an improved analytical approach for the determination of

HAL Id: hal-00895555https://hal.archives-ouvertes.fr/hal-00895555

Submitted on 1 Jan 2005

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

An improved analytical approach for the determinationof bovine serum albumin in milk

Bärbel Lieske, Alfred Jantz, Berndt Finke

To cite this version:Bärbel Lieske, Alfred Jantz, Berndt Finke. An improved analytical approach for the determination ofbovine serum albumin in milk. Le Lait, INRA Editions, 2005, 85 (3), pp.237-248. �hal-00895555�

https://hal.archives-ouvertes.fr/hal-00895555

https://hal.archives-ouvertes.fr

237Lait 85 (2005) 237–248© INRA, EDP Sciences, 2005DOI: 10.1051/lait:2005018

Original article

An improved analytical approach for the determination of bovine serum albumin in milk

Bärbel LIESKEa, Alfred JANTZb, Berndt FINKEb*

a Department of Food Rheology, Technical University of Berlin, Königin-Luise-Str. 22, D-14195 Berlin, Germany

b Humana Milchunion eG, Bielefelder Str. 66, D-32049 Herford, Germany

Received 1 October 2004 – Accepted 2 March 2005

Abstract – The content of bovine serum albumin (BSA) in milk is considered as a marker of themammary gland health of the cow. Based on its physico-chemical properties, BSA tends to formaggregates hindering its exact determination with the previously known colorimetric bromcresolgreen dye (BCG) procedure. The BCG method was improved by adjusting the milk salt buffer sys-tem for the exact determination of BSA in milk. In a parallel approach, size-exclusion chromatogra-phy was applied. It was found that the analytical availability of BSA in milk samples was best whenmilk samples were directly analyzed with BCG without sample treatments such as precipitation andfiltration. Spiked milk samples showed a recovery of 98.8% (+/– 1.5%) with the improved method.The revised BCG method was fast and reliable in comparison with size-exclusion chromatography,allowing an exact determination of BSA in milk. Based on the spectrometric and chromatographicresults it was concluded that the partial oligomerization of BSA occurring in milk depends on dif-ferent factors: pH, ionic strength, concentration of BSA, actual immunological status and substan-ces affiliated to it. The improved BCG method could be effectively used for the control of the mam-mary gland health status of cows.

bovine serum albumin / bromcresol green dye method / udder’s health status

Résumé – Une approche analytique améliorée pour la détermination de la sérum-albuminedans le lait de vache. La teneur du lait en sérum-albumine (BSA) est considérée comme un indicede la santé de la glande mammaire de la vache. En raison de ses propriétés physicochimiques, laBSA a tendance à s’agréger, ce qui en empêche la détermination précise avec la méthode colorimé-trique au vert de bromocrésol (BCG) actuellement connue. La méthode au BCG a été améliorée enajustant le système tampon salin du lait en vue de la détermination exacte de la BSA dans le lait.Parallèlement, on a appliqué la chromatographie d’exclusion. La mise en évidence analytique de laBSA dans l’échantillon de lait était meilleure quand on analysait les échantillons de lait directementavec la procédure au BCG sans les avoir préalablement traités par des méthodes comme la précipi-tation et la filtration. Les échantillons de lait surchargés ont présenté un taux de récupération de 98,8 %(+/– 1,5 %) avec la méthode améliorée. La méthode au BCG améliorée s’est montrée rapide et fiablepar rapport à la chromatographie d’exclusion et a permis la détermination exacte de la BSA dans lelait. On a pu conclure, sur la base des résultats spectrométriques et chromatographiques obtenus, quel’oligomérisation partielle de la BSA qui se produit dans le lait dépend de plusieurs facteurs : le pH,la force ionique, la concentration de la BSA, le statut immunologique effectif et la présence des

* Corresponding author: [email protected] Abbreviation key: BSA = bovine serum albumin/ Ig = Immunoglobulin/ BCG = bromcresol green/ SE-FPLC = size-exclusion fast protein liquid chromatography/ SCC = somatic cell count/ BTM = bulk-tankmilk/ pI = isoelectric point.

Article published by EDP Sciences and available at http://www.edpsciences.org/lait or http://dx.doi.org/10.1051/lait:2005018

http://www.edpsciences.org/lait

http://dx.doi.org/10.1051/lait:2005018

238 B. Lieske et al.

substances qui y sont liées. La méthode au BCG améliorée pourrait donner de bons résultats dans lasurveillance du statut sanitaire des glandes mammaires de la vache.

sérum-albumine bovine / méthode au vert de bromocrésol / statut sanitaire de la mamelle

1. INTRODUCTION

Normal bovine milk contains 0.1 to0.4 g·L–1 bovine serum albumin (BSA). Thisis 0.3 to 1.0% of the total nitrogen of milk [2].BSA contributes to the serum protein andremains in whey after renneting or acid pre-cipitation of milk. It is quite a large molecule(66.3 kg·mol–1) consisting of 582 amino acidsof known sequence including 17 intramo-lecular disulfides and one free sulfhydrylgroup [23]. The disulfides are relativelyclose together in the polypeptide chain andstabilize the molecular structure in the nat-ural BSA. The molecule is suggested to beelliptical in shape with three distinct glob-ular domains [2]. Each of these domainspossesses a differentiated balance of molec-ular binding forces on the surface, allowingthe binding of anions (free fatty acids), cat-ions (Ca2+, Cu2+ and Mg2+) and interactionswith bilirubin, cholesterol and a variety ofhormones and pharmaceuticals. Moreover,BSA plays a major role in regulating andmaintaining the osmotic pressure in theblood [21].

The biological function of BSA in milkis still unclear. It is suggested that it is of lit-tle significance [2]. Increased concentra-tions of BSA in milk are considered to be amarker of the mammary gland health [8].There are also some results showing no sig-nificant correlation between the concentra-tion of BSA and somatic cell count (SCC)[14], which is a known measurement ofmammary gland health and of milk quality[6, 8–10, 18]. The noted differences in thesignificance of BSA may be partly explainedby methodical difficulties in the correctanalysis of BSA. Currently available meth-ods for the determination of BSA includeradial immunodiffusion [3, 22], isoelectricfocusing and SDS-PAGE [16, 20] orSE-chromatography [13].

Recently, Bouchard et. al. [1] deter-mined BSA colorimetrically using brom-cresol green dye (BCG) to investigate effectsof an endotoxin-induced mastitis. The color-imetric assay was applied in parallel withthe estimation of SCC and N-acetyl-β-D-glucosamidase (NAG) activity, resulting inincredibly high estimates for BSA. TheBSA assay applied was a modification ofthe method developed by Koupparis et al.[11] to measure the serum albumin inhuman blood sera. Koupparis et al. reportedon an automated stopped-flow analysis ofalbumin with BCG and purple. Albuminand bromcresol green are allowed to bind atpH 4.2 and absorption of the BCG-albumincomplex is determined spectrophotometri-cally at 640 nm. At pH 4.2, albumin acts asa cation to bind the anionic dye. Non-spe-cific binding with some other serum pro-teins was minimized by keeping the reactiontime in the analyzer as short as possible. Themethod developed showed sufficient agree-ment with the selective immunodiffusionassay.

Our own studies on the modified proce-dure and its efficiency at quantifying BSAin dairy fluids revealed a reasonably goodcorrelation with the estimates obtained bySE-FPLC, but the absolute BSA levelswere significantly higher and in agreementwith those reported by Bouchard et al. [1]and by Guzmán et al. [4]. To explain thosedifferences, better knowledge is requiredabout the influence exerted by the inherentsteps of the BCG method and by the modeof sample preparation prior to analysis, butalso about effects of molecular-structuralchanges of BSA on the results of analysis.Therefore, this study aimed at qualifyingthe colorimetric method as a more efficientand reliable assay for BSA in milk and derivedsubstrates. The methodical research wassupported by chromatographic separations




Assay for bovine serum albumin 239

by use of SE-FPLC for judging the colori-metric results in comparison with molecu-lar-structural data.

2. MATERIALS AND METHODS

2.1. Milk and whey supply

Fresh raw milk from the local dairy wasskimmed at 1075 × g for 30 min. Rennetwhey was prepared from skimmed raw milkwhich had been renneted by the use of0.15 g·L–1 pure Chymogen (clotting activ-ity 1:7500, Chr. Hansen, Lübeck, Germany)at pH 6.7 and 31 °C. The curd was cut after40 min and the rennet whey was isolated bycentrifugation (3200 × g; 10 min) and/or byfiltration using Whatman No. 42 filter paper.

The pH 4.6 filtrate was prepared fromskimmed raw milk using diluted HCl to pre-cipitate casein at ambient temperature. ThepH 4.6 filtrate was separated by centrifuga-tion and/or filtration using Whatman No. 42filter paper. Natural milk serum was sepa-rated from skimmed raw milk at 20 °C bycentrifuging at 50 000 × g for 60 min usinga Sigma 3 K 30 centrifuge (Sigma Laborzen-trifugen, Osterode, Germany). Pure BSAproducts were purchased from Sigma (Deisen-hofen, Germany), Riedel-de Haen (Seelze,Germany) and Impfstoffwerk (Dessau-Tornau, Germany). The purity was deter-mined using three spectrophotometric pro-cedures, UV-spectrophotometry at 280 nm,Bio-Rad protein assay and the revisedcolorimetric method described here.

2.2. Revised bromcresol green dye assay for BSA

2.2.1. Reagents

BCG stock solution

(i) 0.0474 g BCG (Merck, Darmstadt, Ger-many) was dissolved in 2.0 g·L–1 NaOH to100 mL.(ii) 11.81 g succinic acid was dissolved in500 mL distilled water and adjusted to pH4.2 with NaOH.

(iii) Brij-35 (Merck).The BCG stock solution was prepared by

mixing 1 vol. of (i) with 3 vol. of (ii) andwith (iii) to a final concentration of 0.08%Brij-35. The pH of this solution wasadjusted to pH 4.20 ± 0.05 with 40 g·L–1

NaOH. The BCG reagent was stored in aclosed polyethylene bottle at room temper-ature and remained stable for at least fourweeks.

BSA standard

BSA (4.0 mg·mL–1) was dispersed insynthetic milk salt ultrafiltrate [7] at pH 6.7.The standard BSA curve was prepared dailyby diluting the standard solution (20 °C) to0.5, 1.0, 2.0 and 3.0 mg·mL–1. Then 0.2 mLof each solution was mixed with 1.0 mLdeionized water and 1.0 mL BCG reagent.After 10 min the absorbance was measuredat 640 nm against a BSA-free standard.

2.2.2. Procedure

Skimmed milk (0.2 mL) was mixed with1.0 mL deionized water and 1.0 mL BCGreagent using a tube mixer. Then the samplewas centrifuged for 10 min at 2500 × g topellet the casein. After recording the absorb-ance of the supernatant at 640 nm against ablank (deionized water instead of milk), theactual sample concentration was determinedin comparison with the standard BSA curve.

2.2.3. Size-Exclusion Chromatography

The examination of the reliability of thecolorimetric results was carried out by size-exclusion fast protein liquid chromatogra-phy (SE-FPLC; Pharmacia LKB, Uppsala,Sweden) connected to a variable UV-detec-tor and a Shimadzu C-R 3A integrator. Theseparation was performed with a Superose12 column HR 10/30 (Pharmacia). The col-umn was loaded with 50 µL of milk serum,pH 4.6 filtrate, rennet whey or pure BSAdiluted in elution buffer, consisting of6.15 g·L–1 NaH2PO4 and 8.77 g·L–1 NaCl,





pH 6.9. The flow rate was 0.7 mL·min–1 andthe proteins were detected at 280 nm.

Molecular weight standards used to cal-ibrate the Superose 12 column were BSA,α-lactalbumin (α-La), β-lactoglobulin (β-Lg)and bovine immunoglobulin G (IgG). Allstandards were purchased from Sigma andwere diluted in elution buffer before use(2 mg·mL–1).

2.2.4. Statistical analysis

All experiments were replicated threetimes and all measurements were carriedout at least in duplicate. Effects resultingfrom studies in sample pre-treatment werestatistically analyzed, using the Student’st-test. Differences were considered to besignificant at P < 0.05. The least-squaremethod was used to fit the standard curve.

3. RESULTS AND DISCUSSION

3.1. Studies in the reliability of the BCG method

Bouchard et al. [1] determined BSA inmilk by comparison with a standard BSAcurve (0 to 60 mg·mL–1 of reconstitutedskim milk). From this, the actual concentra-tion of BSA in milk was calculated. Apply-ing this procedure, the BSA level in milk waspermanently overestimated (P < 0.05). Theaccuracy of the modified method was provenby adding various known amounts of BSAstandards (1 to 4 mg·mL–1) to different dilu-ents including reconstituted skim milk, nat-ural milk serum or simulated milk saltbuffer according to Jenness and Koops [7].The results of the analytical recovery areshown in Figure 1 and were substantiatedfurther by regression analysis; the relatedregression equations were as follows:

Jenness and Koops-buffer: g·L–1 BSA =Abs.(640nm) × 7.276 – 0.0115 (1)

Natural milk serum: g·L–1 BSA =Abs.(640nm) × 9.079 – 0.0039 (2)

Reconstituted skim milk: g·L–1 BSA =Abs.(640nm) × 14.419 – 0.0811 (3)

In all cases, the increase in absorbance ofthe BCG / albumin complex was linear.However, the quantitative recovery ofadded BSA from the diluents differed sig-nificantly (P < 0.05) as can be concludedfrom the coefficients of regression. From ananalytical point of view it was concludedthat the calibration of the BCG methodshould never be performed in dairy-baseddiluents, as the analytical recovery isreduced by the conditions of this assay.

The pH of the three diluents are on thelevel of the pH of natural milk but theeffects described here also occurred atthe pH of sour milk products (data notshown). From this, it was suggested thatpH-related effects may only play a second-ary role. The different slopes in Figure 1indicate a substrate-dependence via pro-tein-protein interactions and in parallel areduction of analytically reactive BSA mol-ecules. The molecular changes reachedmaximum value when natural milk sampleswere used as diluents. The majority of rawmilk samples in our study revealed differentslopes of the standard BSA curve corre-sponding to regression coefficients between13.8 and 15.0. A small number of the bulk-tank milk samples, however, showed nosubstrate-dependent interactions betweenmilk proteins and added BSA. The slopes ofthe standard BSA curve as well as relatedcoefficients of regression equations weresimilar to Jenness and Koops buffer (Graph3 in Fig. 1 and Eq. 1), indicating that boththe blank of milk and buffer met in a betterway than using phosphate-buffered salineor distilled water (not shown). The samplesof bulk-tank milk (BTM) were collectedfrom a herd which had been suffering fromimmunodeficiency for quite a while. Thepattern of SE-FPLC revealed no molecular-structural changes in added BSA, whichwas different from the majority of milksamples collected in the period of thisstudy, and is a subject of discussion in sec-tion 3.3.





Taking our own findings into consider-ation, Jenness and Koops buffer was rec-ommended for preparing the BSA standardcurve as indicated in the methodical descrip-tion of the revised BCG procedure.

Practically, absorbance is a linear func-tion of BSA concentration from 0.1 to4 mg·mL–1 but it should be confirmed bysuitable dilution. The total coefficients ofvariation calculated for two pools of dairyfluids with BSA concentrations of 1.2 to1.6 mg·mL–1 and 0.3 to 0.6 mg·mL–1 wereat a level of < 2.0% and 2.5%, respectively.

The statistical data are based on eightBTM samples severely affected by immu-nodeficiency. The samples were chosendue to an exceptional long-term stability incold storage (6 to 8 °C) allowing repeated

evaluations of the revised BCG procedureusing one BTM sample several times aweek. Moreover, these milk samples wereused to prove the reliability of the revisedcolorimetric method, as no interferencewith molecular interactions of BSA andother milk proteins were to be considered.The analytical recovery of added BSA (1,1.5, 2 and 2.5 mg·mL–1) in these milk sam-ples was between 97.8 and 100.3%.

3.2. Effects of sample preparation on the quantification of BSA in milk

The response of the colorimetric assay topre-treatment of milk samples prior to anal-ysis of BSA is reported in Table I. The sub-strates studied were skimmed raw milk(No. 1) and derived from it, pH 4.6 filtrates

Figure 1. Plots of observedresponse of the BCG assay versusstandard concentrations of BSA dis-persed in different diluents. Graph 1,

— reconstituted low-heat skimmilk; Graph 2, — natural milkserum (pH 6.8); Graph 3, — Jen-ness & Koops buffer (pH 6.8).





(No. 2 to 4), natural milk serum (No. 5) andrennet whey (No. 6). As dilution effectswere considered in the calculation of BSAresults, the differences are related as aneffect of sample pretreatment alone (Tab. I).As indicated in Table I, the analytical avail-ability of BSA was best when milk sampleswere analyzed directly. Comparable resultswith pH 4.6 filtrates required both the useof diluted acid for pH adjustment and a con-ditioning step for the acidified milk sampleincluding a thermal treatment at 50 °C for30 min while stirring (No. 2). The quantifi-cation of BSA from pH 4.6 filtrate is still ofanalytical interest as it has been the pre-ferred substrate to determine the milk serumproteins.

The sensitive properties of BSA on acid-ification of milk are seen in relationship toits isoelectric point (pI) which is in therange of pH 4.7 to 4.9 [23] and thus closeto the pI of individual caseins. During acid-ification BSA may precipitate partially withthe casein. The extent of coprecipitationdepends on the actual conditions applied forpH adjustment. The resulting deficits of freeBSA might be significantly higher than inthis study, if the acidified milk sampleswere centrifuged immediately after the pHadjustment.

The analytical availability of BSA is sig-nificantly impaired (P < 0.05) in milk serum(No. 5). About 21.5% of the BSA in the rawmilk co-sedimented with casein micellesduring centrifugation, owing to strong non-

covalent binding forces between both pro-teins at the natural pH of milk. These molec-ular interactions were still provable whencasein was clotted by chymosin (No. 6). Inthis study, about 17% of BSA in milkremained associated with the curd (Tab. I).

3.3. Investigations of the biophysical properties of BSA in raw milk

When the BCG method was calibratedwith increasing concentrations of pure BSAdispersed in a dairy-based diluent, then theslope of the resulting standard curvedecreased significantly (Fig. 1) over thatprepared in simulated milk salt buffer. Itcould be concluded that some of the addedBSA is analytically not available by themethod applied. Possibly, this arises fromthe fact that BSA tends to aggregate. Thisfact was recently described by Militello etal. [15] and declared by the partial unfold-ing of the tertiary structure and by thedecrease in alpha-helix and random coilcontents in favor of beta-sheet aggregates.

To study this skimmed raw milk, recon-stituted low-heat skim-milk and naturalmilk serum were spiked with BSA (1 to3 mg·mL–1). The results of spectrophoto-metric analysis of the four BSA productsused in this study are compared in Table II.When BSA was dispersed in skimmed milkan equilibrating step was inserted to allowcomplete dispersion, e.g., at 25 to 30 °C for30 min while stirring. Then the samples

Table I. Influence of sample preparation on the result of colorimetric BSA analysis in skimmedmilk.

No. Substrate Mode of preparation BSA content (mg.mL–1)a

1 Skimmed raw milk, pH 6.7 without 0.941 ± 0.018

2 pH 4.6 filtrate 36.46 g.L–1HCl, 30 min at 50 °C 0.941 ± 0.020

3 pH 4.6 filtrate 36.46 g.L–1HCl 0.909 ± 0.017

4 pH 4.6 filtrate conc. HCl 0.827 ± 0.016

5 Raw milk serum, pH 6.67 50 000 × g, 1 h 0.739 ± 0.014

6 Rennet whey without 0.781 ± 0.015

aarithmetic mean (n = 6).





were prepared for further chromatographicanalysis by SE-FPLC to follow the molec-ular-structural changes in the pattern ofadded BSA. In SE-FPLC, the whole proteinin milk is not resolved to satisfaction due tostrong molecular forces between the indi-vidual caseins as well as interactionsbetween casein and whey protein. For thisreason, the milk samples were centrifugedat 50 000 × g for 60 min at 20 °C to separatethe casein fraction from serum protein at thenatural pH of milk. The serum protein in thesupernatant was diluted further in elutionbuffer and applied onto the column to frac-tionate the proteins according to theirmolecular size in physiological conditions.The resulting chromatographic pattern ofBSA in milk serum was compared with thepattern of the pure BSA products consid-ered next.

The pure BSA products reported inTable II were separated into two distinctfractions. The smaller one corresponded toa molecular weight of about 67 kg·mol–1

(16.5-min retention time) and is assigned toBSA, whereas the unknown protein sepa-rated in the second fraction of about35 kg·mol–1 (18.4-min retention time) wascollected from the Superose column for fur-

ther analysis. The assumption that the35 kg·mol–1 protein belonged to BSA wassupported by two findings. First, this pro-tein responded to the BCG reagent with aA280nm/OD640nm coefficient of between6.10 and 6.17 and thus on a level with thepure BSA products (Tab. II) and secondly,this protein is contained in the pattern ofthe natural milk serum protein as seen inFigure 2a.

The effects of adding pure BSA to askimmed raw milk sample on the SE-chro-matographic profile of the serum proteinare seen in Figure 2. From that, one can con-clude that increasing the BSA concentrationin a natural milk sample from 0.38 g·L–1

(Fig. 2a) to 1.38 g·L–1 (Fig. 2b) leads to anincrease in SEC of both the BSA signals at16.5 and 18.4 min as well as the signal ofthe protein eluted at 11.0 min. This phe-nomenon reflects the inherent molecular-structural property of BSA to build aggregatesafter spiking with pure BSA. If the total con-centration of BSA in milk was experimentallyadjusted to 2.38 mg.mL–1 (Fig. 2c), theresulting milk serum was partially deprivedof both the IgG/BSA fraction (11.0 min)and BSA (16.5 min and 18.5 min), by thecentrifugal forces (50 000 × g), indicating

Table II. Spectrophotometric characterization of some commercial products of pure BSA(3 mg.mL–1 in 6.15 g·L–1 NaH2PO4 / 8.77 g·L–1 NaCl, pH 6.9) used in this study.

No. BSA product UV-spectrophotometry BIO-Rad Assay BCG Assay

A280nm mg.mL–1a OD595nmb OD640nm

c mg.mL–1

1 Biosynth ® 1.877 2.84 0.309 0.288 2.71

Riedel-de Haën

2 “Dessau” 1.980 3.00 0.325 0.326 2.96

Impfstoffwerk

Dessau- Tornau

3 Lot 59 H 7612 1.755 2.66 0.302 0.281 2.63

Sigma

4 Lot 79 H 7614 1.783 2.70 0.305 0.289 2.71

a) A 280nm, 1cm, 1% = 6.6 [22] b) sample dilution prior to analysis was 0.260 mg.mL–1 a) b) c) Values are means of triplicate determinations in two separate experiments .





a deficiency of the colloidal stability of theproteins. Precipitation of some of the addedBSA with centrifugation may follow areduced protein protection from the milksalt system, which differs markedly fromthe salt concentration in inflammatoryudder infections. This assumption is sup-ported by the results of the colorimetricanalysis. When the concentration of BSA inthe same milk sample was experimentallyadjusted between 1 mg·mL–1 and 6 mg·mL–1

the resulting plot of BSA vs. Abs.(640nm)revealed a straight line with a regressioncoefficient of 14.25 (Fig. 1) indicating thatproteins were completely dispersed andthus ready to react with the BCG reagent.The aggregates of BSA are also dispersed,but fail to react with the BCG reagent andundergo acid denaturation and precipitationat pH 4.2 together with the other dairy pro-teins. The extent of polymerization is sub-strate-dependent and decreases in thefollowing order: raw milk > reconstituted

milk > natural milk serum. It seems worthmentioning that BSA/Ig aggregates wereproduced spontaneously. It can be easilyproved with natural milk serum spiked withBSA. The spiked sample is briefly mixedand right away separated by SEC. In thechromatographic pattern the known effectsof polymerization are detectable. Undoubt-edly, the spontaneous formation of aggre-gated BSA at ambient temperature is notimplicated in covalent binding forces viadisulfide bridges.

Even though in the SEC-profiles BSAaggregates eluted together with the immu-noglobulins, it is not likely that BSA acts asa precursor of IgG. Both proteins differ intheir amino acid sequence and their glyco-sylation. BSA contains no glycan chainswhile IgG is partly N-glycosylated, approx-imately 2.9% of the IgG molecule are gly-cans [23]. The aggregates formed by themuch smaller BSA molecules were proved

Figure 2. Effects of adding BSA to raw milk on the resulting serum patterns obtained with aSE-FPLC system, connected to a Superose 12-column. Pattern (a), normal serum. Pattern (b), plus1 mg.mL–1 BSA. Pattern (c), plus 2 mg.mL–1 BSA.





to remain stable and were not disintegratedduring separation by SEC performed at pH7.2, where the natural albumin is an anionwith more than 200 negative charges permolecule [21]. The molecular stability ofthe aggregates was also ensured when SDS-PAGE with brilliant blue staining was usedfor characterization of BSA: two bands atapproximately 67 kg·mol–1 and approxi-mately 130 kg·mol–1 occur [5]. In SDS-PAGE severe disintegrating and denaturingas well as chemical reduction of disulfidesare involved with sample preparation andelectrophoresis. From this, it is unlikely thatthe formation of BSA aggregates is areversible reaction. These aggregates may,however, be of biological significance asthey belong to the main protein in the serumpattern of milk samples collected from indi-vidual udders suffering from clinical mas-titis. Such a chromatographic pattern is shownin Figure 3. The increase in blood protein inmilk due to a disruption to the integrity ofthe mammary epithelia leads to an increasein SEC of both the BSA signals as well asthe signal at the retention time of IgG/aggregated BSA at 11.0 min.

In addition, the new signal seen between12.5 and 13.0 min belongs to the antimicro-bial protein lactoferrin (Lf) (Fig. 3) asascertained by SDS-PAGE [19].

Next, a further qualification of the35 kg·mol–1 protein was necessary to ruleout the existence of β-Lg dimers (36 to38 kg·mol–1). Different methods for quan-tifying β-Lg were used to supply proof ofthe assigning of the 35 kg·mol–1 protein.First, chromatographic analysis of β-Lgwas chosen (Fig. 3). For that, the SEC col-umn was calibrated with pure β-Lg and astandard curve was prepared. The chroma-tographic analysis showed 2.52 g·L–1 β-Lg(SD 0.04) contained in the single fraction ofβ-Lg (Fig. 3). The low value appeared to berealistic due to a decreased synthesis ofmilk proteins involved in cases of severemastitis and it is found to be in agreementwith other chemical parameters determinedin the quarter milk sample with mastitis,

e.g., total protein, lactose and fat amountingto 28.6 g.·L–1 (SD 0.03), 36.4 g·L–1 (SD0.06) and 23.0 g·L–1 (SD 0.05), respec-tively.

The second method to determine appliedβ-Lg is a spectrophotometric assay devel-oped for native β-Lg in raw and processedmilk [12]. The spectrophotometric analysisof β-Lg in the serum of the quarter milksample amounted to 2.55 g·L–1 β-Lg (SD0.04), thus agreeing with the chromato-graphic findings. The results of FPLC were

Figure 3. Effects of a clinical case of mastitison the resulting serum pattern obtained with aSE-FPLC system and a Superose 12- column.(a): 35 kg.mol–1; (b): 67 kg.mol–1.





further verified using SDS-PAGE [19] as athird method. If the 35 kg·mol–1 proteinactually belonged to dimeric β-Lg, then itwould be checkable in the pattern of SDS-PAGE. In the electrophoresis pattern (datanot shown) a pale protein fraction is detect-able with a molecular mass between 33 and40 kg.mol–1 that may be assigned to theBSA fragment in question.

Summarizing the results obtained by thethree methods, it is concluded that theapparent molecular-structural changes ofnatural BSA in milk collected from anudder quarter suffering from severe mastitiswere confirmed. The presence of the BSAresidue in the pattern of SDS-PAGE andSE-FPLC is not limited to the substratesanalyzed in this study. Practically, it wascontained in all sera and whey samples, ina varying concentration but at a lower per-centage compared with BSA. In the litera-ture, the allocation of the protein fractionvaried a lot, and it remained mostly uncon-sidered.

Finally, we focus our attention onceagain on the restricted number of bulk-tankmilk samples (BTM) being remarkable fortheir analytical characteristics in the BCGassay. When these samples were mixedwith added BSA, then in SEC both BSAfractions (67 kg·mol–1 and 35 kg·mol–1) wereincreased; however, no BSA aggregateswere produced. The serum protein patterndiffered significantly from those of themajority of about two hundred BTM sam-ples and were analyzed with regard to boththe concentration of IgG (< 0.25 g.kg–1

BTM ) and BSA (> 1.50 g·kg–1 BTM). Incomparison, the approximate average quan-tities of Ig and BSA in milk are 0.75 g·kg–1

and 0.40 g·kg–1, respectively [23]. From this it was suggested that the for-

mation of aggregated BSA complexes inmilk may be dependent on the presence ofIg, and thus are based on protein-proteininteractions between BSA and Ig. Theseinteractions were at the maximum in rawmilk and decreased in parallel with thedepletion of Ig with thermal treatments or

with the different procedures of separationapplied, and are found to be in good agree-ment with the colorimetric results reportedin Figure 1. Comparable interactions betweenIgG and the much smaller plasma albuminare well known in immunochemistry. Thecourse of this reaction goes on voluntarily,depending on entropy, and is for the mostpart irreversible [5, 17]. It is applicable tothe results reported here. Recently, it wasconfirmed that antigenic determinants ofbovine serum albumin were bound to IgGand IgA antibodies [5]. The reactionsinvolved are part of the natural defence sys-tem against udder infections [22]. In thisstudy, some BTM samples served as anexample of when the natural antibodyresponsiveness is affected and is thus inef-fective.

The BTM samples in question beingdeficient in Ig and without any provablereaction to added BSA were traced back toa weakened immune system on the herdlevel owing to toxic residues in fodder thathad been improperly processed and fed tothem for months. A safe one replaced thisfodder and the BTM was analyzed again13 weeks later to examine the effects ofadapting the new fodder resource. The pro-files of SEC (not shown) revealed a decreasein BSA and in parallel an increase in theconcentration of Ig, indicating an improve-ment of the immune system. Moreover, someaggregated BSA/Ig was detected againwhen this milk was spiked with BSA. Fromthis, it can be concluded that the status ofthe immune system plays the decisive rolein the formation of polymerized protein inmilk.

4. CONCLUSION

The methodical insufficiencies of quan-tifying BSA in milk reported during samplepreparation prior to analysis or with cali-bration of the colorimetric method wereattributable to the outstanding molecular-structural and inherent biological proper-ties of this serum protein.





The evaluation of reliability of the BCGmethod is becoming an analytical problemwith normal milk samples due to incom-plete recovery of the BSA added. Whenevera raw milk sample is spiked with BSA, someBSA molecules will polymerize with thenatural Ig contained in milk. The aggregatesof BSA and Ig are, in all probability, part ofthe immune reactions happening in theudder. It was confirmed by experimentsstudying the character of the molecularbonds stabilizing the aggregates, revealingthat these bonds are produced spontaneouslybased on non-covalent binding forces. Asseen in the chromatographic pattern the for-mation of the polymer structures in milk isdetectable both in vitro and in vivo, inexperimental spiking with BSA and ininfectious udder diseases, respectively.

The proof of reliability of the revisedBCG method was feasible with milk col-lected from udders with an acute immuno-deficiency. In the chromatographic patternof these milk samples no or minimal indi-cation of Ig was seen. When these milksamples were spiked with increasing con-centrations of BSA, the analytical recoveryof BSA by the BCG method was 98.8 ±1.5% because no BSA/Ig aggregates wereformed.

The biological function of BSA is definedby the differences ascertained between thetwo opposite milk samples with addedBSA. From this it was concluded that com-plexed BSA/Ig is one link in the chain ofimmune defence reactions at an early stageof udder infection. That natural defencesystem, however, is switched off by toxicsubstances in the fodder and thus it was notsurprising that relevant changes in the rawmilk are not subjected to the control systemof a dairy. From an interdisciplinary viewit is of great interest to get reliable informa-tion about the actual immune status of onecow or on the herd level. The methodicalapproach of BSA analysis presented heremay be used as a basis for a rapid screeningtest of the immune status of lactating cows.An indication of the actual status of the

immune system may be obtained if milksamples are used for calibration in parallelto the working method as recommendedhere. The colorimetric results are confirmedfurther by SE-FPLC requiring two runs persample, with and without added BSA.

Moreover, the use of BSA as an indicatorof the udder health status is very promisingas BSA levels in milk were correlated withSCC [1]. Giesecke and Viljoen [3] devel-oped a radial immunodiffusion techniquefor quantifying BSA in milk. However, thetedious setting-up time and the measure-ment of the precipitin ring size, togetherwith the 18 to 20 h incubation period,excluded it from large-scale monitoringprograms [9].

All of the methodical insufficiencies ofdetermining BSA in milk were overcomeusing the revised BCG procedure. Moreo-ver, the BCG method is easy to perform,fast and reliable and may be carried out ina continuous analyzer with some signifi-cant advantages if large numbers of sam-ples were to be analyzed.

Acknowledgments: The authors would like toacknowledge C. Falke, A. Kadow, S. Graf andS. Jung for their contribution to the experimen-tal work.

REFERENCES

[1] Bouchard L.S., Blais C., Desrosiers X., ZhaoX., Lacasse P., Nitric oxide productionduring endotoxin-induced mastitis in thecow, J. Dairy Sci. 82 (1999) 2574–2581.

[2] Fox P.F., McSweeney P.L.H., Dairy Chemis-try and Biochemistry, Chapman and Hall,London, UK, 1998.

[3] Giesecke W.H., Viljoen M.H., The diagnosisof subclinical mastitis in lactating cows: acomparison of cytological methods and amonovalent radial immunodiffusion test,Onderstepoort J. Vet. Res. 41 (1974) 51–74.

[4] Guzmán A.A., Peréz M.D., Marquez R.M.,Garza J.R., Colorimetric determination ofbovine serum albumin (BSA) in milk usingbromcresol green dye, J. Dairy Sci. 69(Suppl. 1) (1986) 102.





[5] Hilger C.F., Grigoni G., de Beaufort G.,Michel J., Freilinger F., Hentges F., Differentbinding of IgG and IgA antibodies to antige-nic determinants of bovine serum albumin,Clin. Exp. Immunol. 123 (2001) 387–394.

[6] Jayarao B.M., Wolfgang D.R., Bulk-tankmilk analysis. A useful tool for improvingmilk quality and herd udder health, Vet. Clin.Food Anim. 19 (2003) 75–92.

[7] Jenness R., Koops J., Preparation and proper-ties of a salt solution which simulates milkultrafiltrate, Neth. Milk Dairy J. 16 (1962)153–164.

[8] Kehrli J.M.E., Shuster D.E., Factors affec-ting milk somatic cells and their role in healthof the bovine mammary gland, J. Dairy Sci.77 (1994) 619–627.

[9] Kitchen B.J., Review of the progress of dairyscience: bovine mastitis: milk compositionalchanges and related diagnostic test, J. DairyRes. 48 (1981) 167–188.

[10] Kitchen B.J., Middleton G., Durward I.G.,Andrews R.J., Salmon M.C., Mastitis dia-gnostic tests to estimate mammary gland epi-thelial cell damage, J. Dairy Sci. 63 (1980)978–983.

[11] Koupparis M.A., Diamandis E.P., MalmstadtH.V., Automated stopped-flow analyser inclinical chemistry; determination of albuminwith bromcresol green and purple, Clin.Chim. Acta 149 (1985) 225–235.

[12] Lieske B., Konrad G., Faber W., A new spec-trophotometric assay for native and pro-cessed milk, Int. Dairy J. 7 (1997) 805–812.

[13] Lieske B., Konrad G., Faber W., Chromato-graphic analysis of serum protein in bulkmilk. A possibility to detect changes of themammary tissue as well as of the manufac-turing properties of milk, virtual conference:Sustainable Animal Production. Forum 3:Product Safety, October 1999-April 2000(Internet session).

[14] Lindmark-Månsson H., Svensson U.,Paulsson M., Aledén G., Frank B., Johnsson

G., Influence of milk components, somaticcells and supplemental zinc on milk proces-sability, Int. Dairy J. 10 (2000) 423–433.

[15] Militello V., Casarino C., Emanuele A.,Giostra A., Pullara F., Leone M., Aggrega-tion kinetics of bovine serum albumin studiedby FTIR spectroscopy and light scattering,Biophys. Chem. 107 (2004) 175–187.

[16] Murphy M.C., Howell N.K., Effect of theattachment of valine residues on the physico-chemical and functional properties of bovineserum albumin, J. Sci. Food Agric. 54 (1991)245–253.

[17] Rappoport S.M., Medizinische Biochemie.5. Aufl., VEB Verlag Volk und Gesundheit,Berlin, DDR, 1969, pp. 516–520.

[18] Rogers S.A., Mitchell G.E., The relationshipbetween somatic cell count, composition andmanufacturing properties of bulk milk. 6.Cheddar cheese and skim milk yoghurt, Aust.J. Dairy Technol. 49 (1994) 70–74.

[19] Schägger H., von Jagow G., Tricine-sodiumsulfate-polyacrylamide gel electrophoresisfor the separation of proteins in the rangefrom 1 to 100 kD, Anal. Biochem. 166 (1987)368–379.

[20] Sheffield L.G., Mastitis increases growthfactor messenger ribonucleic acid in bovinemammary glands, J. Dairy Sci. 80 (1997)2020–2024.

[21] Tietz N.W., Textbook of Clinical Chemistry,W. B. Saunders Co., Philadelphia, PA, USA,1986.

[22] Wagter L.C., Mallard B.A., Wilkie B.N.,Leslie K.E., Boettcher P.J., Dekkers J.C.M.,A quantitative approach to classifying Hols-tein cows based on antibody responsivenessand its relationship to peripartum mastitisoccurrence, J. Dairy Sci. 83 (2000) 488–498.

[23] Walstra P., Jenness R., Dairy Chemistry andPhysics. John Wiley & Sons, New York, NY,USA, 1984.




an improved analytical approach for the determination of

Documents