A Rapid Electrophoresis Method on Agarose Gel to Characterise Dairy Protein Aggregates

Heat treatment of milk may cause whey proteins and caseins to form aggregates. These soluble and micellar aggregates and their other properties (size, composition, shape, etc.) can affect the techno-functionalities to the milk, conferring interesting or negative features depending on the application in dairy industries. In this study, we propose a new approach to characterise those protein aggregates. SDS-agarose electrophoresis is followed by the calculation of a retention factor (Rf) for each protein spot. Rf allows milk aggregates to be compared qualitatively under the same conditions. This method could be transposed to the dairy industry for a better knowledge of the milk subsequent to heat treatment.


Introduction
In the dairy industry, heat treatments-such as pasteurization-are widely used to stabilize the microbial evolution of the milk and increase the shelf-life of dairy products, or simply to cook the product. But meanwhile, the milk undergoes physico-chemical and biochemical changes; the proteins, especially, can be modified to a greater or lesser extent. At temperatures above 60˚C, soluble whey proteins (β-Lactoglobulin, β-Lg, α-Lactalbumin, α-La, bovine serum albumin, How to cite this paper: Gemelas, L., Degraeve, P., Morand [2]. Thiol-disulfide reactions mainly occur which lead to covalent bonds. However, ionic and hydrophobic interactions can also happen during the protein aggregation, especially during the initial phase of aggregation [1]. Several mechanisms describe aggregate formation [3]. Aggregate composition depends on the physico-chemical state of the milk and the intensity of the heat treatment (time and temperature). Regardless of mechanisms, different aggregates can be formed depending on the product (whey or milk) and their physico-chemical characteristics: a/denatured whey protein soluble aggregates, b/denatured whey proteins bonded to caseins under soluble or micellar forms.
Soluble aggregates are defined here as protein structures found in the supernatant after centrifugation (25,000 g/30min). The pellets containing the micellar aggregates and the micellar caseins are recovered in the precipitate [4].
Protein aggregates formed in a milk heated between 85˚C and 95˚C for a few minutes lead to technological and rheological modifications thereafter [5]. During acid coagulation-in yoghurt and fermented milk processes-, the gelation pH of a pre-heated milk is higher than the raw milk: 5.3 vs. 4.6. The resultant acid gel is firmer and less porous (syneresis is therefore reduced during storage) than the gel of the raw milk [6]. During a cheese making process, it is well known that adding rennet to a pre-heated milk leads to a less cohesive gel-explained mainly by ion migrations inside casein micelles and also by the alteration of the enzymatic site of the chymosin. The alteration of clotting properties as a consequence of thermal treatment constitutes a technological defect inducing yield losses [7]. It is therefore essential to characterise aggregates formed during the different steps of the milk process following heat treatment.
Agarose gel electrophoresis is frequently used, for instance to separate DNA fragments. But according to Wu et al. [9], it could also be employed to discriminate between and characterise proteins, even those with high molecular mass, i.e.
superior to 330 kDa. In this study, we propose a new approach-cheaper and easier to implement in food companies-to characterize protein aggregates formed during the heat treatment of the milk.

Materials and Methods
Except when specified, all the chemical products were provided by Sigma-Aldrich (St-Quentin Fallavier, France). Milk and whey powders were reconstituted at 10% (w/v) in distilled water and stirred during one hour at a rotation speed of 52 rad/s. Sodium azide (0.02% w/v) was then added to prevent the growth of gram negative bacteria. Samples were given 12 hours to equilibrate at ambient temperature without agitation.

pH and Acidity
pH and acidity were measured in triplicate before the heat treatment of the milk (see below). pH was measured using a PHM210 Standard pH Meter (Meter-Lab®). Titratable acidity was determined according to the Dornic method (˚D, with 1˚D = 0.01% lactic acid) [10].

Heat Treatment
Glass tubes (volume 8 mL, diameter 5 mm) were cleaned with hot nitric acid and rinsed with distilled water. The tubes were then filled with 4 mL of sample and closed with a screw cap. The samples were heated for 6 min in a water-bath at 92˚C and then cooled at once in an ice-water bath.

Preparation of Samples for Electrophoresis
Two fractions of pre-heated milk-the supernatant and the pellet-were obtained by centrifugation as follows: 2 mL of pre-heated milk poured into a 2 mL-tube were centrifuged at 25,000 g during 60 min at 20˚C in an Eppendorf centrifuge Type 5417R (microcentrifuge, Eppendorf, Montesson, France). These last parameters of centrifugation have been optimized in intern but the results are not shown. After centrifugation, the supernatant-containing soluble aggregates-was separated from the precipitate, i.e. the pellets-containing micellar aggregates. The two fractions were diluted equally (1 to 1) in the sample buffer.
This buffer contained 0.6 M Tris-HCl pH 6.8, 10% (v/v) glycerol, 0.036% (w/v) bromophenol blue (Biorad Laboratories, Hercules, US). The supernatant and the control milk were homogenised using an agitator (Edmund Bühler, Hechingen, Germany) rotating at 420 rpm during 1 h at ambient temperature. The pellets were first solubilized in the sample buffer by mechanic action and then agitated at 420 rpm for 12 h. One gram of whey powder was diluted in 10 mL of sterile distilled water before the addition of the sample buffer. The samples-control milk, reference whey, supernatant of pre-heated milk, pellets of pre-heated milk, external standard (thyroglobulin)-were all frozen at −20˚C before use.

Preparation of the External Standard
Five grams of thyroglobulin from bovine thyroid were solubilized in 100 mL of sterile distilled water. An equal volume of the thyroglobulin solution was diluted in the sample buffer. The thyroglobulin, used as an external standard electrophoresis, is formed of two subunits of about 330 kDa.

Characterisation of the Aggregates
For each spot observed on a track on the electrophoresis gel, a retention factor (Rƒ) was calculated (such as in chromatography). In our case, the retention factor was defined as the ratio between the distance covered by the protein aggregates (the middle of the spot) and the total length of the track (Figure 1). The retention factor of each spot-i.e. globular whey proteins, caseins, protein aggregates and thyroglobulin (external standard)-was determined using the Im-ageQuant software. Each spot was also marked out by a front and a back end along the track (Figure 1). We also calculated the retention factors corresponding to these two points, referred to later in this document as Rƒmin and Rƒmax.
For the integration of each spot using ImageQuant software, which appeared as a density curve, 2.5% of the area was removed from each side to avoid taking account of the noise resulting from the gel dye. Figure 1. Graphical representation of an agarose gel. After migration, the retention factors of the protein aggregates (in grey) and the thyroglobulin (external standard, in black) were calculated: Rf (ratio between the distance covered by the centre of the spot and the total length of the track), Rƒmin (ratio between the distance covered by the front end of the spot and the total length of the track) and Rƒmax (ratio between the distance covered by the back end of the spot and the total length of the track).

Statistics
Each sample was analysed ten successive times in order to determine a 95% confidence interval for each retention factor. The confidence interval was calculated as follows: -with m the mean, s² the variance and n the number of sample-for a 5% threshold value for a (n − 1) degree of freedom. Two results were considered as significantly different (p < 5%) if their mean ± the standard-error did not overlap.

pH and Acidity Measurement
pH and acidity means were respectively equal to 6.66 ± 0.05 and 13.5± 0.5˚D for the control milk and 6.72 ± 0.05 and 16 ± 0.5 for the reference whey.

Characterisation of Protein Aggregates
The two extracts of pre-heated milk (supernatant and pellets) were analysed by agarose electrophoresis and compared with the corresponding control milk and reference whey, in order to characterise protein aggregates. The deciphering of the gels was based on the work of [4]. The protein bands on SDS-agarose showing on Figure 2 are sometimes diffuse. The methodology was improved later but this example is given anyway because all the different samples were analysed on the same SDS-agarose gel.
The whey proteins with low (between 14 and 18 kDa, spots d and f) and high (between 66 and 83 kDa, spots b and e) molecular mass were both identified in the control milk sample and in the reference whey sample (Figure 2, tracks 1 and 2). Caseins were also spotted between 19 and 25 Da (control milk: track 1spot c). As expected, aggregates only appeared in the pre-heated milk, in the supernatant as regards soluble aggregates (track 3-spot g) and in the pellets for the micellar aggregates (tracks 4-spot h). On the gel, a protein aggregate was observed both in the control milk sample and in the pellets of the pre-heated milk samples (track 1-spot a and track 4-spot i). We can notice that these spots were located on the gel above the high molecular mass whey proteins (track 1-spot a) and below the micellar aggregates (track 4-spot h). It could be argued, along with Guyomarc'h, et al. and Vasbinder et al. [2] [7] that these aggregates may be composed of CN-κ polymers.

Characterisation of Protein Aggregates on SDS-Agarose Gel
Using a New Approach were also characterised. To compare the results of protein migration across several agarose gels, we used this external standard as a reference. We calculated a ratio

Comparison of Protein Aggregates Distribution
The two spots corresponding respectively to the proteins in the control milk and the pellet of the pre-heated milk ( Figure 2, track 1-spot a and track 4-spot i), both presumed to be CN-κ polymers, had the same Rt, and were consequently the same size (Rt ± confident interval) ( Figure 5). The Rt of these two spots were significantly higher (p < 5%) than the Rt of the soluble and micellar aggregates Concerning the small aggregate fractions, whatever their micellar or soluble Figure 5. Distribution of the size of aggregates in samples (Rt values on Y-axis). This is materialized by the retention factors: Rtmin ( ), Rt ( ) and Rtmax ( ). The letter "a", "b" and "c" show significant differences between the distributions following the respective 95% confident intervals. origin, we were not able to draw any conclusion. For instance, it was impossible to estimate the Rtmax of micellar aggregates since the spot partially overlapped with the one of CN-κ polymers. Further research will have to be carried out to improve the visual resolution in this technique.

Conclusion
In this article, we have proposed a technique to qualitatively evaluate the soluble and micellar aggregates formed during the heat treatment of milk. This methodology is based upon protein electrophoresis in agarose gel. Specific units, Rt, Rtmin, Rtmax, defined in relation to an external standard-the thyroglobulinare calculated for each protein aggregate. Under the same analytical conditions, micellar aggregates appeared bigger than soluble aggregates. This methodology could be helpful in dairy research in order to study the presence of protein aggregates in dairy ingredients or subsequent to a heat treatment even though we are aware of the necessity to go on improving the methodology, especially for the detection of smaller-sized aggregates.