating time and temperature can have a great impact on the denaturation of whey proteins.

Preheating is always done indirectly by heating milk up to 60-95 °C. This step is also called protein stabiliza-

tion step as the preheating temperature is often held from 15 seconds to a few minutes to denature the

whey proteins and thereby reduce their ability to foul the heating system at higher temperatures in indirect

systems. The holding period at preheat temperature is normally not used for direct systems as fouling is not

a general problem due to the fast heating rate (Lewis and Deeth, 2008). The preheating step is also benefi-

cial to minimize the energy used for heating, as it requires large amounts of energy to reach desired tem-

peratures above 100 °C in one step.

The holding time at heating temperatures is easily changed for the heating systems. When the fluid has

reached the desired temperature, the fluid is held at the desired temperature for a fixed time. The holding

time can be varied from a few seconds up to several minutes (Edgerton et al., 1970). After passing the hold-

ing tubes the fluid goes into indirect system for cooling for both direct and indirect heating systems.

- 10 -

2.2.1.1 Plate Heat Exchanger

Plate heat exchangers (PHE) are widely used in the dairy industry for heat treatment of milk. It is an indi-

rect heating system consisting of metal plates placed closely together in a frame. The formation and num-

ber of plates varies according to purpose of the heating. Gasketed plates has a maximum temperature

around 150 °C and a maximum operating pressure of 20.4 bar while welded plates can withstand tem-

peratures from -50 °C to 350 °C and a pressure from vacuum to 40 bar (Abu-Khader, 2012).

Each plate has orifices in the corners going through the whole plate frame. If the orifices are closed the

fluid is forced through small channels which are made by the space between the plates. Wide flow chan-

nels give lower heating rates and a more non-uniform heating. The number of plates and channels vary

according to the aim of the heating process, but the number of channels and plates are the same for both

hot and cold fluids (Gut and Pinto, 2003). Hot and cold fluid flow counter current on each side of the

plates and exchange heat through the plates. This is shown in figure 4. The surface of the plates is often

corrugated in various degrees which causes turbulence in the milk while passing through the channels and

thereby increasing the heat transfer and giving a more uniform heating of the milk. The corrugating de-

pends on the fluid and heat transfer rate. Low corrugation and chevron angles on the surface of the plates

decreases the heat transfer (Abu-Khader, 2012).

The flow pattern can be parallel where you have a single pass between the plates. Here the first and last

plates do not transfer heat. The flow pattern can also be a multi pass flow distribution where the fluid

passes through multiple channels on its way through the plate heat exchanger. This is shown in figure 4.

Figure 4. Schematic representation of plate heat exchanger with counter

current heat exchange. (Karami, 2014).

- 11 -

The flow through the channels is arranged in series. The multi pass flow distribution gives the opportunity

of longer and slower heat transition (Gutierrez et al., 2011).

PHE is a compact system with a relative large surface area for heat transfer compared to the volume of

the system. This gives a high overall heat transfer coefficient (OHTC) and minimum loss of heat through-

out the system (Edmond, 2001). On the other hand, the large surface area makes the system more prone

to fouling as the plates are heated to a higher temperature than the desired temperature for the milk and

deposits are building up on the surface of the plates. This requires more often cleaning especially when

more corrugated plates are used (Deeth and Datta, 2002; Visser and Jeurnink, 1997).

2.2.1.2 Tubular Heat Exchanger

Tubular heat exchangers (THE) are indirect heat exchangers, build up by one or more circular tubes sur-

rounded by a larger pipe. Milk flows inside the tubes while water or steam is flowing outside the tubes.

Heat transfer is done between the two fluids across the tube material. The heat exchange is often oper-

ated in counter–current mode. The tubes can have different shapes, but they are relative smooth and

there are few seals which gives less resistance and a smooth passage through the tubes and this system

can tolerate high pressure compared to PHE (Deeth and Datta, 2002). The overall heat transfer rate is

lower for THE compared with PHE and direct heating systems due to a lower surface area for heat ex-

change, a larger diameter of the fluid flowing through the system and less turbulence within the tubes

(Edmond, 2001). The slower heating and cooling rates gives longer transit time through the heat ex-

changer which makes the fluid more prone to chemical changes and fouling (Lewis and Deeth, 2008).

Figure 5. Schematic overview of a tube in tube heat exchanger with counter-

current heat exchange. (Thermaline Inc., 2014).

- 12 -

THE is a considerable flexible system as the core geometry is quite easily changed according to tube di-

ameter, tube length and arrangement. This also makes the system more suitable for fluids with a higher

viscosity even though the heat transfer to the core of the tube decreases (Ditchfield et al., 2007).

The three main tubular heat exchange systems are tube in tube, shell-and-tube and spiral tube heat ex-

changers. Tube in tube, which is shown in figure 5, consists of one or more tubes surrounded by a pipe. It

is the simplest of the tubular heating systems but is not very efficient in heat transferring and it occupies

large space. Shell-and-tube is an advanced version of double pipe heating system. Shell-and-tube is the

most common used heat exchanger in the industry. It consists of multiple tubes in which the product is

flowing, surrounded by a shell with the heating or cooling material. The shell and tubes can be twisted or

baffles can be placed inside the shell to direct the flow of the fluid outside the tubes and give a greater

turbulence in the fluid to get a more efficient and equally distributed heat transfer (Shah and Sekulić,

2007).

2.2.1.3 Direct Steam Injection

Direct steam injection (DSI) and direct steam infusion are the two main direct heating methods used in

the dairy industry. Direct steam infusion is also called ‘product into steam heating system’ and is the op-

posite of DSI which is called ‘steam into product’ (Campell, 2013). Further on, DSI will be described.

DSI is a good method when a very accurate temperature is needed as no heat is lost in transferring mate-

rial during transfer of energy from steam to milk. The heat transfer is also much more rapid as steam is

mixed with the milk and gives a more uniform heating. DSI can only be used for temperature above 100 °C

as the water has to be steam which is capable of reaching temperatures above 150 °C (Campell, 2013).

Figure 6. Schematic overview of a direct steam injection system with THE as

preheating and cooling system. (Powerpoint international, 2014).

- 13 -

Figure 6 shows a schematic overview of a DSI system. Milk is first preheated to temperatures from 75 °C-

85 °C by using an indirect heat exchanger before the steam can be injected (Schroyer, 1997). This preheat-

ing step is used to minimize the energy use, as it requires large amount of energy to heat the milk to tem-

peratures above 100 °C in one step. Furthermore, heating in one step is found to apply great stress to the

milk (Deeth and Datta, 2002). During stream injection, the steam is injected, under pressure, into the milk

as small bubbles which collapse rapidly as they are mixed with the milk and in this process transfer heat.

The milk is heated but it also becomes diluted with water as the steam condensates due to release of

energy. After heating, the water added as steam is removed from the milk in a vacuum chamber which

also flash cools the milk to temperatures near preheating temperature at the same time. To ensure that

the fluid is not diluted or concentrated during the heat treatment, the total solid content before and after

heating is monitored. An indirect system is used for further cooling to the desired temperature (Lewis and

Deeth, 2008). Limitations of the DSI system are that it cannot be used if the product is sensitive to mixing

with steam and the following condensation. Furthermore, the water used to generate steam must be free

of organic constituents to avoid an incorporation of these in the milk when it is mixed with steam and

thereby contaminate the fluid.

DSI is not very prone to fouling due to shorter total heating time and less surface area is available to initi-

ate the fouling process. This short heating time compared to indirect methods also gives less chemical

changes and protein denaturation(Deeth and Datta, 2002; Lorenzen et al., 2011; Oldfield et al., 1998a).

2.3 Heat induced denaturation of milk proteins

Caseins are very heat resistant due to their loose structure and it is generally accepted that they can with-

stand heating at 140 °C for 15-20 min as the random coiling of the primary chain is generally hard to de-

stroy compared to secondary and tertiary structures. To some extent both dephosphorylation and hydroly-

sis of the caseins has been found in heat treated milk (Belitz et al., 2004; Farrell Jr. et al., 2004; Fox, 1980).

Heat treatment above 100 °C gives a decrease in micellar size due to increase in colloidal phosphate and

dissociation of κ-casein from the micelle surface (Singh and Waungana, 2001).

The globular structure of whey protein makes them heat labile. Heat treatment of the whey proteins above

60 °C, results in unfolding of the globular structure and the proteins thereby denature.

The denaturation of whey proteins is generally considered involving two steps. The first step is an unfolding

of the native globular structure, which leads to exposure of hydrophobic residues and disulphide bonds. If

the heat treatment is minimal, the unfolded protein can refold into native structure. At high temperatures,

the unfolded proteins will form new hydrophobic interactions and disulphide bridges which can result in

- 14 -

refolding the protein, but this is often disordered and gives rise to a random structure. The increase in the

reactivity of the unfolded protein heated at high temperatures can also lead to the second step of the de-

naturation process. The unfolded whey proteins can form aggregates with other molecules, mostly through

disulphide bonding and covalent bonds (Singh and Latham, 1993). Immunoglobulin’s and BSA are the least

stable whey proteins, β-Lg is intermediate and α-La is the most resistant protein to heat denaturation.

These differences in extent of heat denaturation are caused by the differences in structure and strength of

intermolecular bonds (Anema, 2008; Corredig and DaLgleish, 1996).

The dimer of β-Lg dissociates between 30 and 55 °C, but these changes are reversible and the monomers

can rebound by cooling if the temperature does not exceed 60 °C. When heating to temperatures above

60-70°C, the tertiary- and also partly secondary structure of the monomer starts to unfold, leading to expo-

sure of the free thiol group (Cys121) and hydrophobic parts of the residues chain, resulting in a reactive

monomer. (Iametti et al., 1995and Iametti et al. 1996). The formation of these monomers is irreversible

and they cannot refold to native state. Instead there will be formed non-native monomers, which can form

aggregates with other monomers but also aggregates with other types of proteins can be formed (Tolkach

and Kulozik, 2007).

α-La is the least heat resistant whey protein with a denaturation temperature around 62 °C, but the unfold-

ing at this temperature is reversible. It does not form aggregates or modified monomers at heating tem-

perature below 80 °C, at neutral pH (pH 6.6-6.8). This is due to α-La having no free thiol groups which can

change the reactivity of α-La (Eigel et al., 1984). α-La is capable of refolding to its native state in presence of

calcium if the disulphide bonds are still intact (Brew, 2003). The binding of calcium is very pH dependent

and calcium dissociates from the α-La binding site at pH below 5 which makes α-La lose the ability to refold

to its native structure after heat treatment. Severe heating conditions with temperatures above 100 °C for

several minutes disrupt the disulphide bonds and formation disulphide linked aggregates of denatured α-La

occurs (Singh and Havea, 2003).

The main aggregates formed as a consequence of heat treatment of milk, are complexes formed by aggre-

gation of denatured whey proteins and complexes between β-Lg and κ-casein on the surface of the casein

micelles via disulphide bonds and hydrophobic interactions. At temperatures below 70 °C the interaction is

mostly caused by hydrophobic interactions while at higher temperatures it is mostly caused by disulphide

bonds (Corredig and DaLgleish, 1996; O’Connell and Fox, 2011).The κ-casein and β-Lg interactions are most

pronounced when the κ-casein is placed on the surface of casein micelles as the association between β-Lg

and κ-casein is less favourable when κ-casein is dissolved in serum. This can be caused by κ-casein is pre-

sent in a more compact structure when dissolved in serum whereas placed on the surface of casein micelles

- 15 -

the structure of κ-casein is more loose (Donato et al., 2007). The formation of these complexes may be

altered by a slow heating rate or heating for a long time at lower temperatures. This gives longer time for

the β-Lg to unfold and associate with the casein micelles. In contrast, a rapid heating rate to the required

temperature gives a shorter time for unfolding and it is more likely that β-Lg refolds in a non-native struc-

ture or forms aggregates with other unfolded monomers instead of associating with κ-casein (Oldfield et

al., 1998b).

α-La does not associate with the casein micelles on its own like β-Lg; it has to form complexes with β-Lg

which then associates with the casein micelle and it requires a prolonged heating period to start associating

with the casein micelle (Oldfield et al., 1998b; Oldfield et al., 2000).

The rate of denaturation is mainly controlled by heating temperature, heating time and pH but also protein

concentration and ionic strength have been proved to have some effect (McSwiney et al., 1994; Oldfield et

al., 2000; Qi et al., 1995). At neutral pH the free disulphide group of β-Lg is very reactive and this is the

main mechanism for aggregation and gives a faster aggregation of β-Lg. Dissanayake et al. (2013b) have

shown that the denaturation rate is significantly lower at pH 3 compared to pH 6 and the aggregates

formed at pH 3 are caused by non-covalent bonding. This is consistent with free disulphide groups being

inactivated at acidic conditions. At neutral pH most whey protein complexes formed by denaturation are

soluble. A decrease in pH below 6.2 followed by heating gives a faster formation of whey protein/κ-casein

complexes and they were often associated with the casein micelles. Heating at a pH above 6.8 leads to dis-

sociation of the whey protein/κ-casein complexes from the micelle surface (Zúñiga et al., 2010).

The reaction kinetics of the denaturation of β-Lg has been widely investigated in order to predict the extent

of denaturation according to different heat treatments. Dannenberg and Kessler (1988) determined the

reaction order for the denaturation of β-Lg for each temperature by using the model

(1)

where n is reaction order different from 1 and

is the ratio of denatured β-Lg at holding time t and k is

the rate constant at a given temperature. They found that a reaction order of 1.5 gave a linear correlation

between denaturation of β-Lg and holding time in skim milk for various temperatures having the rate con-

stant defined as the slope of the linear graph.

This has been verified since, i.e. Zúñiga et al. (2010) and it is now widely used to report denaturation of β-

Lg in skim milk. Kessler and Beyer (1991) have shown that the reaction number varies between 1.5 and 2

according to the casein/whey protein ratio in the milk with skim milk having a reaction order of 1.5 while

sweet whey has a reaction order of 2.

- 16 -

Factors such as the methods used to detect differences in β-Lg denaturation, lack of enough data to make

an accurate determination, the heating methods and how the samples are heated, including preheat time,

temperature and cooling rate, can have a great impact on the denaturation of β-Lg (Oldfield et al., 1998a).

Dannenberg and Kessler (1988) also investigated the effect of temperature on the rate constant k for de-

naturation of β-Lg in skim milk. The rate constants for various temperatures can med used to make an

Arrhenius plot, equation 2.

(2)

This is visualised plotting the logarithm of rate constants against the inverse temperature in kelvin. From

linear regression of the data points, is it possible to calculate the activation energy for the denaturation

process. The Arrhenius plots are used to detect the effect of temperature on a specific chemical reaction.

Figure 7 shows the Arrhenius plots for β-Lg B, β-Lg A and α-La, which can be used to determine the rate of

denaturation of a given temperature.

From figure 7 it is observed that the relationship is linear, but only for given temperature ranges, namely

70-90 °C and 95-150 °C. The graph shows a bend around 95 °C where the activation energy above 95 °C

decreases significantly. This change in rate constant and activation energy indicates that two reactions are

taking place. At lower temperatures, denaturation and unfolding is the dominating reaction which re-

quires large quantities of energy, while at higher temperatures aggregation dominates which does not

Figure 7. The effect of temperature on rate constant for β-Lg A,

β-Lg B and α-La of skim milk. Remodeled from Dannenberg and Kessler (1988)

- 17 -

require same amount of energy.

The kinetics of α-La has been investigated in similar way and a reaction order of 1 was found and verified

since experiments made by Dannenberg and Kessler (1988) (Oldfield et al., 1998a). As seen in figure 7,

α-La also shows a change in activation energy when looking at the temperature effect on the rate con-

stant, but this change occurs abound 80 °C compared to 95 °C for β-Lg. The activation energy for α-La in

the temperature range 85°C-150°C is higher than the activation energy for β-Lg which supports the theory

of α-La being less heat stable but less prone to heat denaturation. However, this pattern for α-La is only

seen when it is solution with other denatured whey proteins. Calvo et al. (1993) found no thermal aggre-

gation of α-La in absence of other whey proteins and heat treated at 90 °C for 24 min. Addition of β-Lg

caused aggregation of α-La which was depended on the content of free thiol groups present on β-Lg.

2.4 Protein separation methods

Protein separation and identification can be done in various ways, depending on purpose of the separa-

tion. In the following, two often used techniques will be described.

2.4.1 Liquid Chromatography

A widely used method for separation is Liquid Chromatography (LC). LC comprises separation techniques

that can be used separate various compounds, such as proteins, carbohydrates, fatty acids and amino

acids, according to the desired characteristics, such as size and mass (Size Exclusion Chromatography

(SEC)), hydrophobicity or biological function (affinity chromatography). There are many different LC sys-

tems referring to the pressure and phase used, such as e.g. Reverse Phased High Performance Liquid

Chromatography (RP-HPLC) and Ultra High Performance Liquid Chromatography (UPLC). RP-HPLC is a fast

an accurate method for separation and quantification of milk proteins separating according to hydropho-

bicity (Aguilar, 2004). It contains a column with a stationary phase which is non-polar and hydrophobic

while the mobile phase is polar, which is the opposite of normal phased LC. The length of the carbon chain

bound to the stationary phase depends on the sample to be separated. The longer carbon chain, the more

hydrophobic is the stationary phase. The molecules in a sample mixture binds to the stationary phase if

they are non-polar and hydrophobic. Molecules that have hydrophobic groups and longer alkyl chains will

bind stronger to the stationary phase than molecules having polar groups in their structure. The more

hydrophobic solute, the higher is the affinity of the solute to the stationary phase and the more time the

solute spends in the stationary phase and the later it leaves the column. On the other hand, polar mole-

cules will not bind to the column to the same extent and will leave the column earlier. As a result,

- 18 -

molecules with different hydrophobicity elute at different times and are thereby separated (Giacometti

and Josić, 2013). In order to be able to separate all proteins and having no proteins bound to the column

at the end of analysis, a gradient of organic solvent is used. By use of a gradient, it is possible to control

the separation to certain extent in a specific area by changing the solvent from being non-polar to be

more polar over time, and more proteins will be detach the stationary phase and thereby be eluted and

detected. Even small changes in the gradient will change the separation and thereby the elution time. For

detection, for example a UV detector can be used at 214 nm to detect peptide bonds or 280 nm to detect

aromatic residues (Berg et al., 2006). The detection is visualized as a peak in a chromatogram showing

detection time since starting the analysis and the intensity of the detection. The time from start of analy-

sis until maximum detected of a fragment by the detector is called retention time (RT). If retention time is

used to characterize molecules and peptides fragments, the analysis conditions have to be carefully con-

trolled to be able to compare fragments from different samples. These conditions are; the pressure used,

the column dimensions and stationary phase, the composition of solvent and temperature (Aguilar, 2004;

Schlüter, 1999). Even small changes in these parameters will give a shift in retention time.

The separation of milk proteins in skimmed milk can be seen in figure 8. The area of each peak on the

chromatogram refers to the intensity of peptide bonds detected at a given time. The more intense peak,

the more peptide bonds are detected. These chromatograms and peaks for each protein indicate the ratio

between the different proteins, but it does not give a quantification of the concentration of each protein

present in milk.

Figure 8. An example of a chromatogram separating proteins from milk according to their hydrophobicity. Peak 1-3: κ-casein,

peak 4: αs2-casein, peak 5: αs1-casein with 8 phosphorylations, peak 6: αs1-casein with 9 phosphorylations, peak 7: β-casein

variant A1, peak 8: β-casein variant A

2, peak 9: α-Lactalbumin, peak 10: β-Lactoglobulin B, peak 11: β-Lactoglobulin A (Rauh et

al., 2014).

- 19 -

Due to high resolution, RP-HPLC is capable of differentiating between the various genetic variants of both

casein and whey proteins. Different structural isomers on the other hand will not be separated as they

have same hydrophobicity.

During irreversible denaturation of whey proteins, the globular structure is changed due to refolding in a

non-native structure. The denatured whey proteins are said to sediment when acidifying the solution to

pH 4.6 mainly due to aggregation, and they will therefore not occur when analysing the soluble phase

from the acidification.

By investigating the protein composition, it is possible to observe the amount of each protein that is in it

native form from RP-HPLC data. The amount of denatured protein can be calculated by comparing the

amount of native protein in non-heat treated samples with native protein in heat treated samples.

2.4.2 Gel electrophoresis

Protein separation using one dimensional gel electrophoresis (1DGE) is another widely used method. One

of the common gels used contains polymerization of acrylamide and bis-acrylamide. Polymerization of

these introduces crosslinking and pores are formed between the crosslinks. The size of the pores is de-

termined by the ratio between bisacrylamide and the concentration of acrylamide. This is used when pre-

paring the gel, by making a gradient of acrylamide to make a more specific separation in a certain protein

size area (Berg et al., 2006).

Sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) is a commonly used method due

to its high resolution of protein separating by size and charge which was firstly published by Laemmli,

(1970) and this is still the basis of SDS-PAGE used today (Jones, 1993). SDS denatures the non-disulphide

tertiary and secondary structure and covers the protein in negative charges corresponding to the length

of the protein chain. If proteins in a protein mixture have to be separated completely, the disulphide

bridges have to be reduced to separate eventually protein aggregates and unfold the proteins (Berg et al.,

2006). The proteins migrate differently in an electric field due to the different charges and this migration

is thus proportional with the molecular mass of the protein.

Proteins can also be separated according to their isoelectric point (pI), called isoelectric focussing (Radola,

1984). The protein mixture is loaded into a pH gradient in an electric field. The electric field makes the

proteins migrate towards their isoelectric point, pI, as a cause of their own charge and the migration stops

when reaching the pI and they have no net charge.

Isoelectric focusing has also become widely used as the first dimension of two-dimensional gel electro-

phoresis (2DGE), which is then followed by SDS-PAGE in second dimension. This technique is a powerfull

- 20 -

tool for protein separation and commonly used in proteomic profiling (Rabilloud et al., 2010). A major

advantage of 2DGE is the opportunity to separate different modifications of a protein from each other, as

these often vary in pI while the difference in mass can be difficult to separate as done in 1DGE. 2DGE also

gives the possibility to separate many different proteins and peptides in a complex matrix as milk, in one

single run. Each spot on the gel can be identified by different types of mass spectrometry (MS) (Jensen et

al., 2012a; O’Donnell et al., 2004).

Limitations on separation with both gel electrophoresis methods, is the limited amount of sample loading,

and the gels are often not capable to separate molecules with a mass above 250 kDa. Furthermore is gel

electrophoresis time consuming, especially for 2DGE where only one sample can be analysed at one gel.

2.5 Protein identification Protein separation described in section 2.4 can be coupled to MS analysis to identify the separated frac-

tions. This coupling is also useful for identifying unknown peaks and to show how purified the peaks from

LC and gel electrophoresis are (Lacorte and Fernandez-Alba, 2006). The principle of MS is to separate pro-

tein fractions according to their mass by use of an electric field. The proteins become ionised in an electric

field and the movements in this electric field depends on the molecular weight of the protein (Berg et al.,

2006). Small molecules move faster and are detected earlier by a UV-detector compared to the heavy

molecules.

Different molecular weights are often found when looking at the mass spectrometer of one fraction sepa-

rated with LC or gel electrophoresis, which can correspond to genetic variants or attachment of e.g. one or

more lactose units (Czerwenka et al., 2006).

2.6 Rennet coagulation of milk proteins

Raw milk is very prone to spoilage during storage and it is therefore necessary to process the milk to pro-

long shelf life. Cheese is one of many processing opportunities of milk. The first step in the cheese produc-

tion is the coagulation of casein micelles. This can be done enzymatic by adding different kinds of enzymes

which causes the caseins to clot. One of the most used enzymes is the enzyme Chymosin, also called ren-

net, which is a milk clotting enzyme originally obtained from calf stomach. Alternatively, pH can be low-

ered to the isoelectric point for casein which also makes the caseins coagulate. Chymosin cleaves κ-casein

between amino acid residue 105 and 106, resulting in a hydrophilic part, caseinomacropeptide, and a

hydrophobic part, para-κ-casein (Belitz et al., 2004).By removing the hydrophilic part of κ-casein, the ca-

- 21 -

sein micelles lose their solubility and when approximately 85 % of κ-casein is hydrolyzed, the colloidal

stability of micelles is reduced in such an extent that they will start coagulate (Fox and McSweeney, 1998).

It is widely known that milk treated at high temperatures has longer coagulation times, reduces the firm-

ness and forms weaker gels when manufacturing cheese (Waungana et al., 1996a). This can be caused by

complexes of denatured whey protein and κ-casein, leading to the formation of appendages on the mi-

celle surface which makes the Phenylalanine 105-Metionine 106 bond of κ-casein less susceptible to hy-

drolysis by rennet and thereby decreases the coagulation abilities according to cheese production

(Tolkach and Kulozik, 2007). Singh and Waungana (2001) observed that heat treatments which resulted in

denaturation degrees of less than 60 % of β-Lg had little effect on gelation time of skim milk, while gel

strength decreased for all detected denaturation degrees. These impairments in rennet coagulation prop-

erties can be restored in some extent after severe heating by addition of CaCl2 and lowering pH (Hougaard

et al., 2010; Waungana et al., 1996b).

Several modifications of milk have been investigated to reduce rennet coagulation time and improve tex-

ture. One method is to concentrate milk to achieve higher protein content. This gives firmer gels com-

pared to skim milk and reduced decrease in coagulation properties when heated at high temperatures,

while having the same degree of whey protein denaturation. This is due to the increase of caseins which

are placed closer because of reduced volume and thereby can the coagulation be altered (McMahon et

al., 1993). Another method is microfiltration of milk to reduce the amount of whey proteins, such as Mi-

cellar Casein Isolate (MCI). By removal of the whey proteins, a reduction in coagulation time and an in-

crease in gel firmness can be observed, compared to skim milk, also after heating (Pierre et al., 1992;

Wang et al., 2007). This supports the theory of whey proteins have a negative effect on the rennet coagu-

lation abilities of milk.

2.6.1 Measurement of coagulation properties

Knowledge about milk coagulation is of great importance to be able to make the best quality cheese.

Throughout time many different methods used to investigate rennet coagulation of milk, like oscillation

rheometry (Bohlin et al., 1984), dynamic light scattering (Horne and Davidson, 1990) and ultrasound

(Beguigui et al., 1994). One of the recently new methods to measure the coagulation of the milk is by a

ReoRox rheometer which is based on free oscillation rheometry. The ReoRox system is formerly used only

for determining blood coagulation properties but is has been shown that the results from milk coagulation

analysis on the ReoRox are consistent with analysis made on a rheometer which is well known and com-

monly used for milk coagulation in literature (Andersen, 2013; Frederiksen et al., 2011). The ReoRox con-

- 22 -

tinuously measures the elasticity and viscosity of a sample throughout the sampling time. Figure 9 shows

an example of a coagulation analysis of pasteurized skim milk with the four main characteristics shown.

These are the rennet coagulation time (RCT) which is the point of which the milk goes from fluid like to

solid like characteristics, the curd firming rate (CFR) which describes the speed of gel formation, and time

at which the gel contains gel strength of 200 Pa. It is possible to measure multiple samples at the same

time which makes it less time consuming when having many samples. The ReoRox is a single frequency

oscillation test which operates with a fixed strain which is expected to be within the linear viscoelastic

region for the given product. The ReoRox also uses a fixed frequency at 10 Hz. These two parameters fixed

to fit standard milk, which makes it is possible to determine when the milk goes from fluid to more solid

during time with the least variables to take into account (Andersen, 2013).

Figure 9. Rennet coagulation analysis of low pasteurized skim milk on ReoRox 4.Clot onset time: time at which first clots

are detected. Rennet coagulation time: time at which the milk goes from being more solid than fluid. Curd firming rate: the

rate of curd formation, measured in Pa/min. Time at 200 Pa: the time when the milk gel reaches strength of 200 Pa. (Ander-

sen, 2013.)

- 23 -

3 Materials and methods

3.1 Milk types

Milk used in the trials was skim milk delivered from Brabrand Dairy (Brabrand, Denmark) and micellar

casein isolate milk (MCI) from Arla Foods Innovation Centre Nr. Vium (Nr. Vium, Denmark). Skim milk from

Brabrand Dairy contained 3.47-3.79 % protein, 4.75-4.88 % lactose, 9.26-9.61 % total solid and less than

0.08 % fat.

MCI was produced from skim milk which was microfiltered (MF) to remove whey proteins. Less than 4 %

of the total protein content in MCI was whey protein. One portion MCI was adjusted to a total protein

content equal to skim milk (referred to as MCI) and one portion MCI was adjusted to a casein content

equal to skim milk (referred to as MCIc).The mineral, lactose and total solid content was equal in all three

milk types.

Pasteurised milk was used each trial day and delivered to Arla Strategic Innovation Centre (ASIC)

(Brabrand, Denmark). Milk which did not receive further heat treatment is referred to as control milk. The

overall milk composition of the control milk was measured by FT120 Milkoscan (Foss, Denmark).

3.2 Heat treatments

A laboratory scale UHT heat exchanger (Powerpoint International Ltd, Japan) with 3 different possibilities

to make heat treatments with both direct and indirect heating systems was used for heat treatments in

the pilot plant at ASIC. The three heating methods were tubular heat exchanger (THE), plate heat ex-

changer (PHE) and direct stream injection (DSI). The plates and tubes are composed of two heating units

and three cooling units. For temperatures above 100 °C, backpressure is used to avoid milk evaporating.

The pressure increases with increasing temperature and milk flow through the system and varied from

0.5 bar at 80 °C to 4.2 bar at 140 °C. The DSI system also uses vacuum down to 400 mbar, to be able to

withdraw the injected water from milk after heating. The preheating is always indirect, while the second

stage can be direct or indirect heating up to 150 °C. The three cooling units are indirect where the first

cooling unit uses tap water for cooling while second and third cooling unit uses ice water for cooling.

The Plate heat exchanger contains corrugated flow plates, specialized for this heat exchanger. The tubes

in the tubular heat exchanger are corrugated tubes in tubes made of stainless steel with an internal finish

of 1 micron in a soft spiral pattern with a tube diameter of 8 mm. The passage through one plate or tube

unit is approximately 20 s with a flow of 20 L/h. For DSI, PHE was used for preheating to 75 °C and cooling

- 24 -

was done with flash cooling to 65 °C followed by plate cooling system to 4 °C.

Preheat temperature for trials with an end temperature below 95 °C was 60 °C while a preheat tempera-

ture of 75 °C was used for trials with an end heating temperatures of 95 °C and above. Temperatures be-

tween 80 °C and 145 °C were used with a desired holding time between 2 s and 300 s. After heat treat-

ment the samples were immediately cooled to the end temperature of 4 °C.

Table 2 shows the various temperature and holding time combinations for skim milk samples. MCI was

not heated at 85 °C and 105 °C while the other combinations were equal. MCIc was only heated with PHE

at temperatures of 80 °C 115 °C and 130 °C with equal holding times compared to skim milk.

For all skim milk samples, and MCI heated with PHE, the temperature and holding time combination was

performed as biological duplicates at different trial days. For MCI heated with THE and DSI, each tempera-

ture and holding time was preformed once and for MCIc temperature and holding time was preformed

was performed once using PHE.

In total 135 skim milk samples, 68 MCI milk samples and 18 MCIc milk samples were produced and ana-

lysed.

Table 2. Temperature and holding time combinations used for heat treatment of skim milk. The method is stated if the combi-

nation has been used for the particular method. The milk flow through the system for each holding time is given.

Temperature (°C) Holding time (s) Flow (L/h)

80 85 95 105 115 130 140 145

4 20 DSI DSI DSI DSI 5 20 PHE

THE PHE PHE

THE PHE PHE PHE

THE PHE PHE

10 30 PHE THE

PHE PHE THE

PHE PHE PHE THE

PHE PHE

30 40 PHE THE

PHE PHE THE

PHE PHE PHE THE

PHE PHE

60 20 PHE THE

PHE PHE THE

PHE PHE PHE THE

120 20 PHE THE

PHE PHE

300 20 PHE THE

PHE

- 25 -

3.3 Protein analysis

Protein analysis was performed by a LC-MS system consisting of Agilent 1290 LC infinity system with

Agilent 6530 Accurate-Mass Q-TOF system (Agilent Technologies, USA). The MS detection was equal for all

LC-MS analyses. Mass scans of the protein fractions separated on the LC system were continuously re-

corded to detect fractions with a mass to charge ratio between 300 and 3500. Data analysis of MS data

was performed by using Mass Hunter (Agilent Technologies, USA) to identify the content of the peaks

found through the LC analysis of the milk samples.

Sample vials were kept 4°C and injected via an auto-sampler. The analysis method was a modification of

the model used by Bonfatti et al. (2008).

3.3.1 Analysis of total protein content

For each milk sample, 200 mg milk was frozen at -20 °C until further use.

Prior to analysis samples were defrosted and added a reduction buffer containing 6 M urea, 0.1 M triso-

dium citrate and 0.5 M dithiothreitol (DTE). The samples were incubated at 30 °C for 60 min while stirring

to reduce non-covalent bonds, crosslinks and disulphide bonds.

Hereafter the samples were centrifuged for 10 min at 4°C, 9300 ×g, and collected for analysis.

A volume of 5 μL sample was injected into the LC-MS system. A Biosuite column C18 PA-B 2.1x250 mm,

particle size of 3.5 μm and pore size of 300 Å was used (Waters, USA) is used. Buffer A contained 0.05 %

triflouroacetic acid (TFA) in milliQ water and buffer B contained 0.1 % (TFA) in acetonitrile. The column

temperature was 40 °C. A linear gradient of buffer B from 34.8 % to 46.5 % from 2 to 16.5 min was applied

with a flow rate of 0.35 mL/min and total analysis time was 21 min. UV detection of 214 nm was used to

detect the protein fractions.

3.3.2 Analysis of pH 4.5 soluble protein

20 mL of milk sample was adjusted to pH 4.5-4.55 with 1M hydrochloric acid (HCl) to sediment the caseins

and denatured whey proteins completely due to reaching their isoelectric point. This was done using a pH

meter (Knick GmbH,Germany). pH was measured before adding HCl and the pH was lowered while stirring

at room temperature. After reaching the pH the samples were stirred for at least 15 min to make sure

that the buffer effect of milk was removed.

The samples were then centrifuged 10 min, 4 °C at 17090 ×g to separate the precipitated caseins and de-

natured whey proteins in the pellet and the soluble whey proteins in the supernatant. The supernatant

- 26 -

was collected in eppendorf tubes and frozen at -20 °C until further analysis. Duplicates pH adjustments

were made for each milk sample collected.

Prior to LC analysis, one of each duplicate from all milk samples, pH 4.5 soluble protein solutions were

centrifuged for 10 min at 4°C and 9300 ×g to separate eventually precipitated casein and fat particles from

the supernatant. Approximately 1 mL of the supernatant was transferred to vials and analysed.

5-20 μL of each sample was injected into the LC-MS system. For milk samples heated less than 100 °C, 5 μL

was injected, for samples heated at 100-115 °C, 10 μL was injected and for samples heated at 130-145 °C,

20 μL was injected. These differences in injection volumes were caused by differences in content of pro-

tein left in the supernatant after adjusting the pH to 4.5. An Xbridge BEH300 C18, 2.1x250 mm column

with diameter of 5 μm, particle size of 3.5 μm and pore size of 300 Å was used (Waters, USA). Buffer A

contained 0.05 % TFA in milliQ water and buffer B contained 0.1 % TFA in acetonitrile. A gradient was ap-

plied with 17 % buffer B at2 min, 40 % buffer B at min 8, 44 % buffer B in min 14. The column temperature

was 45 °C. The flow rate was 0.35 mL/min and total time for each analysis was 17 min. The UV detection

of 214 nm was used to detect protein fractions.

Peak areas for β-Lg B, β-Lg A and α-La was calculated for each sample. All peak areas were adjusted to

same injection volume. The degree of denaturation of the whey proteins was calculated as the difference

in peak area between reference sample and heat treated sample.

3.4 Measurement of rennet coagulation

The coagulation properties of milk were performed using a ReoRox G2 Rheometer (Medirox, Sweden) one

day after heat treatment. The oscillation frequency is fixed at 10 Hz and amplitude of 2° which means that

the sample cup is swung 2° for each 2.5 seconds. This gives a corresponding strain of 0.07 which is the

strain within the LVR for standard milk.

The analysis was performed as described in Frederiksen et al. (2011). 103 g milk, corresponding to 100 mL

milk, was adjusted to pH 6.5 with 10 % lactic acid while stirring. The milk samples were incubated in water

bath at 33 °C for 30 min. Chy-Max Extra® was addition to the milk to a final concentration of 0.038

(Chr. Hansen, Denmark) was made. The Chy-Max Extra used for all samples, was from the same batch. The

addition of Chymosin defines starting point for the ReoRox analysis. 1 mL of the milk sample was trans-

ferred to a sample cup placed in the ReoRox while the remaining sample was placed in water bath for

visual analysis and confocal laser scanning microscopy (CLSM).The ReoRox measures the rennet coagula-

tion time, gelation point, curd firming rate, gel strength at 45 min analysis and time for reaching gel

- 27 -

strength of 200 Pa.

The rheological measurements were performed for 2 hours at 33 °C.

Duplicates of all milk samples were measured.

To avoid the impact of variation between the milk batches used, the coagulation properties for all heat

treated milk samples were calculated as the relative difference to the corresponding control sample.

3.5 Analysis of protein structure

For each milk type, three heat treated samples with a holding time of 10 s, and the matching control sam-

ples were analysed for protein structure 2 hours after addition of rennet while incubated at 33 °C (see

section 3.4 for renneting procedure).

A thin slice of milk coagulate was placed on a microscope slide and added a droplet 0.02 % (v/v) fluo-

rescein -5-isothiocyanate (FITC) solution in acetone. When the acetone was fully evaporated, the sample

was analysed using a 40x and 100x oil immersion objective coupled to a Leica DMIRE2 inverted confocal

laser scanning microscope (CLSM) (Leica Microsystems GmbH, Germany). The laser used was an Ar-

gon/Krypton ion laser. FITC excite UV light at 485 nm and the emission UV light is recorded in the green

area, between 500 and 545 nm. Triplicates of each sample were performed.

3.6 Analysis of protein aggregates

Heat induced protein aggregates were analysed by 1DEG, 2DGE and size exclusion chromatography (SEC).

Table 3 shows the detailed description of the samples used for analysis of protein aggregates. 4 selected

heat treated samples and 2 control samples of skim milk and MCI were analysed under reducing and non-

reducing conditions. The non-reducing conditions did not contain DTE to maintain the disulphide bonds,

while the reducing conditions contained DTE in the analysis process to break disulphide bonds.

- 28 -

Table 3. Samples used for analysis of heat induced protein aggregates

ID # Milk type Method Temperature (°C)

Holding time (s)

Protein % (milkoscan)

105 Skim Control 3.59

107 Skim DSI 130 4 3.56

139 MCI PHE 130 5 3.75

147 Skim THE 130 5 3.64

158 MCI Control 3.62

213 Skim PHE 130 5 3.66

3.6.1 1DGE

In 1-D analysis, the samples were analysed on a Criterion TGX any kDa 1-D gel containing 18 30 μL, 1.0 mm

wells, (Biorad,USA). The milk was diluted to three different concentrations with MilliQ water. These dilu-

tions were mixed with Laemmli sample buffer (65.8 mM Tris-HCl, pH 6.8, 2.1% SDS, 26.3% (w/v) glycerol,

0.01% pyronin Y) in the ratio 1:1. For each dilution, one fraction was reduced by adding 20mM DTE while

the other fraction remained unreduced and was only added Laemmli sample buffer All sample solutions

were incubated at 70°C for 1 min. 20 μL sample solution was loaded at the gel in two protein concentra-

tions, 20 μg, 10 μg. 10 μL Spectra multicolour broad range protein ladder marker was loaded as marker

(Pierce Biotechnology, USA).

The gel was placed in criterion Dedeca cell system (Biorad, USA) with 1M running buffer (19.2 mM glycin,

0.01 % SDS, Tris-base) and migration of the proteins was achieved by applying 200 V for 35 min. The gel

was fixed in Fix solution with 50 % ethanol and 8 % phosphoric acid for at least 2 hours while stirring. The

gel was coloured according to the colloid coomassie staining method (Kang et al. 2002). Gel images were

captured using an ImageScanner (Amersham Biosciences, Sweden).

The different protein bands were identified according to (Souza et al., 2000).

3.6.2 2DGE

The 2-D gel analysis was done according to (Jensen et al., 2012). For isoelectric focusing in 1st dimension,

samples were diluted 10 times with immobilised pH gradient (IPG) lysis buffer (7M urea, 2 M Thiurea,

40 mM Trisbase). For reducing conditions, 1 % DTE was added to the IPG lysis. Precast 11 cm strips, pH 4-7

(Biorad, USA), loaded with an amount of dilution corresponding to 100 μg protein and left for rehydration

- 29 -

for 16-18 h at room temperature in darkness. Isoelectric focusing was carried out using an electrophoresis

chamber (Biorad, USA) with an increasing volt gradient with a maximum of 57.6 kV hours.

After focusing, the strips were incubated in equilibration buffer 1 (6M Urea, 30 % (v/v) glycerol, 2% SDS

and 0.05 % Tri-HCl) for 15 min. 65mM DTE was added reduced samples. The strips were incubated in

equilibration buffer 2 (6M Urea, 30 % (v/v) glycerol, 2% SDS and 0.05 % Tri-HCl) and for reduced samples

270 mM Iodoacetamide, for 15 min. The strips were then placed on top of a Criterion Precast gel, 8-16 %

tris-HCl, 1.0 mm IPG+1 well comb (Biorad, USA) using 0.5% agarose coloured with bromophenolblue to

bind the strip. The gels were electrophoresed, stained and coloured in same procedure as 1-dimensional

gels (section 3.6.1).

The different protein bands were identified according to Jensen et al. (2012b) and Larsen et al, (2010).

3.6.3 Analysis of protein aggregate composition

Size exclusion chromatography (SEC) was performed on a 1260-infinity semi-preparative LC system

(Agilent Technologies, USA). Reduction of samples was performed as described in section 3.3.1 was used.

For the non-reduced samples, same procedure was used, but no DTE was added to the buffer.

150 μL of each sample was injected on a Bio SEC-3 Column (3μm, 7.8 × 300 mm, 300 Å, Agilent Technolo-

gies, USA) at 30 °C. The mobile phase contains of 0.2M sodium phosphate pH 7 and 0.2 M Sodium chlo-

ride. This column separates in the molecular range of 1200000 to 5000 Da. A flow rate of 1 mL/min was

applied and UV signals were detected at 214 nm.

For 2 skim milk samples, respectively heated at 130 °C for 5 s at PHE and THE, the protein aggregations

were analysed.

1 mL milk sample was added 5 mL buffer containing 6 M urea and 0.1 M trisodium citrate. The samples

were incubated at 30 °C for 60 min while stirring followed by centrifugation for 10 min at 4°C, 17090 ×g

and analysed as described above. 1 mL of four fractions in the time frame 8-12 min was collected by a

fraction collector from each milk sample.

The fractions were freeze dried under vacuum (0.1 bar) overnight in a Hetosicc freeze-dryer (Pfeiffer Vac-

uum GmbH, Germany). The freeze dried fractions were dissolved in 100 μL reduction buffer containing

6 M urea, 0.1 M trisodium citrate and 20 mM DTE. These samples were incubated at 30 °C for 60 min

while stirring. 20 μL of each sample was injected and analysed on LC as described in section 3.3.1.

- 30 -

3.7 Statistical analysis

One way-ANOVA analysis was performed to determine significant differences (p<0.05) among milk types

and heating methods, carried out using MATLAB (Mathworks, USA).

- 31 -

4 Results

4.1 Variation in milk protein composition Milkoscan was used for a fast determination of milk components. There were small variations in the

amount of protein, lactose, dry matter and fat between the milk batches, but these differences were not

significant and no pattern for the variation in composition was found.

Further analysis of total protein was made of all control samples for all milk types, to investigate differences

in the protein content between trial days. Figure 10A shows the analysis of total protein content of three

Figure 10. Total protein analysis of control samples of the three milk types. A: three skim milk control samples from three

different batches. B: Total protein analysis of control samples of skim milk, MCI and MCIc.

Peak 1-3: κ-casein, peak 4: αs2-casein, peak 5: αs1-casein with 8 phosphorylations, peak 6: αs1-casein with 9 phosphorylations,

peak 7: β-casein variant A1, peak 8: β-casein variant A

2, peak 9: α-La, peak 10: β-Lg B, peak 11: β-Lg A.

- 32 -

control skim milks from three different batches from three different weeks, analysed on LC-MS. The peaks

on the chromatogram were analysed by use of the MS data for each fraction. From figure 10A It is clear

that there are small variations in the protein content, especially in the casein composition, which are ob-

served in fraction 1-8. The total protein content measured on the Milkoscan also showed these variations,

respectively 3.471 %, 3.598 % and 3,663 % protein for samples from sampling week 2, 14 and 21.

Figure 10B shows analysis of total protein of control samples of the three milk types, from the same trail

week but from different batches. Small variations were found for the caseins fractions, which are seen in

fraction 1-8. The content for casein is higher in the MCI milk, while MCIc and skim milk are equal in casein

content. There are only small amounts of whey protein present in MCI and MCIc compared to skim milk.

The amount of β-Lg B, β-Lg A and α-La, shown in fraction 9-11, represent 2-4 % of the total protein content

in MCI while it was around 2 % of total protein content for MCIc. Due to this low content for whey proteins

in MCI and MCIc, the denaturation degree of whey protein was only analysed for skim milk samples.

Variations in the milk composition also gave rise to variations in enzymatic coagulation. Table 4 shows the

average relative standard deviation, minimum and maximum for the four measured parameters from 38

coagulation analyses on control skim milk from 19 different trial days. A large variation was found in the

measured properties between the control skim milk samples. A variation of RCT from 11.6 min to 19.6 min

and CFR varying from 10 Pa/min to 24 Pa/min had great influence on gel strength at 45 min and also on the

time reaching gel strength of 200 Pa. The relative change in coagulation properties for heat treated samples

according to the control sample was therefore calculated and used throughout the report, unless otherwise

it stated.

Table 4 Rennet coagulation properties of 38 control skim milk samples. Rennet coagulation time, curd firming rate, gel

strength at 45 min and time at gel strength of 200 Pa is shown.

Rennet

coagulation

time (min)

Curd firming

rate (Pa/min)

Gel strength,

45 min (Pa)

Time at 200 Pa

gel strength

(min)

Average 15.37 14.76 398.06 29.56

Relative standard

deviation (%)

11.51 23.63 21.64 13.84

Minimum 11.60 10.10 284.00 20.20

Maximum 19.60 24.40 641.30 35.90

- 33 -

4.2 Effect of indirect heating on skim milk For each skim milk sample, pH 4.5 soluble protein fractions were analysed on LC –MS in duplicates and

integration of peak areas of each whey protein fraction was used to calculate the percentage of whey pro-

tein in the heat treated samples compared to the control sample. The denaturation degree is defined as the

percentage of whey protein not appearing in pH 4.5 soluble protein analysis compared to the control sam-

ple, which is stated to have a native whey protein percentage of 100 % and thereby no denaturation.

Heat treatment using THE and PHE were compared according to temperature and holding time to investi-

gate differences in denaturation degree of whey proteins and coagulation properties between the two

heating methods. Figure 11 shows the denaturation degree of β-Lg B and α-La in skim milk heat treated

using PHE and THE. The denaturation pattern of β-Lg A is shown in Appendix 1 which shows similar results

as denaturation of β-Lg B. The denaturation of β-Lg B is shown in figure 11A, for heating temperatures

Figure 11. Denaturation degrees of β-Lg B and α-La for skim milk heated at PHE and THE.

A: Denaturation of β-Lg B heated at temperatures below 100 °C. B: Denaturation of β-Lg B heated at temperatures above 100 °C.

C: Denaturation of α-La heated at temperatures below 100 °C D: Denaturation of α-La heated at temperatures above 100 °C

Different letters indicate significant differences were found between heating methods for each given temperature (p<0.05).

- 34 -

below 100 °C and in figure11B for heating temperatures above 100 °C. For all temperatures, an increase in

denaturation degree was observed as the holding time increases. However, the greatest effect was seen by

increase in temperature, especially when temperatures above 85 °C are used. A heating temperature of 85

°C resulted in denaturation degree of 25 % of β-Lg B with a holding time of 5 s, while heating at 130 °C and

140 °C resulted in denaturation degrees above 90 % even at very low holding times.

This was applicable for both heating methods. Comparing the denaturation degrees for heating tempera-

tures at 80°C and 95 °C using PHE and THE from figure 11A, it is clear that THE results in significant higher

denaturation degree than PHE for holding times above 10 s for 80 °C (p<0.04) and for holding times below

120 s for 95 °C (p< 0.009) as both methods have denaturation degrees above 90 % at holding times of 120 s

and 300 s for heating temperature of 95 °C. Figure 11B shows heating temperatures above 115 °C, which

results in denaturation degrees of β-Lg B above 90 % for both heating methods.

The denaturation of α-La is shown in figure 11C and figure 11D. It is observed that the denaturation of α-La

is also affected by temperature and holding time. The degree of denaturation of α-La is however lower than

the degree of denaturation of β-Lg B, which is shown in figure11 A and figure 11B. Figure 11C shows the

denaturation degree of α-La for temperatures below 100 °C. From this, it is observed that the denaturation

degree does not exceed 50 % at any of the investigated holding times for both PHE and THE. No significant

difference in denaturation of α-La was found when heating at 80 °C between methods. Heating tempera-

ture at 95 °C results in significant higher denaturation degree was found for heat treatment using THE,

compared to PHE, with a holding time of 60 s (p<0.002). Figure 11D shows the denaturation degree of α-La

for heating at temperatures above 100 °C. Heating at 130 °C and 140 °C gave the highest denaturation de-

gree, but no denaturation degrees above 90 % were found for any of the investigated temperature and

holding time combinations. Heating temperatures of 130 °C results in significant higher denaturation de-

gree for heat treatment using THE compared to PHE for all holding times (p<0.001).

Rennet coagulation analysis of skim milk heat treated using PHE and THE are shown in figure 12. Figure 12A

and figure 12B shows the relative RCT of heat treated skim milk samples, heat treated below and above

100 °C, respectively. From figure 12A, it is observed that heat treatment at temperatures below 85 C only

have slight increase in coagulation time at all holding times. Heating temperatures of 95 °C had a significant

increase in relative RCT, even at the short holding times for both heating methods (p< 0.006). Comparing

the two methods, there is a tendency towards PHE having less increase in relative RCT for temperatures of

80 °C and 95 °C, but there is no significantly difference between the two methods at these temperatures

for all holding time (p = [0.52-0.86]). Heating at temperatures above 115 °C shown in figure 12B, resulted in

very long RCT and samples heated at 140 °C using PHE did not reach RCT within 2 hours for any of the in-

- 35 -

vestigated holding times. Samples heated at 130 °C resulted in large increase in relative RCT and there is a

significant difference between the methods for all holding times (p <0.03]). THE had less increase in relative

RCT compared to PHE and RCT was detected within two hours with use of holding times of 30 and 60 s,

which was not observed for heat treatment using PHE.

The relative CFR for skim milk samples heated at PHE and THE is shown in figure 12C and figure 12D. The

relative CFR decreases with increasing temperature and holding time. Comparing relative CFR for the two

heating methods at temperatures below 100 °C from figure 12C, there is a significantly lower CFR for THE

heated at 80 °C with holding times above 120 s (p<0.025). This is also observed for heating at 95 °C with

holding time less than 60 s (p<0.048). Holding times above 60 s results in very low relative CFR for both

methods, with a CFR corresponding to less than 5% of control samples. Comparing heating at 80 °C and

85 °C at PHE, small differences in relative RCT was found in figure 12A, but when comparing CFR, the de-

Figure 12. The relative Rennet coagulation time and relative curd firming rate for skim milk samples heated at PHE and

THE.

A: Relative RCT of skim milk heated at PHE and THE at temperatures below 100 °C. B: Relative RCT of skim milk heated at

PHE and THE at temperatures above 100 °C. Graphical lines continuing out of the visualized graph indicates that the RCT is

not reached within two hours of measurements. C: Relative CFR for skim milk heated at PHE and THE for temperatures be-

low 100 °C. D: Relative CFR for skim milk heated at PHE and THE for temperatures above 100 °C. Graphical lines going out of

the visual range indicates that no CFR was detected within two hours of measurement.

Different letters indicate significant differences were found between heating methods for each given temperature, at a given

holding time (p <0.05).

- 36 -

crease in relative CFR is significantly lower for 85 °C compared to 80 °C, at holding times of 30 s (p=0.002)

and 60 s (p=0.003).

Heating at temperatures above 100 °C is shown in figure 12D had strong pronounced decrease in relative

CFR, with only low CFR detected, even though the RCT was reached within two hours of measurement, as

observed in figure 12B. The UHT treated samples did not reach RCT within two hours, and therefore no

curd firming rate was detected. PHE had significant lower CFR when heating at 130 °C for all holding times

(p<0.03), compared to THE.

4.3 Effect of indirect and direct heating systems on heat treatment of skim milk To investigate how different heating method affects the milk properties , skim milk samples heated at DSI,

PHE and THE were compared. Milk samples heated using DSI were only heated with one holding time,

namely 4s. Milk samples heated indirectly used a holding time of 5 s. When comparing the three methods,

the variation in holding time should be kept in mind.

Figure 13A shows the denaturation degree of β-Lg B in skim milk heated at DSI, PHE and THE at various

temperatures. Heating at 105 °C, no denaturation was found when heating at DSI while heating at PHE was

significantly higher (p<0.0001), which has a denaturation degree of 70 %. For indirect heating, a denatura-

tion degree above 90 % was observed for heating temperatures of 115 °C or more. There is significant dif-

ference between the two indirect heating methods and DSI (p<0.002) with DSI having lowest denaturation

degree for all comparable temperatures. The denaturation of α-La is seen in figure 13B. No denaturation

Figure 13. Denaturation degrees of whey protein in skim milk heated with DSI for 4 s and skim milk heated with PHE and THE

for 5 s. A: denaturation degree of β-Lg B. B: denaturation degree of α-La. Significant differences were found between heating

methods for each given temperature. Different letters indicate significant differences were found between heating methods for

each given temperature (p <0.05).

- 37 -

was detected for heating at 105 °C using DSI and a low increase in denaturation is observed as temperature

increases. The denaturation of α-La is significantly lower for DSI compared to the indirect heating at all

temperatures (p<0.0001). These variations in denaturation degrees between indirect heating and DSI are

large, which makes it reasonable to say these large differences would also be present if the indirect heating

samples had a holding time of 4 s.

The rennet coagulation properties were compared for the three heating methods. Figure 14A shows the

relative RCT for skim milk samples using PHE and THE heated for 5 s and DSI heated for 4 s. There is a clear

significant difference between PHE and DSI for all temperatures (p<0.002), with DSI having the lowest in-

crease in RCT. Skim milk was heated at 140 °C using PHE and this did not reach RCT within two hours. DSI

samples heated at 145 °C reached RCT within two hours and the relative increase in RCT was lower than

samples heated at 130 °C using PHE. THE and DSI can only be compared at heating of 130 °C but here is the

difference also significant.

In figure 14B, the relative CFR is shown. Both PHE and THE have significantly lower CFR for all temperatures

compared to DSI (p<0.024). The DSI sample heated at 145 °C reached RCT within two hours, but no CFR was

detected. The differences in RCT and CFR between the indirect and direct methods are large, which makes

it reasonable to say that there would still be a significant difference between the methods if they had the

same holding time.

Figure 14. Relative RCT and CFR of skim milk heated with PHE for 5 s and DSI for 4 s.

A: relative RCT for skim milk heated at DSI, PHE and THE. PHE heated 140 °C is not within two hours, which is indicted with a bar

going out of scaled area. B: relative CFR for skim milk heated at DSI, PHE and THE. Samples with negative CFR indicate that RCT was

not detected within two hours of measurement. Different letters indicate significant differences were found between heating methods for

each given temperature (p <0.05).

- 38 -

4.4 Effect of whey protein denaturation on rennet coagulation

The relative RCT and denaturation degree of β-Lg B of heated skim milk was compared to investigate the

correlations between degree of denaturation and relative RCT. Figure15 shows the relative RCT as a func-

tion of denaturation degree of β-Lg B for skim milk samples heat treated using PHE, THE and DSI.

It is observed that the denaturation degree of β-Lg B in some extent explain the increase in relative RCT.

Denaturation degrees of 50 % or less gives only small changes in relative RCT for the two indirect heating

methods, while it is seen that THE has lower relative RCT as the denaturation degrees of β-Lg B exceeds

70 %. Heat treatment using PHE at temperatures of 130 °C and 140 °C did not reach RCT within two hours

of measurement and these samples had denaturation degrees of 92 % or more. These are imagined in

figure 15 by exceeding relative RCT of 800 %. Heat treatment using THE can be described as a two phased

linear correlation with a break at denaturation degrees around 95 %, while heat treatment using PHE can

be explained exponentially.

Heat treatment using DSI is also shown in figure 15. It is seen that a denaturation degree for β-Lg B around

40 % resulted in significantly longer relative RCT compared to samples heated indirectly with same denatu-

ration degree (p<0.001). From section 4.3, it is clear that this denaturation degree is observed at low tem-

peratures for heat treatment using PHE and THE, while DSI was heated at 145 °C to obtain same denatura-

tion degree. This indicates that it is not only the denaturation degree of whey proteins that has an impact

Figure 15. The denaturation degree of β-Lg B shown as a function of relative RCT of skim milk heated at DSI, PHE and THE.

Data points exceeding relative RCT of 800 % did not reach RCT within two hours of measurement.

- 39 -

on the RCT, but that the temperature has a great impact, as it gives rise to other chemical changes in the

milk which has an impact on coagulation.

4.5 Effect of heat treatment of MCI milk samples on rennet coagulation The rennet coagulation properties of MCI milk samples heat treated using PHE and THE were analysed and

compared to investigate the effect of heating method. Only one trial day was used to collect MCI samples

heated using THE and thereby only temperatures with holding times of 5 and 10 s, and for 80 °C also one

sample with a holding time of 120 s, were collected. The data points for heat treatment using THE are

therefore an average of two rennet coagulation analysis while milk heated at PHE is an average of four ren-

net coagulation analysis from two trial days. The relative RCT and relative CFR for MCI samples heat treated

using PHE and THE are shown in figure 16.

Figure 16. RCT and relative CFR for MCI samples heated at PHE and THE. A: Relative RCT of MCI samples heated at PHE and THE

at temperatures below 100 °C. B: Relative RCT of MCI samples heated at PHE and THE at temperatures above 100 °C. C: Relative

CFR for MCI samples heated at PHE and THE below 100 °C. D: Relative CFR for MCI samples heated at PHE and THE above 100 °C.

Graphical lines going out of the visual range indicates that no CFR was detected within two hours of measurement.

Different letters indicate significant differences were found between heating methods for each given temperature, at a given

holding time (p <0.05).

- 40 -

Figure 16A shows the relative RCT for MCI samples heated at temperatures below 100 °C. The relative RCT

follows a parable formation for heat treatment using PHE at temperatures 100 °C. The relative RCT in-

creases for holding times up to 60s and then decreases when a holding time of 120 s is used. Here, the rela-

tive RCT was less than RCT of control samples. As not all holding times were measured for heat treatment

using THE, it is not possible to say if heat treatment using THE follows the same pattern. The relative RCT is

significantly larger for heating at THE than PHE when comparing each holding time between the methods

(p<0.03). MCI samples heat treated at temperatures above 100 °C are shown in figure 16B. Increases in

relative RCT was observed for all temperatures and holding time combination, but all measured samples

reached RCT within two hours of measurement. Heating at PHE was observed to give significant lower rela-

tive RCT compared to heating at THE (p<0.04).

Figure 16C and figure 16D shows the relative CFR for MCI samples heat treated at temperatures below and

above 100 °C, respectively. A decrease in relative CFR is observed for all MCI samples heat treated using

PHE at temperatures below 115 °C with a holding time of 5s, but the relative CFR then increases as the

holding time is increased from 5s up to 30 s. Holding times above 60s resulted in a decrease in the relative

CFR. Heat treatment using PHE has higher relative CFR compared to the corresponding control sample.

Comparing the two heating methods, there is significant lower relative CFR for heat treatment using THE

compared to PHE at temperatures below 100 °C (p<0.03). Heating at temperatures of 130 °C and 140 °C

resulted in a fast decrease in relative CFR for heat treatment using PHE, but a CFR was detected for all MCI

samples measured. No CFR was detected for heat treatment at 130 °C using THE.

Table 5. Relative RCT and relative CFR for MCI samples heated at PHE, 5s and THE, 5 s and DSI, 4s. PHE samples are heated at

140 °C and not at 145 °C ad DSI. *: No CFR was detected within two hours of measurement. Different letters indicate significant

differences were found between heating methods for each given temperature (p <0.05).

Relative RCT (%) Relative CFR (%)

Temperature (°C) DSI, 4 s PHE, 5 s THE, 5 s DSI, 4 s PHE, 5 s THE, 5 s

105 103.23 108.24

115 111.69 120.76 105.88a 78.60c

130 114.11a 138.87b 168.44c 93.82a 44.82b 0*

140 168.82 32.73

145 124.19 64.85

- 41 -

The two indirect heating methods with a holding time of 5 s were compared with direct heat treatment

using DSI with a holding time of 4s, which is shown in table 5.

It is observed that the relative RCT increases slightly as the temperature increases for heat treatment using

DSI. Only two temperatures could be compared directly between the heating methods, namely tempera-

tures of 115 °C and 130 °C using PHE and 130 °C using THE.

There is a significantly difference between DSI and the two indirect methods for heat treatments at 130 °C

(p<0.002). MCI samples were heated at 140 °C using PHE while samples using DSI were heated at 145 °C.

Comparing the relative RCT of these, there are significant differences between PHE at 140 °C and DSI at

145 °C (p<0,003) and it is therefore reasonable to state that a MCI sample heat treated at 145 °C using PHE

would have an even longer relative RCT and thereby also be significant different from MCI samples heat

treated using DSI.

The relative CFR for MCI heated using DSI decreased slightly as temperature increased. Heating at 145 °C

had a relative CFR of 64.85 % while heating at 130 °C using THE, no CFR was detected within two hours of

measurement. There is significant larger decrease in CFR for the indirect methods compared to DSI for all

comparable temperatures (p<0.02).

4.6 Effect of milk type on rennet coagulation Control milk samples from the three milk types were compared to investigate the differences in coagulation

in relation to milk type. The three milk samples are from the same trial week, but from different milk

batches, to avoid seasonal variation. From figure 17A, it is observed that the RCT for the three milk types

were similar and no significant variation was found, even though skim milk as a tendency to have a longer

RCT. Figure 17B shows the curd firming rate. From this, is it clear that MCI, with a casein content of 3.5 %,

had faster gel formation compared to skim milk and MCIc which have casein contents of 3.05 % (p<0.04).

The CFR of MCIc milk is larger than skim milk but this difference was not found to be significant.

These results indicate that the removal of whey proteins from low pasteurized skim milk gives a slight

faster gel formation, while removal of whey proteins and increasing the casein content increases the gel

formation significantly. This is important to have in mind, as the relative difference of control samples can

be equal for the milk types, but it does not consider MCI having faster curd firming rate.

- 42 -

The effects of heat treatment on the three milk types at PHE were analysed. Heat treatment of MCIc milk

samples was only performed using PHE and each temperature and holding time combination was only

measured once due to lack of time. These data points are therefore only the average of two ReoRox meas-

urements while skim milk and MCI samples are an average of four ReoRox measurements from two trial

days. Figure 17A and figure 17B shows the relative RCT for skim milk, MCI and MCIc samples heat treated

using PHE. Significant variations were observed between the three milk types heated at 80 °C (p<0.001).

The MCI and MCIc heated at temperatures of 115 °C and 130 °C reached RCT within two hours, while skim

milk samples with a holding time exceeding 30 s did not reach RCT within two hours for the same tempera-

tures. No significant difference was found between MCI and MCIc but skim milk increased significantly in

relative RCT compared to MCI and MCIc heat treated at temperatures of 115 °C and 130 °C (p<0.03).

The relative CFR of heat treatment of the three milk types are shown in figure 18C and figure 18D. MCIc

heated at 80 °C resulted in large increase in CFR with a holding time of 5s and afterwards decreased until a

holding time of 30 s from where it kept fairly constant. MCI and skim milk decreased in CFR with holding

times below 10 s while a slight increase was observed for longer holding times. Overall, significant differ-

ences in relative CFR between all three milk types was found (p<0.03).

Heat treatments at temperatures of 130 °C resulted in significant difference between the three milk types

(p<0.02). MCI had the least decrease in relative CFR which was significantly lower than for MCIc, even

though CFR was detected for both milk types within two hours of measurement. CFR was only observed for

heating at 130 °C for 5 s for skim milk and this is significantly different from MCI and MCIc (p<0,001). Sig-

Figure 17. RCT and CFR of control milk samples for the three milk types, from the same trial week.

A: RCT of skim milk, MCI and MCIc. Skim milk has the greatest RCT, but no significant differences between the three milk types.

B: CFR of skim milk, MCI and MCIc. MCI has the CFR rate, but no significant differences between skim milk and MCIc. Different

letters indicate significantly different values (p<0.05)

- 43 -

nificant difference between the milk types heated at 115 °C was found, with skim milk having the most

intense decrease in CFR and MCI having the least decrease (p<0,01).

4.6.1 Formation of protein network

The formation of protein network induced by rennet was investigated for six samples – three skim milk and

three MCI samples. The skim milk samples are from same batch, which is also valid for the MCI samples.

Figure 19 shows images of the protein structure of the six samples captured with CLSM. FTIC is bound to

the proteins and this gives the green emission of the proteins. The heat treated samples were all heated

using PHE with a holding time of 10 s. For all MCI samples and for control skim milk, it is seen that a strong

and compact protein network was formed, which indicates that the majority of the caseins are bound in the

network. This can be seen by the clear separation of the proteins coloured green and the dark background.

The dark background indicates that the most protein detected on the images, are bound in protein net-

work.

Figure 18. Relative rennet coagulation time and curd firming rate for skim milk, MCI and MCIc samples heated at PHE.

A: Relative RCT for the three milk types heated at 80°C. B: Relative RCT for the three milk types heated at 115 °C and 130 °C.

Graphical lines continuing out of the visualized graph indicates that the RCT is not reached within two hours of measurements

C. Relative CFR for the three milk samples heated at 80°C. D: Relative CFR for the three milk samples heated at 115 °C and

130 °C. Graphical lines going out of the visual range indicates that no CFR was detected within two hours of measurement.

Different letters indicate significant differences were found between the milk types for each given temperature (p <0.05).

- 44 -

The heat treated MCI samples shown in figure 19B and figure 19C form dense networks compared with the

control sample, shown in figure 19A, having very compact and strong network. This is observed, as strong

networks are compact and little space is found in between the protein networks and less protein not bound

in the network. The skim milk sample heated at 80 °C for 10 s shown in figure 19E, has no large contrast

between the green colour between the network and background. This indicates that there are proteins not

bound in the network. Comparing the sample heated treated at a temperature of 80 °C with the corre-

sponding control sample, shown in figure 19D, t is clear that the control sample has more compact net-

work. The skim milk sample heated at 130 °C, shown in figure 19F, contains many small networks, but the

formation of one large protein gel has not occurred yet. This can also be seen as the solution surrounding

the small networks contains large amounts of proteins due to the background is quite green. These results

are consistent with the results shown in figure 17 and figure 18. An increase in heating temperature gives

decrease in RCT and CFR, which therefore gives prolonged protein network formation. This is more pro-

nounced in skim milk compared to MCI.

Figure 19. CLSM images of the protein network in three skim milk samples and three MCI samples heated at PHE. The

proteins are colored green. A: MCI control sample. B: MCI sample heated at 80 °C, 10 s. C: MCI sample heat at 130 °C, 10 s.

D: Skim milk control sample. E: skim milk sample heated 80 °C, 10 s. F: skim milk sample heated 130 °C, 10 s.

- 45 -

4.7 Heat induced protein aggregation Protein aggregation was investigated with three different analytical techniques, namely 1DGE and 2DGE

and SEC. The main focus is on the skim milk samples, investigating the aggregate size and composition,

mainly for indirect heating.

4.7.1 1-DGE

Two 1-D gels with four skim milk samples and two MCI samples, analysed reduced and non-reduced,

are shown in figure 20. The milk samples were analysed under reducing conditions to disrupt the S-S

bridges within the proteins structure and between various proteins. Milk samples analysed under

non-reducing conditions have intact S-S bridges and differences between the reduced and non-

reduced analysis can be used to identify protein bands containing S-S bridges in the structure. The

protein bands were identified according to (Souza et al., 2000).

The whey protein bands of heat treated skim milk, which are observed at the low molecular mass

range of the gels in figure 20 are of low intensity, both under reduced and non-reduced conditions.

These bands are less intensive than the whey protein bands for the control skim milk. Decreases in

intensity of these bands were most pronounced in heat treated samples using PHE while heat treat-

ment using DSI having the least reduction. This is consistent with results presented in section 4.2 and

section 4.3. No protein bands were found in the top of the gel, which indicates that there are no

large protein aggregates present in the reduced samples.

For all non-reduced fractions, there is a clear band at the top of the gels above the marker band of

300 kDa. This indicates there are protein complexes that do not migrate on the gel. These are various

large aggregates bound together with disulphide bonds as these do not appear in the reduced sam-

ples. The non-reduced samples of skim milk heated using the three different methods, shown in

figure 20B, indicate that there are more complexes reaching the gel in the sample heat treated using

DSI compared to the indirect methods, while there is a tendency toward more protein complexes

that are not migrating at the gel for the indirect methods. The band for κ-CN is not very pronounced

in the non-reduced samples for all milk samples but becomes clearer in the reduced samples for all

six samples. The same tendency is seen for β-Lg even though the amount of β-Lg is low in the heat

treated samples due to denaturation. This indicates that κ-CN and β-Lg is present in the large com-

plexes. For the heat treatment of the MCI sample, shown in figure 20A, the complexes not migrating

on the gel in the non-reduced sample may also contain other caseins as the fraction of whey protein

is very small and it seems like the casein bands become slightly more intense.

- 46 -

Figure 20. 1DGE of six milk samples in reduced and non-reduced form, visualized by colloid Coomassie Brilliant Blue

G-250 staining. The molar masses of the marker used are given, as well as the amount of protein loaded in each well. The

most intense bands are identified. A: three samples; control skim milk, control MCI and MCI heated at 130 °C, 5 s using

PHE are shown in a reduced and non-reduced form. B: three skim milk samples; 130 °C, 5 s at PHE, 130 °C, 5 s using THE

and 130 °C, 4 s using DSI are shown in a reduced and non-reduced form.

- 47 -

4.7.2 2DGE

2DGE was performed on same samples as for 1D gel electrophoresis, section 4.7.1. The protein sport on the

gel were identified according to Jensen et al. (2012b) and Larsen et al, (2010). Four 2D gels are shown for in

figure 21. These gels contain skim milk samples heat treated using PHE and THE, and analysed under reduc-

ing in both dimensions and under non-reducing conditions in both dimensions. The 2D gels that are non-

reduced in both dimensions were less clear in the separation between protein bands compared to the re-

cued 2D gels. It is observed for both milk samples that the spots of β-Lg and α-La were moved towards a

higher pI in the non-reduced samples compared to the reduced samples. This can be due to refolding of the

denatured protein into a non-native structure and thereby changing the pI. β-casein was not affected by

electrophoresis method, while some αs1-casein multimers (sport 6) and αs2-casein dimmers (spot 7) were

observed on the non-reduced 2D gels. The 2D gels of the remaining milk samples are shown in Appendix 2.

For the skim milk sample heat treated using PHE, figure 21A and figure 21B, there is a clear difference in

Figure 21. 2D gel electrophoresis on skim milk samples heated with PHE and THE run with and without DTE in both dimen-

sions. A: Reduced skim milk sample heated with PHE at 130 °C for 5 s. B: Non-reduced skim milk sample heated with PHE at

130 °C for 5 s. C: Reduced skim milk sample heated with THE at 130 °C for 5 s. D: Non-reduced skim milk sample heated with

THE at 130 °C for 5 s. 1: genetic variants of κ-CN. 2:β-Lg. 3: α-La. 4: β-CN. 5: αs1-CN. 6: αs1-CN aggregates. 7: αs2-CN dimers.

- 48 -

intensity of the whey protein bands and also κ-casein bands at the non-reduced and reduced 2D gels. The

κ-casein bands are not visible on the non-reduced 2D gel and the whey protein spots are weak. This indi-

cates that the whey proteins and κ-casein are bound in various complexes which are not seen on the non-

reduced 2D gels, while these complexes are broken down in reduced 2DGE analysis where each protein

fragment is seen individually on the gel. Furthermore can the low intensity of κ-casein be caused by

κ-casein fund as multimers in milk.

The skim milk sample heat treated using THE, shown in figure 21C and figure 21D, shows the same ten-

dency as the skim milk samples heated using PHE. One major difference though, is that THE skim milk sam-

ple has visible κ-casein bands in the non-reduced sample but the same intensity pattern for whey protein

bands. This indicates that the complexes made of THE skim milk contain less κ-casein but the same amount

of whey protein, as the protein content is the two samples are similar, shown in table 3, and it is therefore

reasonable to believe that β-Lg forms aggregates with other β-Lg proteins.

4.7.3 Size exclusion chromatography

Size exclusion chromatography was performed on milk samples in reduced and non-reduced form to inves-

tigate the amount of aggregates formed in each sample. Figure 22 shows the chromatograms of the re-

duced and non-reduced samples. The reduced samples in figure 22A results in peaks observed at retention

time 7-9 min for all skim milk samples. The MCI samples have very low intensity of this peak indicating that

there are some aggregates in this area containing whey proteins which are not S-S bound.

For all samples under non-reduced conditions samples shown in figure 22B, large peaks are observed with

retention times of 4.5-7 min. This indicates a high amount of disulphide bound aggregates which contain

caseins are present in this area. Compared to the reduced samples, no peaks at retention time 12-14 min

can be observed in the non-reduced samples.

The control skim milk has a higher overall absorbance in the non-reduced form, figure 22B, compared to

the heat treated samples which could indicate that there are even larger aggregates present in the heat

treated milk which cannot be detected by the used column.

Comparing the three heat treated skim milk samples, it is seen that THE has larger amount of the large ag-

gregates and also a lower amount of the intermediate size proteins for reducing and non-reducing condi-

tions while the amount of small peptides are equal for all heating methods. DSI has more intermediate

aggregates, retention time 7-9 min, compared heat treatment using PHE and THE.

- 49 -

4.7.3.1 Identification of aggregates

For two of the samples heat treated using PHE and THE at 130 °C for 5 s, the large aggregates were col-

lected and analysed to identify the content of the aggregates containing disulphide bonds.

Figure 23 shows the chromatogram from SEC analysis. The total protein content measured on Milkoscan

was 3.64% and 3.66% for the heat treated sample using THE and PHE, respectively.

The chromatograms for the two samples show similar pattern, but differences in intensities. Heat treat-

ment using PHE had higher absorbance and more protein is thereby detected by the use of the specific

column compared to heat treatment using THE. This could indicate that there is a higher content of large

protein complexes that cannot be detected with the used column for heat treatment using THE, as the total

protein content in the two samples are similar. The used column separates molecules in the weight range

from 5kDa to 1200 kDa. From the results obtained from 1DGE shown in section 4.7.1, could it be concluded

that protein aggregates above 300 kDa were present in heat treated milk. Results presented in this section

show that protein aggregates exceeding a molecular mass 1200 kDa are present.

Figure 22. Chromatograms for milk samples analyzed with SEC under reduced and non-reduced conditions. A: Reduced

milk samples analyzed on SEC. B: non-reduced milk samples analyzed on SEC.

- 50 -

Figure 24. Total protein analysis of skim milk fractions collected at SEC. Peak 1-3: κ-casein, peak 4: αs2-

casein, peak 5: αs1-casein, peak 6: β-casein, peak 7 : α-la, peak 8: β-Lg B, peak 9: β-Lg A.

Figure 23. Chromatograms for skim milk samples analysed with SEC in non-reduced

conditions. The peak area from 8-12 min was collected into four fragments for

further analysis.

- 51 -

The protein aggregates appeared on the chromatogram with RT 8-12 min were analysed on LC-MC.

Figure 24 shows the chromatogram of UV detection from LC analysis of the four fractions collected for both

heat treatments. The four fractions all show similar pattern but are of different intensities. This is caused by

the amount of protein in each fragment is not equal due to differences in intensity of the SEC analysis of

which the four fragments were collected from. Comparing the chromatograms of the two heating methods,

there were found variations in the protein content in the aggregates. The aggregates formed by heat

treatment using PHE, contained large amounts of glycosylated κ-casein (peak 1-3), β-casein (peak 6) and

whey proteins (peak 7-9). The aggregates formed by heat treatment using THE, contained large amount of

αs1-casein (peak 5) while almost no κ-casein and whey proteins are present compared to heating using PHE.

- 52 -

4.8 Kinetics of denaturation of whey proteins

The effect of heating temperatures and holding times were investigated to determine the rate of denatura-

tion of β-Lg and α-La. The order of reaction used for β-Lg was 1.5 and for α-La it was 1 according to previ-

ous studies (Dannenberg and Kessler, 1988; Kessler and Beyer, 1991; Oldfield et al., 1998a; Zúñiga et al.,

2010), using the rate equations

, for n = 1.5

and

, for n = 1

Figure 25. Denaturation degree of β-Lg B and α-la for skim milk samples heat treated using PHE at temperatures from

80 °C to 140°C at various holding times. For each temperature, a linear regression is fitted and these data are shown in

table 5. A: denaturation of β-Lg with a reaction order of 1.5. B: Denaturation of α-La with a reaction order of 1.

- 53 -

Figure 25A shows graphical representation of the denaturation of β-Lg B and figure 25B shows denatura-

tion of α-La for skim milk samples heated at PHE. For each temperature, the best fitted straight line was

plotted. This line indicates the kinetic fit which is used to calculate the rate constants shown in table 6. β-Lg

A follows same pattern as β-Lg B and is therefore not shown graphically.

The rate constant k was calculated from the slope of regression from the best fitted straight line for each

temperature. The slope is , for a reaction order of 1.5 and –k, for a reaction order of 1.

Table 6. Rate ate constant k and correlation coefficient of reaction kinetics on denaturation of β-Lg B, β-Lg A and

α-La in skim milk heated at PHE and THE. The values for β-Lg B and α-La heated at PHE are obtained from figure

25. Values for β-Lg A heated with PHE and all values heated at THE are obtained from graphical analysis, which are

not shown.

Tempera-ture (°C)

k 103 (S-1

) R2 k 103

(S-1)

R2

PHE THE β-Lg B n = 1.5 80 1.64 0.90 3.00 0.94

85 9.39 0.88

95 47.08 0.99 89.46 0.93 105 117.18 0.94

115 160.52 0.93

130 213.96 0.86 226.06 0.80

140 396.08 0.90

β-Lg A

n = 1.5 80 1.29 0.85 2.23 0.95

85 5.28 0.99

95 33.70 0.99 55.07 0.93

105 89.97 0,95

115 162.93 0.91

130 251.16 0.91 304.62 0.90

140 454.68 0.94

α-la

n = 1 80 0.77 0.84 0.71 0.90

85 1.92 0.83

95 2.40 0.93 6.82 0.83

105 7.87 0.92

115 9.72 0.94

130 27.75 0.84 28.13 0.89

140 42.08 0.85

- 54 -

Table shows the achieved data from linear regression and the correlation coefficient for each temperature

of β-Lg B and α-La from figure 25The values for β-Lg A heated at PHE and all values for heat treatment using

THE were obtained in similar way. The reaction order of 1.5 for β-Lg B and β-Lg A and reaction order of 1

for α-La obtained good correlation for all measured temperatures, with R2 from 0.80-0.99. Comparing the

reaction constants for heating at PHE and THE, it is observed that the reaction constant is larger for β-Lg B

and β-Lg A when heating using THE which indicates that the denaturation of β-Lg B and β-Lg A is faster

when heating using THE.

The obtained rate constants were plotted against the reciprocal of the absolute temperature. Figure 26

shows the effect of temperature on the rate constant of denaturation of β-Lg B, β-Lg A and α-La for skim

milk samples heated at PHE.

For β-Lg B and β-Lg A heated at PHE, it is possible to make linear regression in the temperature range from

80 °C to 95 °C and again from 95 °C to 140 °C. For α-La, this break is found at 85 °C. From the linear regres-

sion of each temperature range, the activation energy is calculated from the Arrhenius equation

which is shown in table 7.

Figure 26. The effect of temperature on rate constant for denaturation of β-Lg B, β-Lg A and α-La for skim milk heated at PHE.

Linear regressions are made by fitting rate constants for heating at PHE.

- 55 -

The activation energy for β-Lg at heating temperatures below 95 °C obtained activation energies of 250

,

while heating temperatures above 95 °C obtained activation energies of 55-72

.similar shift in activation

energy is observed for α-La, in the temperature ranges 80-85 °C and 85-140 °C.

As can be seen in figure 26, the measured rate constants are fitted to a linear regression, which is stated in

table 7. The calculation of reaction kinetics for α-La at the low temperature range, only two rate constants

were available and thereby could the uncertainty of the fit not be given. It was not possible to calculate the

reaction kinetics for heat treatment using THE as only three heating temperatures were used and the ob-

tained kinetic result would be very uncertain.

Table 7. Reaction kinetic for denaturation of β-Lg B, β-Lg A and α-La for skim milk heated at PHE. The values

are calculated from data obtained from figure 26.

Order (n) Temperature

range °C

ln(k0) Ea

R2

β-Lg B 1.,5 80-95 73.25 247.66 0.96

95-110 14.42 56.14 0.94

β-Lg A 1.,5 80-95 72.05 245.07 0.99

95-140 19.03 72.19 0.96

α-La 1 80-85 58.11 203.57 1

85-140 18.01 77.15 0.97

- 56 -

5 Discussion In this study, three different types of heat treatments were performed on milk with various whey protein

and casein content, and these were examined for coagulation properties, whey protein denaturation and

formation of heat induced aggregates.

The denaturation degree of whey proteins in skim milk increased with increasing holding time for all inves-

tigated temperatures. β-Lg B showed a higher degree of denaturation compared to β-Lg A, while α-La had

the least denaturation at all measured temperature and holding time combinations and heating methods.

This is consistent with theory, as β-Lg A has a slightly lower denaturation temperature, but is less reactive

due to a higher negative charge compared to β-Lg (O’Connell and Fox, 2011). α-La has a tendency to reform

into native structure at low temperatures and it higher temperatures are required to form aggregates with

other proteins which can explain the lower degree of denaturation observed for α-La.

The reaction kinetics of denaturation of whey proteins was extensively investigated by Dannenberg and

Kessler (1988) and their results are widely accepted and commonly used as reference for the effect of heat

treatment on whey protein denaturation. Since the publication of their study, various different heating

systems and other analytical methods have been used to analyse the denaturation degrees of whey pro-

tein.

Figure 27. Effect of heat treatment with PHE on the denaturation of β-Lg B in skim milk. The lines represent the calcu-lated rate of denaturation measured by Dannenberg and Kessler (1988) and each point represents the measured dena-turation degrees in the present study.

- 57 -

From an industrial perspective, it is of great interest to know how the denaturation degree of whey pro-

teins is affected in pilot scale heating systems. This allows mimicking the industrial used heating systems

and thereby enables the transfer of knowledge achieved in the laboratory to the production sites.

Figure 27 shows a comparison of obtained denaturation degrees of β-Lg B of skim milk heated at PHE from

the present study and the denaturation degrees observed by Dannenberg and Kessler (1988). From

figure 27 it observed that the denaturation degrees obtained in the present study for heating temperatures

below 90 °C show similar tendencies as the results by Dannenberg and Kessler (1988). Heating at tempera-

tures of 90 °C and above, the degree of β-Lg B differs substantially. Heating at a temperature of 140 °C for

5s gave a denaturation degree of 94 % for β-lg B in the present study, while Dannenberg and Kessler (1988)

only achieved a denaturation degree slight above 60 %. These variations were also found when comparing

heating using THE in the present study compared to Dannenberg and Kessler (1988). The same tendency

can be found comparing denaturation degrees of β-lg A and α-la with results from Dannenberg and Kessler

(1988).

The calculated kinetic parameters for denaturation of whey proteins in skim milk, presented in section 4.8,

were calculated on the basis of previous findings and compared with these (Dannenberg and Kessler, 1988;

Kessler and Beyer, 1991; Oldfield et al., 1998a; Zúñiga et al., 2010). The reaction order for β-Lg of 1.5 and 1

for α-La gave good correlation for the calculation of the rate constant, with R2 ranging from 0.8 and 0.99 for

all temperatures measured. All previous studies and the present study show similar patterns in the activa-

tion energies, with large activation energy for temperatures below 95 °C and low activation energies above

95 °C. The denaturation degree and reaction kinetics detected are observed to vary according to heating

system and analytical method used and also according to the milk type used (Anema and McKenna, 1996;

Dannenberg and Kessler, 1988; Oldfield et al., 1998a; Tolkach and Kulozik, 2007). The activation energy

found in the present study was lower at temperatures below 95 °C and higher for temperatures above

95 °C compared to previous studies (Corredig and Dalgleish, 1996b; Donato et al., 2007; Singh and Latham,

1993). This implies that less energy is required to unfold the whey proteins; while more energy is needed to

form aggregates containing these unfolded whey proteins.

The variations in degree of denaturation of whey proteins can be caused by differences in the heating sys-

tems used. A preheating period is often not used in existing publications in the area of research and often

small amount of skim was heated in tubes in water baths. Dannenberg and Kessler (1988) used a very small

pilot plant tubular heat exchanger with extremely short heating and cooling time and very small tubes for

heating (Dannenberg and Kessler, 1986).

The heating system in the present study contained a preheating step with preheating temperatures of 60 °C

or 75 °C which was estimated to be reached within the first minute of heat treatment. Based on the results

- 58 -

obtained by Dannenberg and Kessler (1988) approximately 5 % of the total β-lg content was denatured

during the preheating period. This means that the preheating process contributed to the denaturation of

whey proteins, but it could not explain the large variations in denaturation when heating at temperatures

above 90 °C.

By using a preheating section and three cooling sections, the total heating profile and heating period is

prolonged compared to previous studies. In this project, a small pilot plant heat exchanger was used which

is similar to a large industrial scale heat exchanger system. The heating and cooling period for used in the

present study was longer than the holding time at a desired temperature and this plays a significant role in

the denaturation. The overall heating profile is thereby larger compared to Dannenberg and Kessler (1988)

and which could affect the milk properties.

The desired temperatures in the present study were reached within approximately 1.5 min, while Dannen-

berg and Kessler (1988) reached the desired temperature in 0.3s. This short heating profile could favour

refolding of denatured protein in a non-native structure, which can be difficult to separate from non-

denatured whey proteins (Oldfield et al., 1998b). Aggregation of denatured whey proteins with other pro-

teins is expected to be more pronounced by using a larger heating profile and longer total heating time. An

important factor to take into account is the size of the heating system. The results made on small pilot sys-

tems could be difficult to reproduce in large scale production, as the heating process will for example have

different flow rates, heat transfer rates and turbulence. It could be argued that the heat transfer and heat

distribution is better in the present study due to higher milk flow and corrugations of the plates and tubes.

Another cause of differences in denaturation degrees are the methods for analysing the denaturation de-

gree of the whey proteins. Zúñiga et al. (2010) used SDS-Page and HPLC for analysing β-Lg denaturation and

detected a lower degree of denaturation when analysing heated β-Lg dispersions. Dannenberg and Kessler

(1988) used gel electrophoresis with isoelectric focusing which also resulted in lower denaturation degrees,

as shown in figure 27. This could be caused by gel electrophoresis providing a more limited separation than

HPLC, as denatured whey protein with closely related structures and no change in charge will migrate to

the same extent on the gel and will not be separated. The same insensitivity can be found for HPLC analysis

if there is no change in hydrophobicity if the whey proteins are denatured.

Furthermore, the integration and quantification of protein bands on gel electrophoresis can be difficult if

the protein concentrations are either high due to broadening of the protein band and interference with

other protein bands, or too low at which the background noise can be hard to separate from the protein

bands. This can be solved by using various concentrations of proteins in each sample. This was also neces-

sary in the present analysis of whey protein denaturation. For Control milk samples and milk samples which

were heated at temperatures below 100 °C, a lower injection volume was used due to the amount of pH

- 59 -

4.5 soluble proteins was higher in these samples. The integration of the peaks from these samples had to

be adjusted to be able to compare all samples. This means that small variations found for integration of

peak areas can lead to large variations when multiplying. These problems could also have affected the re-

sults of previous studies since the protein content changes in the soluble phase at pH 4.6 when the milk has

received heat treatment.

One major advantage of LC compared to gel electrophoresis are the reduced working hours, due to auto

sampling, auto cleaning and also the data analysis afterwards.

The effect of heat treatment of milk and how this affects rennet coagulation properties have been investi-

gated (Anema et al., 2011; Blecker et al., 2012; Waungana et al., 1996b) It was found that a denaturation

degree below 60 % did not affect the rennet coagulation time, while curd firming rates were much more

sensitive to denaturation. The present results are in agreement with these results. In the present study,

rennet coagulation properties for skim milk samples was impaired as the holding time increased for all

temperatures and all three heating methods. The rennet coagulation properties of skim milk did not show

any significant difference between the methods at temperatures for heat treatment using PHE and THE

below 100 °C. At temperatures above 100 °C, heating using THE resulted in worse rennet coagulation prop-

erties compared to PHE. It is generally accepted that at temperatures above 100 °C, most whey proteins are

instantly unfolded and the limiting step of irreversible denaturation is the aggregation of unfolded whey

proteins. At temperatures below100 °C, while it is the unfolding step that is the limiting factor shown by a

large activation energy (Dannenberg and Kessler, 1988; Oldfield et al., 1998a; Tolkach and Kulozik, 2007). It

can be speculated that heat treatment of skim milk using THE gives larger conformation changes of caseins

which could lead to heat induced aggregates of whey proteins not attached to casein in the casein micelle,

but with free caseins, and these complexes are therefore not as big a hindrance for rennet cleavage of

κ-casein on the casein micelle structure (Anema et al., 2007).

As can be seen in the analysis on composition of aggregates shown in section 4.7.3, it becomes clear that

there are great differences in the aggregates formed with heating using PHE and THE. Heating using THE

resulted in large protein complexes with a size above 1200 kDa which were not possible to detect on the

current method used in this study, while heating using PHE resulted in a larger content of smaller com-

plexes that were possible to detect with the used column. As the complexes found in skim milk heated us-

ing THE did not contain large amounts of whey protein and κ-casein and the pH 4,5 soluble analysis showed

low content of native whey proteins, it could be argued that the whey proteins are bound in large com-

plexes exceeding 1200 kDa. This is supported by the findings of Guyomarc’h et al., (2003), who investigated

protein aggregation in reconstituted skim milk and found whey protein-κ-casein complexes in the size

- 60 -

range of 3500-5000 kDa.

In order to obtain knowledge on the size and content of the heat induced complexes formed by heating

using THE, a SEC column which can separate even larger aggregates would be preferable. Furthermore, the

amount of protein in each collected fraction was fairly low (section 4.7.3.1) and it would have been a good

idea to collect a larger amount of each fraction to be able to obtain a better LC analysis of these.

The RCT and denaturation degrees of the differently heat treated milks were compared. As shown in

figure 15, only slight changes in RCT were seen for denaturation degrees up to 60 % when using indirect

heating. This is consistent with previous studies (Singh and Waungana, 2001; Waungana et al., 1996a). This

indicates that when more than 50 % of the whey proteins are denatured, the amount of whey protein

bound to the casein micelle is high enough to result in a steric hindrance for rennet cleavage. The correla-

tion between rennet coagulation time and denaturation degree is not similar for the three heating meth-

ods. Denaturation degrees above 70 % resulted in lower increase in relative RCT was found for heat treat-

ments using THE compared to PHE with the same denaturation degree. It seems like it is not only the de-

gree of whey protein denaturation that is important, but also how these denatured proteins interact with

other proteins.

Heat treatment of skim milk is shown to lead to formation of aggregates containing k-casein and whey pro-

teins, which is also shown by various authors (Graveland-Bikker and Anema, 2003; Tran Le et al., 2008). The

unfolding of β-lg exposes the reactive thiol groups which can bind the para κ-casein region of κ-casein

through disulphide linkages (Jean et al., 2006). As the formation of these complexes are altered by a large

heating profile with slow heating, the formation of the heat induced aggregates would be expected to be

more pronounced in heat treatments using PHE and THE compared to DSI. Dalgleish (1990) observed an

increase in size of heat induced aggregates as the denaturation of whey proteins increased. Formation of

large whey protein complexes could also in some extent be bound to the casein micelle in a different way

or not bound at all due to size, and they are thereby thought to give less steric hindrance while smaller

complexes might bind easier to the casein micelle.

The attachment of these heat induced aggregates on the casein micellar surface creates large steric hin-

drance for rennet induced cleavage. When the RCT is eventually reached for high heat treatment, the curd

firming process was slow for milks with a high denaturation degree. This could be caused by the attach-

ment of β-Lg to the k-casein which occurs on the para κ-casein region and this complex there stays on the

casein micelle after cleavage (Anema et al., 2007, 2011).

Differences in whey protein denaturation and formation of aggregates are caused by differences in the

heating method. The heating profile for heating using THE and PHE are similar when the same milk flow

- 61 -

was used, with THE having a slight longer preheating and cooling period. The heat transfer is expected to be

better in PHE due to corrugation of the plates which gives greater turbulence. The twisted shape of the

tubes decreases the heat transfer and turbulence and thereby resulted in a more unequal heat distribution,

compared to heating using PHE. These observations in differences in heating methods are consistent with

previous literature (Deeth and Datta, 2002). This could indicate that the high heat transfer in PHE and

slightly shorter heating profile leads to formation of smaller aggregates compared to THE. Fouling was ob-

served when the UHT plant was using PHE at high temperatures for several hours without cleaning, which

was not seen when heating with THE. Fouling on the surface of the plates mainly consists of whey protein

aggregates and calcium phosphate particles (Visser and Jeurnink, 1997). Formation of a fouling deposit on

the surface of the plates thereby binds denatured whey proteins which are not present in the milk any

longer. Although fouling was no large problem in this study, it could be expected that this could have an

effect on the content of denatured whey protein and thereby also an effect on the content of whey protein

that can form large aggregates.

The DSI heating system was found to be very different from the two indirect heating methods. DSI has

shown to have a significant lower denaturation degree of β-Lg and α-La and the rennet coagulation proper-

ties were also significantly less affected by the heat treatment compared to indirect heating. The heating

profile of DSI is different from the two indirect methods by having a very short heating time from preheat

temperature to the desired temperature and a flash cooling to 65 °C before reaching the three plate cool-

ing sections. The injection of steam into the preheated milk gives a good heat transfer due to mixing of milk

and steam, but this also induces more stress to the milk. This can be seen when heating at temperatures of

145 °C, where DSI had a low denaturation degree compared to skim milk samples heated using PHE and

THE, but this gave rise to a three times higher relative RCT compared to samples heated indirectly with

same denaturation degree. The increase in relative RCT can thereby not be the only explained by the dena-

turation of whey proteins, but the temperature gives rise to other changes milk, such as mineral precipita-

tion and lactose degradations (Lewis and Deeth, 2008).

No whey protein denaturation was observed upon heating skim milk at 105 °C for 4s using DSI. As the pre

heating and cooling sections are PHE, it could be expected that whey protein denatured to some extent.

The fast heating and flash cooling times gives less time for formation of aggregates and denatured whey

proteins could refold into non-native structures as most whey protein are instantly unfolded at tempera-

tures above 100 °C.

The flow rate of milk through the system had a great impact on the heating profile in the same extent as

the length of the holding section. The general flow rate in this project was 20 L/h. Heating using PHE and

THE, with holding times of 10 s and 30 s, operated with flow rates of 30L/h and 40 L/h, respectively. This

- 62 -

gave a faster heating and cooling time and thereby a different total heating profile, even though the hold-

ing time at the desired heating temperature was increased. These variations in heating profile could affect

the whey protein denaturation. This can in some extent be seen in figure 11, where the denaturation of β-

lg B and α-la, heated at temperatures of 95 °C or below, show a decrease in denaturation compared to the

heat treatment with a holding time of 5s. This can also in some extent explain why the relative RCT for the

three milk types varies at these holding times. This is shown in figure 12. Holding times of 30 s results in

decrease in the relative RCT while the relative CFR increases compared to heating with a holding time of 10

s and 60s.

Variations in whey protein denaturation and rennet coagulation properties between skim milk samples

with the same heat treatment combination and heating method were found between different trial days.

This was caused by variations in milk batches on various trail days. Even though there were only small varia-

tions in the milk composition analysis of total protein on LC-MS and Milkoscan, Law and Leaver (1997) have

shown that the level of denaturation decreases as the total protein content decreases and reverse. The

ratio between casein and whey protein also has a great impact on the denaturation of whey protein. Re-

moving casein from the milk and keeping the whey protein content constant, less denaturation of whey

protein was found, but these whey proteins formed large aggregates in the milk as they could not aggre-

gate with caseins in the casein micelle.

The content of casein and whey protein and the ratio between these is also important for coagulation

properties. The casein:whey protein ratio of 80:20 in skim milk had a prolonged RCT and lower CFR com-

pared to MCIc and MCI with a casein:α-la, β-lg ratio of 96:4. This was seen for all samples, both control and

heat treated samples. Skim milk and MCIc had the same casein content, with MCIc having lower total pro-

tein content. Still, MCIc showed slightly better coagulation properties when comparing control samples,

which indicates that whey proteins had an impact on CFR, even on low pasteurized (72 °C, 15s) skim milk.

Since 85 % of all κ-casein has to be hydrolysed to initiate casein micelle aggregation, it could be thought

that MCI would have longer RCT as the content of casein is higher and thereby more κ-casein has to be

cleaved. This was not seen and the results therefore indicate that the amount of Chymosin added to the

milk in these analyses was enough even though the casein content is 13 % higher in MCI compared to skim

milk. Removal of whey proteins from low pasteurized skim milk and increasing the casein content gave

significant improvements of the rennet coagulation properties.

The coagulation properties for heat treated MCI by the indirect heating methods showed the opposite of

heating of skim milk. For skim milk, heat treatment using THE improved the rennet coagulation properties

than PHE at high temperatures to the milk, but the opposite was found for heating of MCI. For tempera-

- 63 -

tures below 100 °C, PHE was better for both milk types. This could indicate that the heating system of THE

with larger flow area and lower heat transfer results in larger changes of the casein micelles compared to

heat treatment using PHE which is seen to affect the rennet coagulation properties. In order to make a

proper conclusion, more heat treatment combinations using THE are necessary. Analysis of the aggregate

content in heat treated MCI samples have to be analysed for all heat treatment systems, which could give

indications of which heat induced casein aggregates are formed by heat treatment of MCI using the differ-

ent heating systems.

The rennet coagulation properties of MCI and MCIc change during heat treatment which indicates that the

heat treatment results in structural changes of the casein micelle, even though the caseins and casein mi-

celle are stated to be quite heat stable. The CFR was increased when heating MCI and MCIc at tempera-

tures below 100 °C. Heat treatment above 100 °C was shown to have a negative effect on the coagulation

properties. As only 4 % of the total protein content was β-Lg and α-La, these are not the only explanations

for the decrease in coagulation properties. Heat treatments above 100 °C increases the amount of soluble

casein in solution and k-casein can be removed from the micellar surface at normal pH. The amount of

soluble phosphate and calcium decreases when increasing temperature (O’Connell and Fox, 2003; Sauer

and Moraru, 2012). This exposes the calcium sensitive caseins in casein micelle which likely could form

small aggregates. These aggregates have structures different from the casein micelle and could be less af-

fected by rennet cleavage. Mohammad and Fox (1987) also reported precipitation of calcium phosphate on

the surface layer of casein micelles, which hinders the rennet coagulation when heating at 140 °C.

Bulca and Kulozik, (2004) found crosslinking between casein and also dissociation of caseins from the mi-

celle when heating whey protein free casein milk solutions at high temperatures.

The degradation of lactose and binding of lactose to casein occurs in a higher degree than maillard reac-

tions. The binding of lactose to caseins on the casein micellar structure can form steric hindrance for rennet

cleavage of κ-casein. Furthermore is degradation of lactose responsible for formations of formic acids

which can give small decreases the pH in milk and reduce the micellar calcium phosphate (Martinez-Castro

et al., 1986; Turner et al., 1978).

These structural changes also appear in skim milk, as the only difference between the milks used is the

protein content, and this can help explaining why high heated treated skim milk with equal denaturation

degrees have variations in coagulation properties, even though the whey protein denaturation and forma-

tion of heat induced aggregates containing these are the main cause of impairments of rennet coagulation

properties of skim milk.

- 64 -

6 Conclusion The aim of this study was to investigate the effect of three different heating systems used in the dairy in-

dustry for heat treatment of milk, and how denaturation of whey proteins and rennet coagulation proper-

ties of skim milk were affected by the heating process. Further on, the effect of heat treatment on rennet

coagulation properties were analysed in skim milk with reduced content of whey proteins and various ca-

sein concentrations. The content of β-Lg and α-La in MCI and MCIc was less than 4 % of the total protein

content, and it was therefore not possible to make proper calculations on the denaturation of an already

small content of whey proteins.

The degree of denaturation of whey proteins in skim milk increased when increasing holding time for all

temperatures measured and heating methods used. Heat treatment using indirect heating at temperatures

of 115 °C or more resulted in denaturation degrees above 90 % for β-Lg for all holding times, while the de-

naturation degree of α-La did not exceed 90 % at any temperature and holding time combination. Heating

using THE gave rise to significantly higher degrees of denaturation at temperatures below 100 °C, com-

pared to PHE, while when heating at 130 °C, most β-Lg was denatured in both indirect heating methods.

Heating using DSI is gentle when comparing the denaturation degree with the indirect heating methods. A

denaturation degree of 40 % was found for β-Lg B when heating at 145 °C, while the indirect methods

showed denaturation degree of >95 % when heating at 140 °C.

The reaction kinetics of denaturation of whey proteins in skim milk was found to follow reaction order 1.5

for β-Lg and 1 for α-La. The denaturation degrees in this study were found to be higher than what was

found in previous studies, and the activation energy was found to be lower at temperatures below 95 °C

and higher at temperatures above 95 °C. This was caused by using a different heating system and a

generally longer and larger heating profile which gives rise to a larger extent of denaturation.

The rennet coagulation properties of skim were impaired when increasing holding time for all temperatures

measured for all heating methods, which is linked to the whey protein denaturation. An increase in degree

of denaturation gave impairments in rennet coagulation properties. Heating skim milk using DSI gave sig-

nificantly lower decrease in rennet coagulation properties compared to the indirect heating methods, and

RCT was reached within two hours when heating at 145 °C. Heating at PHE gave better rennet coagulation

properties when heating at temperatures below 100 °C, compared to THE, while THE reached RCT within

two hours at temperatures of 130 °C which heating at PHE did not. The removal of whey proteins from milk

improved the rennet coagulation properties compared to skimmed milk. MCI, with casein content equal to

total protein content of skim milk, showed the largest improvements in rennet coagulation properties at all

temperatures, but MCIc, with a casein content equal to skim milk, was also significantly better than skim

milk. When heating using PHE with temperatures below 100 °C with holding times below 60 s, the RCT and

- 65 -

CFR were improved compared to the control milk for MCI and MCIc. Rennet coagulation properties de-

creased when heating above 100 °C, but RCT and CFR were detected for all measured temperature and

holding time combinations. Heating MCI using DSI was also a gentle heating method compared to the indi-

rect heating. Heating using PHE shows a tendency towards improved rennet coagulation properties com-

pared to THE.

To investigate why THE and PHE showed similar degree of denaturation when heating at 130 °C, in spite of

THE having significantly better rennet coagulation properties, the content of heat induced aggregates in

skim milk were analysed. THE caused formation of a high level of large whey protein and κ-casein aggre-

gates, the sizes of which was larger than the column could detect, while heating using PHE resulted in lar-

ger amounts of smaller aggregates of whey protein and κ-casein, which could be detected.

- 66 -

7 Perspectives Heat treatments used in this project are performed on heating systems that are supposed to mimic the

large heating systems used in the dairy industry in the preheating and cooling sections. The results on how

the three heating methods affect milk proteins can thereby easily be transferred to large scale productions

without experiencing large changes in these parameters which can occur when upscaling productions.

Heat treatment using temperatures above those used in traditional pasteurization (72 °C for 15 s) is nor-

mally not used in cheese production due to the negative effect on whey protein denaturation, although

these temperatures are not sufficient to inactivate all undesirable microorganisms (Belitz et al., 2004). The

rennet coagulation properties of milk with low content of whey proteins, following high heat treatment can

be useful if the milk has to receive severe heat treatment to destroy higher contents of bacteria and spores

and still be able to make a hard or semi hard cheese.

The three methods used and their resulting differences in denaturation of whey protein and rennet coagu-

lation can be used in production of products with special characteristics which can be improved when using

a more specialized heating. This can be cheeses with certain hardness, incorporation of whey proteins in

the cheese without large change in functionality of the cheese, or decreasing the heating time in the proc-

essing of yogurt.

7.1 Future research Future research in this area could be to look more into the particle size of the heat induced protein aggre-

gates for various temperatures and also for the various milk types, to make a full picture of the content of

the aggregates. It could also be interesting to look into the casein micellar structure to see which changes

happen in the various milk types at the different heating methods and relate this with rennet coagulation.

Furthermore, more heat treatments could be made for heat treatment using THE and additionally also us-

ing direct steam injection for all milk types, to make at full picture of the effect of each heating system on

heat treatment of milk.

- 67 -

8 References Aguilar, M.S. (2004). HPLC of Peptides and Proteins: Methods and Protocols (Humana Press).

Andersen, P.N. (2013). ReoRox G2 and Rheometer comparison (Arla Foods Amba).

Anema, S.G. (2008). Chapter 8 - The whey proteins in milk: thermal denaturation, physical interactions and effects on the functional properties of milk. In Milk Proteins, A. Thompson, M. Boland, and H. Singh, eds. (San Diego: Academic Press), pp. 239–281.

Anema, S.G., and McKenna, A.B. (1996). Reaction Kinetics of Thermal Denaturation of Whey Proteins in Heated Reconstituted Whole Milk. J. Agric. Food Chem. 44, 422–428.

Anema, S.G., Kim Lee, S., and Klostermeyer, H. (2007). Effect of pH at heat treatment on the hydrolysis of κ-casein and the gelation of skim milk by chymosin. LWT - Food Sci. Technol. 40, 99–106.

Anema, S.G., Lee, S.K., and Klostermeyer, H. (2011). Rennet-Induced Aggregation of Heated pH-Adjusted Skim Milk. J. Agric. Food Chem. 59, 8413–8422.

Bansal, B., and Chen, X.D. (2006). A Critical Review of Milk Fouling in Heat Exchangers. Compr. Rev. Food Sci. Food Saf. 5, 27–33.

Belitz, H.D., Grosch, W., and Schieberle, P. (2004). Food Chemistry (Springer).

Benguigui, L., Emery, J., Durand, D., and Busnel, J., P. (1994). Ultrasonic study of milk clotting. Lait 74, 197–206.

Berg, J.M., Tymoczko, J.L., and Stryer, L. (2006). Biochemistry (New York: W.H.Freeman and Company).

Blecker, C., Habib-Jiwan, J.-M., and Karoui, R. (2012). Effect of heat treatment of rennet skim milk induced coagulation on the rheological properties and molecular structure determined by synchronous fluorescence spectroscopy and turbiscan. Food Chem. 135, 1809–1817.

Bohlin, L., Hegg, P.-O., and Ljusberg-Wahren, H. (1984). Viscoelastic Properties of Coagulating Milk. J. Dairy Sci. 67, 729–734.

Bonfatti, V., Grigoletto, L., Cecchinato, A., Gallo, L., and Carnier, P. (2008). Validation of a new reversed-phase high-performance liquid chromatography method for separation and quantification of bovine milk protein genetic variants. J. Chromatogr. A 1195, 101–106.

Boxler, C., Augustin, W., and Scholl, S. (2014). Composition of milk fouling deposits in a plate heat exchang-er under pulsed flow conditions. J. Food Eng. 121, 1–8.

Brew, K. (2003). α-lactalbumin. In Advanced Dairy Chemistry, Volume 1, Proteins, P.F. Fox, and P.L.H. McSweeney, eds. (Kluwer Academic/Plenum Plubishers),.

Bulca, S., and Kulozik, U. (2004). Heat-induced changes in native casein micelles obtained by microfiltration. Bull. IDF 389, 36–39.

Caffin, J.P., Poutrel, B., and Rainard, P. (1985). Physiological and Pathological Factors Influencing Bovine α-Lactalbumin and β-Lactoglobulin Concentrations in Milk. J. Dairy Sci. 68, 1087–1094.

- 68 -

Calvo, M.M., Leaver, J., and Banks, J.M. (1993). Influence of other whey proteins on the heat-induced ag-gregation of α-lactalbumin. Int. Dairy J. 3, 719–727.

Campell, M. (2013). Heating methods for Ultra-High temperature pasteurization (http://www.foodprocessing.com/assets/wp_downloads/PickHeaters_eHandbook_02-13.pdf).

Considine, T., and Flanagan, J. (2008). Chapter 13 - Interaction between milk proteins and micronutrients. In Milk Proteins, A. Thompson, M. Boland, and H. Singh, eds. (San Diego: Academic Press), pp. 377–407.

Considine, T., Patel, H.A., Anema, S.G., Singh, H., and Creamer, L.K. (2007). Interactions of milk proteins during heat and high hydrostatic pressure treatments — A Review. Innov. Food Sci. Emerg. Technol. 8, 1–23.

Corredig, M., and Dalgleish, D.G. (1996a). Effect of temperature and pH on the interactions of whey pro-teins with casein micelles in skim milk. Food Res. Int. 29, 49–55.

Corredig, M., and Dalgleish, D.G. (1996b). Effect of temperature and pH on the interactions of whey pro-teins with casein micelles in skim milk. Food Res. Int. 29, 49–55.

Croguennec, T., O’Kennedy, B.T., and Mehra, R. (2004). Heat-induced denaturation/aggregation of β-lactoglobulin A and B: kinetics of the first intermediates formed. Int. Dairy J. 14, 399–409.

Czerwenka, C., Maier, I., Pittner, F., and Lindner, W. (2006). Investigation of the Lactosylation of Whey Pro-teins by Liquid Chromatography−Mass Spectrometry. J. Agric. Food Chem. 54, 8874–8882.

Dalgleish, D.G. (1990). Denaturation and aggregation of serum proteins and caseins in heated milk. J. Agric. Food Chem. 38, 1995–1999.

Dalgleish, D.G. (1993). Bovine milk protein properties and the manufacturing quality of milk. Proc. Jt. EAAPASAS Workshop 35, 75–93.

Dannenberg, F., and Kessler, H.G. (1986). Quantitative densitometry of heat-induced changes in whey pro-teins following ultrathin-layer isoelectric focusing. ELECTROPHORESIS 7, 67–72.

Dannenberg, F., and Kessler, H.-G. (1988). Reaction Kinetics of the Denaturation of Whey Proteins in Milk. J. Food Sci. 53, 258–263.

Deeth, H.C., and Datta, N. (2002). ULTRA-HIGH TEMPERATURE TREATMENT (UHT) | Heating Systems. In Encyclopedia of Dairy Sciences, H. Roginski, ed. (Oxford: Elsevier), pp. 2642–2652.

Dissanayake, M., and Vasiljevic, T. (2009). Functional properties of whey proteins affected by heat treat-ment and hydrodynamic high-pressure shearing. J. Dairy Sci. 92, 1387–1397.

Dissanayake, M., Ramchandran, L., Donkor, O.N., and Vasiljevic, T. (2013b). Denaturation of whey proteins as a function of heat, pH and protein concentration. Int. Dairy J. 31, 93–99.

Ditchfield, C., Tadini, C.C., Singh, R.K., and Toledo, R.T. (2007). Heat transfer during thermal processing of a temperature dependent non-Newtonian fluid in a tubular heat exchanger. Adv. Appl. Chem. Eng. Princ. Food Ind. 46, 472–476.

- 69 -

Donato, L., Guyomarc’h, F., Amiot, S., and Dalgleish, D.G. (2007). Formation of whey protein/κ-casein com-plexes in heated milk: Preferential reaction of whey protein with κ-casein in the casein micelles. Int. Dairy J. 17, 1161–1167.

Edgerton, E.R., Jones, V.A., and Warren, J.A. (1970). Effect of Process Variables on the Holding Time in an Ultrahigh-Temperature Steam Injection System1. J. Dairy Sci. 53, 1353–1357.

Edmond, S.. (2001). Continious heat processing. In Thermal Technologies in Food Processing, P. Richardson, ed. (Elsevies), pp. 29–48.

Eigel, W.N., Butler, J.E., Ernstrom, C.A., Farrell, H.M., Jr., Harwalkar, V.R., Jenness, R., and Whitney, R.M. (1984). Nomenclature of Proteins of Cow’s Milk: Fifth Revision1. J. Dairy Sci. 67, 1599–1631.

Farrell Jr., H.M., Jimenez-Flores, R., Bleck, G.T., Brown, E.M., Butler, J.E., Creamer, L.K., Hicks, C.L., Hollar, C.M., Ng-Kwai-Hang, K.F., and Swaisgood, H.E. (2004). Nomenclature of the Proteins of Cows’ Milk—Sixth Revision. J. Dairy Sci. 87, 1641–1674.

Farrell Jr., H.M., Malin, E.L., Brown, E.M., and Qi, P.X. (2006). Casein micelle structure: What can be learned from milk synthesis and structural biology? Curr. Opin. Colloid Interface Sci. 11, 135–147.

Fox, P.F. (1980). Heat-Induced Changes in Milk Preceding Coagulation. J. Dairy Sci. 64, 2127–2137.

Fox, P.F., and Kelly, A.L. (2006). Chemistry and Biochemistry of Milk constituents. In In Food Biochemistry & Food Processing, (Blackwell Publishing),.

Fox, P.F., and McSweeney, P.L.. (1998). Dairy Chemistry and Biochemistry (London: Blackie Academic and Professional).

Frederiksen, P.D., Hammershøj, M., Bakman, M., Andersen, P.N., Andersen, J.B., Quist, K.B., and Larsen, L.B. (2011). Variations in coagulation properties of cheese milk from three Danish dairy breeds as determined by a new free oscillation rheometry-based method. Dairy Sci Technol 91, 309–321.

Gaucheron, F. (2011). MILK SALTS | Distribution and Analysis. In Encyclopedia of Dairy Sciences (Second Edition), J.W. Fuquay, ed. (San Diego: Academic Press), pp. 908–916.

Giacometti, J., and Josić, D. (2013). Chapter 7 - Protein and Peptide Separations. In Liquid Chromatography, S. Fanali, P.R. Haddad, C.F. Poole, P. Schoenmakers, and D. Lloyd, eds. (Amsterdam: Elsevier), pp. 149–184.

Ginger, M.R., and Grigor, M.R. (1999). Comparative aspects of milk caseins. Comp. Biochem. Physiol. B Biochem. Mol. Biol. 124, 133–145.

Graveland-Bikker, J.F., and Anema, S.G. (2003). Effect of individual whey proteins on the rheological proper-ties of acid gels prepared from heated skim milk. Int. Dairy J. 13, 401–408.

Gut, J.A.W., and Pinto, J.M. (2003). Modeling of plate heat exchangers with generalized configurations. Int. J. Heat Mass Transf. 46, 2571–2585.

Gutierrez, C.G.C.C., Dias, E.F.T.S., and Gut, J.A.W. (2011). Investigation of the residence time distribution in a plate heat exchanger with series and parallel arrangements using a non-ideal tracer detection technique. Appl. Therm. Eng. 31, 1725–1733.

- 70 -

Guyomarc’h, F., Law, A.J.R., and Dalgleish, D.G. (2003). Formation of Soluble and Micelle-Bound Protein Aggregates in Heated Milk. J. Agric. Food Chem. 51, 4652–4660.

Heck, J.M.L., van Valenberg, H.J.F., Dijkstra, J., and van Hooijdonk, A.C.M. (2009). Seasonal variation in the Dutch bovine raw milk composition. J. Dairy Sci. 92, 4745–4755.

Horne, D.S., and Davidson, C.M. (1990). The use of dynamic light-scattering in monitoring rennet curd for-mation. Milchwissenschaft 45, 712–715.

Hougaard, A.B., Ardö, Y., and Ipsen, R.H. (2010). Cheese made from instant infusion pasteurized milk: Ren-net coagulation, cheese composition, texture and ripening. Int. Dairy J. 20, 449–458.

Iametti, S., Cairoli, S., De Gregori, B., and Bonomi, F. (1995). Modifications of High-Order Structures upon Heating of .beta.-Lactoglobulin: Dependence on the Protein Concentration. J. Agric. Food Chem. 43, 53–58.

Iametti, S., De Gregori, B., Vecchio, G., and Bonomi, F. (1996). Modifications Occur at Different Structural Levels During the Heat Denaturation of β-Lactoglobulin. Eur. J. Biochem. 237, 106–112.

Jakob, E., and Puhan, Z. (1992). Technological properties of milk as influenced by genetic polymorphism of milk proteins — A review. Int. Dairy J. 2, 157–178.

Jean, K., Renan, M., Famelart, M.-H., and Guyomarc’h, F. (2006). Structure and surface properties of the serum heat-induced protein aggregates isolated from heated skim milk. Int. Dairy J. 16, 303–315.

Jensen, H.B., Holland, J.W., Poulsen, N.A., and Larsen, L.B. (2012a). Milk protein genetic variants and isoforms identified in bovine milk representing extremes in coagulation properties. J. Dairy Sci. 95, 2891–2903.

Jensen, H.B., Poulsen, N.A., Møller, H.S., Stensballe, A., and Larsen, L.B. (2012b). Comparative proteomic analysis of casein and whey as prepared by chymosin-induced separation, isoelectric precipitation or ultra-centrifugation. J. Dairy Recearch 79, 451–458.

Jones, A.J.S. (1993). Analysis of polypeptides and proteins. Protein Ther. Using Recomb. Technol. 10, 29–90.

Kang, D., Gho, Y.S., Suh, M., and Kang, C. (2002). Highly Sensitive and Fast Protein Detection with Coomassie Brilliant Blue in Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis. Bull. Korean Chem. Soc. 23, 1511–1512.

Kessler, H.-G., and Beyer, H.-J. (1991). Thermal denaturation of whey proteins and its effect in dairy tech-nology. Int. J. Biol. Macromol. 13, 165–173.

Abu-Khader, M.M. (2012). Plate heat exchangers: Recent advances. Renew. Sustain. Energy Rev. 16, 1883–1891.

Kontopidis, G., Holt, C., and Sawyer, L. (2004). Invited Review: β-Lactoglobulin: Binding Properties, Struc-ture, and Function. J. Dairy Sci. 87, 785–796.

Lacorte, S., and Fernandez-Alba, A.R. (2006). Time of flight mass spectrometry applied to the liquid chroma-tographic analysis of pesticides in water and food. Mass Spectrom. Rev. 25, 866–880.

- 71 -

Laemmli, U.K. (1970). Cleavage of Structural Proteins during the Assembly of the Head of Bacteriophage T4. Nature 227, 680–685.

Larsen, L., Wedholm-Pallas, A., Lindmark-Månsson, H., and Andrén, A. (2010). Different proteomic profiles of sweet whey and rennet casein obtained after preparation from raw versus heat-treated skimmed milk. Dairy Sci. Technol. 90, 641–656.

Law, A.J.R., and Leaver, J. (1997). Effect of Protein Concentration on Rates of Thermal Denaturation of Whey Proteins in Milk. J. Agric. Food Chem. 4255–4261.

Lewis, M.J., and Deeth, H.C. (2008). HeatTreatment of Milk. In Milk Processing and Quality Management, A.Y. Tamime, ed. pp. 168–204.

LORENZEN, P.C., CLAWIN-RÄDECKER, I., EINHOFF, K., HAMMER, P., HARTMANN, R., HOFFMANN, W., MAR-TIN, D., MOLKENTIN, J., WALTE, H.G., and DEVRESE, M. (2011). A survey of the quality of extended shelf life (ESL) milk in relation to HTST and UHT milk. Int. J. Dairy Technol. 64, 166–178.

Martinez-Castro, I., Olano, A., and Corzo, N. (1986). Modifications and interactions of lactose with mineral components of milk during heating processes. Food Chem. 21, 211–221.

McMahon, D.J., Yousif, B.H., and Kaláb, M. (1993). Effect of whey protein denaturation on structure of ca-sein micelles and their rennetability after ultra-high temperature processing of milk with or without ultrafil-tration. Int. Dairy J. 3, 239–256.

McSwiney, M., Singh, H., and Campanella, O.H. (1994). Thermal aggregation and gelation of bovine β-lactoglobulin. Food Hydrocoll. 8, 441–453.

Miranda, G., and Pelissier, J.. (1983). Kinetic studies of in vivo digestion of bovine unheated skim-milk pro-teins in the rat stomach. J. Dairy Recearch 50, 27–36.

Mohammad, K.S., and Fox, P.F. (1987). Heat-induced microstructural changes in casein micelles before and after heat coagulation. NZ J Dairy Sci Technol 22, 191–203.

Mozhaev, V.V., Heremans, K., Frank, J., Masson, P., and Balny, C. (1996). High pressure effects on protein structure and function. Proteins Struct. Funct. Bioinforma. 24, 81–91.

O’Connell, J.E., and Fox, P.F. (2011). Heat Treatment of Milk | Heat Stability of Milk. In Encyclopedia of Dairy Sciences (Second Edition), J.W. Fuquay, ed. (San Diego: Academic Press), pp. 744–749.

O’Connell, J.., and Fox, P.F. (2003). Heat-induced coagulation og milk. In Advanced Dairy Chemistry, Volume 1, Proteins, P.F. Fox, and P.. McSweeney, eds. (Kluwer Academic/Plenum Plubishers),.

O’Donnell, R., Holland, J.W., Deeth, H.C., and Alewood, P. (2004). Milk proteomics. Int. Dairy J. 14, 1013–1023.

Oldfield, D.J., Singh, H., Taylor, M.W., and Pearce, K.N. (1998a). Kinetics of Denaturation and Aggregation of Whey Proteins in Skim Milk Heated in an Ultra-high Temperature (UHT) Pilot Plant. Int. Dairy J. 8, 311–318.

Oldfield, D.J., Singh, H., and Taylor, M.W. (1998b). Association of β-Lactoglobulin and α-Lactalbumin with the Casein Micelles in Skim Milk Heated in an Ultra-high Temperature Plant. Int. Dairy J. 8, 765–770.

- 72 -

Oldfield, D.J., Singh, H., Taylor, M.W., and Pearce, K.N. (2000). Heat-induced interactions of β-lactoglobulin and α-lactalbumin with the casein micelle in pH-adjusted skim milk. Int. Dairy J. 10, 509–518.

Papiz, M.Z., Sawyer, L., Eliopoulos, E.E., North, A.C.T., Findlay, J.B.C., Sivaprasadarao, R., Jones, T.A., New-comer, M.E., and Kraulis, P.J. (1986). The structure of [beta]-lactoglobulin and its similarity to plasma reti-nol-binding protein. Nature 324, 383–385.

Pérez, M.D., and Calvo, M. (1995). Interaction of β-Lactoglobulin with Retinol and Fatty Acids and Its Role as a Possible Biological Function for This Protein: A Review. J. Dairy Sci. 78, 978–988.

Pierre, A., Frauquant, J., Le Gaët, Y., Poit, M., and Maubois, J.L. (1992). Préparation de phosphocaséinate natif par microfiltration sur membrane . Le Lait 72, 461–474.

Qi, X.L., Brownlow, S., Holt, C., and Sellers, P. (1995). Thermal denaturation of β-lactoglobulin: effect of protein concentration at pH 6.75 and 8.05. Biochim. Biophys. Acta BBA - Protein Struct. Mol. Enzymol. 1248, 43–49.

Rabilloud, T., Chevallet, M., Luche, S., and Lelong, C. (2010). Two-dimensional gel electrophoresis in prote-omics: Past, present and future. Model Org. Proteomics 73, 2064–2077.

Radola, B.J. (1984). [13] High-resolution preparative isoelectric focusing. In Methods in Enzymology, William B. Jakoby, ed. (Academic Press), pp. 256–275.

Reithel, F.J., and Kelly, M.J. (1971). Thermodynamic analysis of the monomer-dimer association of .beta.-lactoglobulin A at the isoelectric point. Biochemistry (Mosc.) 10, 2639–2644.

Sauer, A., and Moraru, C.I. (2012). Heat stability of micellar casein concentrates as affected by temperature and pH. J. Dairy Sci. 95, 6339–6350.

Schlüter, H. (1999). Chapter 3 Reversed-Phase Chromatography. In Journal of Chromatography Library, Michael Kastner, ed. (Elsevier), pp. 147–234.

Schmid, R.F. (2003). The untold story of milk: green pastures, contented cows and raw dairy foods (New Trends Pub.).

Schroyer, J.A. (1997). Understand the Basis of Steam Injection Heating. Chem. Eng. Process.

Shah, R.K., and Sekulić, D.P. (2007). Classification of Heat Exchangers. In Fundamentals of Heat Exchanger Design, (John Wiley & Sons, Inc.), pp. 1–77.

Singh, H. (2011). Milk Protein Products | Functional Properties of Milk Proteins. In Encyclopedia of Dairy Sciences (Second Edition), J.W. Fuquay, ed. (San Diego: Academic Press), pp. 887–893.

Singh, H., and Havea, P. (2003). Thermal denaturation, aggregation and gelation of whey proteins. In Ad-vanced Dairy Chemistry, Volume 1, Proteins, P.F. Fox, and P.L.H. McSweeney, eds. (Kluwer Academ-ic/Plenum Plubishers),.

Singh, H., and Latham, J.M. (1993). Heat stability of milk: Aggregation and dissociation of protein at ultra-high temperatures. Int. Dairy J. 3, 225–237.

- 73 -

Singh, H., and Waungana, A. (2001). Influence of heat treatment of milk on cheesemaking properties. Cheese Ripening Technol. 11, 543–551.

Souza, E.M.T. de, ARRUDA, S.F., BRANDÃO, P.O., and SIQUEIRA, E.M. de A. (2000). Electrophoretic analisys to detect and quantify additional whey in milk and dairy beverages.

Swaisgood, H.. (2003). Chemistry of the caseins. In Advanced Dairy Chemistry, Volume 1, Proteins, P.F. Fox, and P.L.H. McSweeney, eds. (Kluwer Academic/Plenum Plubishers), pp. 139–202.

Thulukkanam, K. (2013). Heat Exchanger Design Handbook (Taylor & Francis Inc.).

Tolkach, A., and Kulozik, U. (2007). Reaction kinetic pathway of reversible and irreversible thermal denatur-ation of beta-lactoglobulin. Lait 87, 301–315.

Tran Le, T., Saveyn, P., Hoa, H.D., and Van der Meeren, P. (2008). Determination of heat-induced effects on the particle size distribution of casein micelles by dynamic light scattering and nanoparticle tracking analy-sis. Int. Dairy J. 18, 1090–1096.

Turner, L.G., Swaisgood, H.E., and Hansen, A.P. (1978). Interaction of Lactose and Proteins of Skim Milk during Ultra-High-Temperature Processing. J. Dairy Sci. 61, 384–392.

Visser, J., and Jeurnink, T.J.M. (1997). Fouling of heat exchangers in the dairy industry. Heat Exch. Fouling 14, 407–424.

Walstra, P. (1990). On the Stability of Casein Micelles. J. Dairy Sci. 73, 1965–1979.

Walstra, P. (1999). Casein sub-micelles: do they exist? Int. Dairy J. 9, 189–192.

Wang, Q., Bulca, S., and Kulozik, U. (2007). A comparison of low-intensity ultrasound and oscillating rheolo-gy to assess the renneting properties of casein solutions after UHT heat pre-treatment. Int. Dairy J. 17, 50–58.

Waungana, A., Singh, H., and Bennett, R.J. (1996a). Influence of denaturation and aggregation of β-lactoglobulin on rennet coagulation properties of skim milk and ultrafiltered milk. Food Res. Int. 29, 715–721.

Waungana, A., Singh, H., and Bennett, R.J. (1996b). Influence of denaturation and aggregation of β-lactoglobulin on rennet coagulation properties of skim milk and ultrafiltered milk. Food Res. Int. 29, 715–721.

De Wit, J.N. (1998). Nutritional and Functional Characteristics of Whey Proteins in Food Products. J. Dairy Sci. 81, 597–608.

Zúñiga, R.N., Tolkach, A., Kulozik, U., and Aguilera, J.M. (2010). Kinetics of Formation and Physicochemical Characterization of Thermally-Induced β-Lactoglobulin Aggregates. J. Food Sci. 75, E261–E268.

- 74 -

- 75 -

Appendix 1 Denaturation degree of β-Lg A in heat treated skim milk at PHE and THE, shown for temperatures below

100 °C and above 100 °C.

- 76 -

- 77 -

Appendix 2 2DGE for milk samples under non-reduced and reduced conditions.

A: control skim milk, non-reduced. B: control skim milk, reduced. C: skim milk heated at DSI, 130 °C for 4s, non-reduced.

D: skim milk heated at DSI, 130 °C for 4s, reduced. E: control MCI, non-reduced. F: control MCI, reduced. G: MCI heated at

PHE 130 °C for 5s, non-reduced. H: MCI heated at PHE 130 °C for 5s, reduced.