«Lars Wiking Faculty of Natural Resources and Agricultural Sciences Department of Food Science Uppsala Doctoral thesis Swedish University of ...»
Triglycerides crystallise in different polymorphic forms (i.e. different crystal lattice types), which have different distances between layers of molecules. The three main polymorphic forms of triglycerides are α, β´ and β. All three types are found in milk fat (Walstra and van Beresteyn, 1975; Lopes et al. 2001) and 2005).
Each polymorphic form is characterized by its own melting point. This contributes to making the overall melting point of milk fat even more complex. The α and β´ forms are metastable. Furthermore, a low melting modification of the β´ forms, is called γ or sub-α form and is reported to have been found in milk (ten Grotenhuis et al. 1999 and Lopez et al. 2005). By using X-ray diffraction the different polymorphic forms can be monitored. Lopez et al. (2001) observed that the nucleation starts in α the polymorphic form at 18 °C by slow cooling (0.15 °C/min) of cream. At 9 °C the formation of the β´ forms begins, and the α + β´ polymorphic forms coexist until the end of cooling (–8 °C). In anhydrous milk fat, nucleation begins in β´ form and thereafter α + β´ polymorphic forms coexist 18 (Lopez et al 2001). By heating cream from –8 to 50°C, the dominant form between -8 to 5 °C is the α form (Lopez et al. 2000). Thereafter and up to 17.4 °C the dominant form is coexistence of α + β´ forms. After 15 °C and up to the final melting, the major form is β´. Similar results were found by heating anhydrous milk fat (Lopez et al. 2001). The transformation of α crystals into β´ crystals means that once β´ crystals are formed, triglycerides will dissolve from the α crystals and crystallise onto the β´ crystals. The cooling rate has an impact on the onset of crystalisation of milk fat, polymorphic crystal transitions and crystal size (Lopez et al. 2005).
Crystallisation of milk fat in MFGs occurs later than in a continuous milk fat phase (Söderberg et al. 1989). Buchheim (1970) showed by electron microscopy that crystallisation in MFGs starts predominantly with the high melting triglycerides at the inside of the membrane. In milk and cream the crystal growth is dependent on only a few catalytic impurities being available for starting the nucleation in every single fat globule and the crystals are disturbed by the curvature. Furthermore, the presence of phospholipids in the MFG affects the crystallisation. Vanhoutte et al. ( 2002a & 2002b) reported that the addition of small amounts of phospholipids into anhydrous milk fat delays the onset of crystallisation upon isotermal cooling at 25°C. They suggested that the phospholipid will be absorbed to the initial crystals due to their lower solubility in the melt and the absorbed phospholipids will block further growth of the crystals.
19 Objective The implementation of automatic milking systems initially caused a lowering of milk quality regarding total bacterial count, somatic cells count, freezing point of milk, antibiotic residues and FFA. During the recent years quality in terms of the parameters mentioned has generally improved, except FFA. A high level of FFA increases the risk of rancid flavour in dairy products.
The main objective of the present thesis was to evaluate different factors in automatic milking systems which could potentially cause increased level of FFA.
The stability of the MFG was described by using model systems.
The specific objectives of this thesis were:
to describe the factors affecting the size distribution of MFGs (paper I).
to evaluate the influence of the temperature of the milk and of feed on MFG stability during mechanical stress of raw milk (paper II and III).
to study the impact of increased milking frequencies of cows on the FFA content in milk (paper IV).
20 Material and methods Animals and feed Milk from Danish Holstein cows was used in paper I, II and III. Swedish Red and White cows were used in paper IV. In paper I milk was randomly collected from cows not administrated any experimental diet. In paper II and III the groups of cows in mid-lactation were offered the diets as shown in Table 5.
Pumping experiments The pumping system (Figure 3) used in paper II and III consisted of 9.5 m pipeline (diameter 2.25 cm), one valve, a balance tank and a centrifugal pump (Alfa-Laval, Sweden). The inlet pipe was fitted at the top of the balance tank. In each treatment, seven litres of raw milk were pumped through the system for 450 seconds. The flow rate was regulated by an frequency converter (ABB ACS 140, Denmark).
21 Figure 3. Picture of the pumping system used in paper II and III.
Milk fat quality Determination of size distribution of milk fat globules Particle size distributions were determined by integrated light scattering using a Mastersizer 2000 (Malvern Instruments Ltd, UK). The refractive indexes according to Michalski, Briard & Michel (2001) were used. The volume-based
is the number of globules in a size class of diameter di). The volume-based diameter, d(4,3) was used because it is more sensitive to the presence of coalescence of MFG.
Analysis of free fatty acids in milk In paper II and III the BDI-method (IDF, 1991) is used to determine the levels of FFA in milk. The advantages of this method are that it is used in many countries and is the most frequently used in combination with sensory tests. The disadvantage is that the recovery of the short-chain fatty acids is poor due to that these are dissolved in the aqueous phase during extraction (Duncan & Christen, 1991; IDF, 1991; Evers, 2003). The method involves an extraction of milk with the BDI reagent (Triton-x-100 and sodium tetraphosphate) to separate the fat. The released fat is dissolved in 2-propanol /ethanol and titrated with ethanolic KOH under nitrogen. Normally, thymol blue is used as indicator, but in the present studies an automatic titration (ABU 96 Tribuette, Radiometer, Denmark) was used. The level of FFA (acid degree value) is expressed as mmol KOH used to neutralize 100g fat.
22 In paper IV the content of FFA was determined by the Auto-Analyzer II method (Lindqvist, Roos & Fujita, 1975). The advantage of this method is that it is cheap and fast, 40 samples/h. Furthermore, the recovery of short chain fatty acids is higher compared with the BDI-method (IDF, 1991). The method is based on an extraction of the milk sample with a solution containing 2-propanol, heptane and 1 N H2SO4. In this method the lipase (is in the aqueous phase) and the lipid are separated, thus the sample can be stored for a week at room temperature without further lipolytic activity. In the auto-analyser, the solution is mixed with the indicator reagent (phenol red, sodium barbital and ethanol) and finally the absorbance is recorded at 560 nm in a colorimeter. These analyses were conducted by STEINS laboratories (Holstebro, Denmark).
Fatty acid composition The fatty composition was determined by gas chromatography (GC). Prior to GC separation and quantification, milk lipids were trans-esterified to methyl esters in a sodium methylate solution (2 g/l methanol). Analysis of the fatty acid methyl esters was carried out with a GC (6890 series, Hewlett-Packard Co., USA) using an FFAP-column (terephtalic acid modified polyethylene glycol 25m x 200 µm x
0.30 µm) (Hewlett-Packard Co., USA) and helium as carrier gas and a flame ionisation detector. Injection was splitless with an injector temperature of 250 °C.
The detector temperature was 300°C. The initial column temperature was 40 °C which was held for 4 min. The temperature was then raised by 10 °C /min to 240°C, and held there for 1 min.
Assays for activity of MFGM enzymes The activity of xanthine oxidase (paper II) and γ-glutamyl transpeptidase (paper III) in raw milk was determined as markers for disrupted MFGM or amount of MFGM, respectively. These MFGM enzymes were chosen, since the assays are very convenient compared with analysis with immunological methods, e.g.
Analysis of xanthine oxidase activity in the milk serum Xanthine oxidase activity was determined by the method of Cerbulis and Farrell (1977). The activity was measured just after treatment so the temperature was kept at the exact pumping temperatures. Milk sample (0.2 mL) was mixed with 1 mL
0.05M-disodium hydrogen phosphate buffer (pH 7.4), 0.8 mL H2O, and 1 mL xanthine solution (20 mg/L H2O). The mixture was incubated at 25 °C for 5 minutes. The reaction was stopped by addition of 1 mL of 20% trichloroacetic acid followed by centrifugation (2000 x g). The accumulation of uric acid was expressed as the absorbance of the supernatant at 290 nm. The molar extinction coefficient of 1.22 x 104 l mol-1 cm-1 for uric acid was used. Samples were analysed in triplicate.
Analyses of γ-glutamyl transpeptidase activity in milk Whole milk (50 µL) was pipetted into 96-well microplates (Nunc, Denmark) and 200 µL substrate mix (0.75M Tris-HCl, pH 6.7, 15mM glycylglycine, 25mM 23 EDTA, 1 mM γ-glutamyl p-nitroanilide) was added. Immediately, absorbance at 412 nm was measured in kinetic mode using an automatic Powerwave microplate reader (Bio-Tek Instruments, USA). Activity of γ-glutamyl transpeptidase in milk was measured as release of p-nitroanilide at 412 nm could be completely inhibited by the specific inhibitor acivicin, proving that the activity observed was due to γglutamyl transpeptidase.
Milk fat crystallisation Determination of liquid fat by nuclear magnetic resonance (NMR) In paper III, the level of liquid fat in milk fat globules was determined by NMR.
Cream was produced by centrifugation (1000 x g) of the milk for 10 minutes at 4°C. Subsequently the cream samples were incubated at 31 °C for 2 hours before measurements. The amount of liquid fat was determined by NMR using a Maran Benchtop Pulsed NMR Analyzer (Resonance Instruments, UK) with a resonance frequency for protons of 23.2 MHz. The NMR instrument was equipped with an 18 mm variable temperature probe. Approximately 2-3 g cream was placed in sealed NMR tubes and upon temperature equilibration at 31°C the free induction decay (FID) was measured. Subsequently the sample was cooled down to 4°C in ice water, and an FID acquisition at 4°C was carried out immediately and thereafter repeated each 10 min for a total of six times. Liquid fat content was determined as signal amplitude of the FID according to the principles described by Samuelsson & Vikelsøe (1971) and the liquid fat content is expressed in paraffin
oil units according to following equation:
24 Results and discussion Influence of feed composition on milk fat globule size A positive correlation was found between diurnal fat yield of cows and average volume-weighted diameter of MFG (paper I). The effect of this new knowledge was used to produce raw milk with different average MFG sizes based upon various fat contents of the milk (paper II and III), as shown in Figure 4. Three groups of cows were fed diets with different amount and source of lipids. One group was given a diet containing a high level of roasted soybeans which is rich in C18:2 (paper II). The result of this diet was production of milk with an average fat content of only 3.7% as the polyunsaturated fatty acids inhibit the formation of precursors for milk fat in the rumen (Børsting, Hermansen & Weisbjerg, 2003).
The fatty acid composition of the produced milk was also affected by the roasted soybeans in the diet.
Figure 4. Effect of fat content and diet on average volume-weighted diameter (d(4,3)) of MFG.
Unsaturated and saturated, refer to the type of lipid in the concentrate. The diets are specified in Material and methods. Grey bars: paper II. Black bars: paper III.
The second group of cows was offered a diet which a included large amount of saturated fat supplement. The consequence of this feeding was that the cows produced milk with 5.0% fat. It is well-known that a saturated fat diet supplement increases the fat content in milk (Chilliard, 1993; Shroeder, 2004) but this was a very high fat content for Holsteins cows. The third group of cows was fed a low fat diet that stimulate high de novo synthesis of the milk fat, resulting in milk with an average fat content of 4.0%. A similar diet was used in the experiment presented in paper III, where it resulted in an average fat content of 3.9%. The 25 second diet in this experiment (paper III) contained a large part of saturated fat supplement with a higher proportion of palmitic acid (C16:0) at the expense of stearic acid (C18:0) compared with the experiment presented in paper II. This diet resulted in that the cows produced milk containing on average 4.6% fat. However, in the experiment presented in paper III, the variation in fat content and average diameter of MFG between cows was larger. In both experiments (papers II and III) the milk type with the highest fat content contained larger MFGs than the milk with a lower fat content. These results from the feeding experiments confirmed the correlation between diurnal fat yield of cows and average volume-weighted diameter of MFG, previously described (paper I). Although, the milk yield was not registered (paper II and III), the large increases in fat percentages upon feeding saturated fat supplements also increased diurnal fat yield, since similar studies have shown that the milk yield slightly increases upon this type of supplements to the diet (Chilliard, 1993).
The mechanisms responsible for the increase in average diameter of MFG when the cow produce more fat were studied in the experiment presented in paper I. It was concluded that the increase in average MFG size is due to the limited amount of membrane material during milk fat synthesis. This was indicated by the significantly decreased activity of the MFGM enzyme, γ-glutamyl transpeptidase, in whole milk with increasing MFG size. Furthermore, it was also demonstrated that when the diurnal fat yield increased, the medium size fat globules were transformed into larger fat globules with an average diameter 8 µm.
The fatty acid composition in the milk was affected by the diets (paper II and III).