J Sci Food Agric 1993, 61, 175-180

Detection of Cows’ Milk in Ewes’ Milk and Cheese by a Sandwich Enzyme-Linked Immunosorbent Assay (ELISA) Elena Rodriguez, Rosario Martin,” Teresa Garcia, Isabel Paloma Morales, Bernabe Sanz and Pablo E Hernandez Departamento de N u t r i c h y Bromatologia 111, Facultad de Veterinaria, 28040 Madrid, Spain (Received 8 June 1992; revised version received 23 September 1992; accepted 10

Gonzalez,

Universidad Complutense,

November 1992)

Abstract: A sandwich enzyme-linked immunosorbent assay (ELISA) has been successfully developed for the detection of defined amounts of cows’ milk in ewes’ milk and cheese. Polyclonal antibodies were raised in goats against bovine caseins (BC). The resultant antibodies were recovered from the crude antiserum by ammonium sulphate precipitation and further purified by immunoadsorption of the cross-reacting antibodies onto columns containing immobilised ovine, caprine and bovine caseins, followed by elution of the bovine caseins specific antibodies (anti-BC) from the column containing the bovine caseins. The anti-BC bound to the wells of a microtitre plate were used to capture the BC from milk and cheese mixtures. Further immunorecognition of the captured proteins was attained with the same specific antibodies conjugated to biotin. ExtrAvidin-peroxidase was used to detect biotinylated antibodies bound to their specific antigens. Subsequent enzymic conversion of substrate gave clear absorbance differences when assaying mixtures of ewes’ milk and cheese containing variable amounts of cows’ milk.

Key words : biotinylated antibodies, enzyme-linked immunosorbent assay (ELISA), ewe’s milk and cheese adulteration.

INTRODUCTION electrophoresis (Aschaffenburg and Dance 1968 ; Pierre and Portmann 1970; Ramos et a1 1977; Addeo et a1

Substitution of expensive ovine milk with cheaper bovine 1981 ; Furtado 1983; Amigo et al 1986), and isoelectro- milk for a greater profit needs to be regulated. However, focusing (Addeo et al 1989; Rispoli and Saugues 1989). the adulterants are so similar in appearance, chemical All these methods require both technical expertise and composition, and organoleptic properties that identifi- laboratory facilities, being relatively expensive to carry cation of the substitution remains a problem for the out. analyst. Accordingly, there is a need to develop analytical Immunological methods developed for milk species procedures for use by dairy factories and law en- analysis include techniques such as the agar-gel immuno- forcement agencies in their attempts to combat increasing diffusion (Levieux 1977; Gombozc et a1 198 1 ; Amigo et problems of adulteration of ovine milk with bovine milk a1 1989), and its ‘dry disc’ modification known as the intended for use in cheese manufacture. COMIT test (Garcia et a1 1989), haemagglutination

Several methods are currently used for milk species inhibition (Levieux 1980), immunoelectrophoresis (Rad- identification. Non-immunological methods include ford et a1 1981 ; Bernhauer et a1 1983; Krause et a1 1988), gas-liquid chromatography (Iverson and Sheppard 1989 ; immunodotting (Aranda et a1 1988) and several forms of Prager 1989), high-performance liquid chromatography the enzyme-linked immunosorbent assay (ELISA) (Kaiser and Krause 1985; Haasnot et al 1986), gel (Garcia et a1 1990, 1991; Rodriguez et a1 1990, 1991;

Sauer et a1 1991). * To whom correspondence should be addressed. Immunoassay techniques provide complementary and/

J Sci Food Agric 0022-5142/93/$06.00 0 1993 SCI. Printed in Great Britain 175

176 E Rodriguez et a1

or alternative approaches to reduce the use of costly, sophisticated equipment and time of analysis, while still maintaining reliability and improved sensitivity. ELISA is the most widely used form of immunoassay in food analysis and has advantages of being low in cost and fast to operate; it is also easy to use, reliable, rapid and readily automated (Allen 1990). It can be designed either as a cheap qualitative method or as a reliable quantitative procedure. The development of immunoassay techniques useful to the food industry may contribute to the quality control and safety of our food supply.

Reported here are the use of affinity purified antibodies against bovine caseins (BCs) and a sandwich ELISA to detect and quantify the presence of bovine milk in ovine milk and cheese.

MATERIALS AND METHODS

Preparation of the antigenic extracts

Bovine milk was defatted by centrifugation at 13000 x g for 10min at 4°C prior to passage of the skim phase through filters containing glass wool. Skim milk was heated at 37°C and acidified to pH 4.6 with 0.2 M HCl, then maintained at 40°C for 30 min and centrifuged at 6000 x g for 30 min to recover the precipitated caseins. The pellet was washed three times with distilled water to eliminate remaining whey proteins and centrifuged at 6000 x g for 10 min. The supernatant was discarded and the pellet containing the precipitated caseins was lyophil- ised, placed in an airtight container and stored at - 20°C until use.

Production of bovine-casein specific antibodies

Polyclonal antibodies against BCs were raised in goats. Immunisation was commenced by intramuscular in- jection into each leg of 12 mg lyophilised BCs emulsified in 1 ml of Freund's Complete Adjuvant (Difco Lab- oratories, Detroit, MI, USA) and 1 ml of deionised and distilled water. Booster injections made in Freund's Incomplete Adjuvant (Difco) were given at l-week intervals for 4 months. Trial bleedings were made monthly. The blood was clotted at room temperature and the serum collected by centrifugation at 2000 x g for 10min. Each batch of antiserum was tested against bovine, ovine and caprine caseins by the agar-gel immunodiffusion technique of Ouchterlony (1 948). Two months after the first immunisation the antiserum showed enough titre for further analysis. Several bleed- ings (up to 1000 ml) were taken from each animal while the immunisation titres remained high. Samples of the crude antiserum (10 ml) were stored frozen at -20°C.

To minimise non-specific interactions the immuno-

globulins were recovered from the crude antiserum by ammonium sulphate precipitation (Harlow and Lane 1988). Samples (10ml) of antisera were centrifuged at 3000 x g for 30 min. The pellet was discarded and 10 ml of a saturated solution of ammonium sulphate (pH 7.4) was added to the supernatant. After overnight incubation at 4"C, the precipitate was collected by centrifugation at 3000 x g for 30min, dissolved in 5 ml of phosphate buffered saline (PBS, pH 7.2), and dialysed overnight against three changes of PBS.

The crude antibodies against bovine caseins were made species-specific by immunoadsorption chromat- ography. Isolation of bovine milk-specific antibodies was carried out by passing 5 ml samples of the precipitated antibodies, diluted 1 :2 with PBS, pH 7.2 containing Tween 80 (0.5 YO), through three affinity chromatography columns at 4°C. Each column contained 10 g of CNBr- activated Sepharose 4B (Pharmacia Fine Chemicals, Uppsala, Sweden) coupled, according to the manu- facturer's instructions, to 175 mg of lyophilised ovine, caprine or bovine casein. After passage of the antibodies through the three columns, specific antibodies against BCs were released from the bovine column by elution with 0.05 M diethylamine (Fluka Chemie, Buchs, Switzer- land) buffer, pH 11.5. The eluted fractions with an absorbance at 280nm higher than 0.1 were pooled, adjusted to pH 7.2 with 0.5 M NaH,PO, (Merck, Darm- stadt, Germany), dialysed overnight against PBS and lyophilised.

Preparation of biotinylated antibodies

The affinity recovered antibodies (anti-BC) were conju- gated to biotinamidocaproate N-hydroxysuccinimide ester (Sigma Chemical Co, St Louis, MO, USA), following the method of Bonnard et a1 (1984), as previously described (Rodriguez et a1 1991). Aliquots of biotinylated antibodies (0.1 ml) were stored at -20°C until use.

Reference milk and cheese samples

Reference milk samples were prepared mixing bovine and ovine milk at different percentages (1-100 % v), and maintained frozen at - 20°C until used. Mixtures from two different batches of each type of milk were prepared to be tested by ELISA.

Cheese samples were prepared using representative samples (25 g) of a known ovine cheese containing weighed portions (1-100 YO w) of bovine cheese. The mixtures were homogenised in 500 ml of PBS, pH 7.2 with a mechanical blender. Portions ( I ml) of the extracts were diluted 1 : 5 in PBS and maintained frozen at - 20°C until use. Mixtures of two different batches of each type of cheese were prepared to be tested by ELISA.

Sandwich ELISA ident8cation of cow's milk 177

Sandwich ELSA procedure

Flat-bottomed micro-ELISA plates (Costar, Cambridge, MA, USA) were filled with 0.1 ml of the capture anti-BC antibodies (1 mg ml-l) diluted in carbonate coating buffer (sodium carbonatebicarbonate buffer, pH 9.6) and incubated for 1 h at 37°C. The wells were washed five times with PBST (PBS containing 5 % Tween 20) and coated with 0-2 ml of 1 YO gelatin in PBS for 30 min at 37°C. After five washes with PBST, 0.1 ml of a milk or cheese sample diluted in PBST was added to the wells and the plates were incubated on a plate shaker (LKB, Pharmacia) for 1 h at 20°C. After being washed a further five times with PBST, 0.1 ml of the biotinylated anti-BC antibodies (1.05 mg ml-') in PBST was added to the wells and the plates incubated with shaking for 1 h at 20°C. After washing five more times with PBST to remove free antibodies, 0.1 ml of ExtrAvidin-peroxidase (Sigma) diluted in PBST was added to the wells, and the plates incubated with shaking for 1 h at 20°C. The wells were washed five more times with distilled water before addition of 0.15 ml of the substrate solution consisting of 0.6 mg ml-' of 2,2'-azinobis-3-ethyl-benzthiazoline sul- phonic acid (ABTS, Sigma) in citric acid-phosphate buffer (pH 3.9) with hydrogen peroxide (1 10 vol, 0.2 pl ml-' buffer). After 30 min incubation, the reaction was stopped by addition to each well of 0.05 ml of 2% sodium fluoride. The green colour which developed was measured at 405nm with a Titertek Multiscan Plus Spectrophotometer (Flow Laboratories, McLean, VA, USA).

RESULTS

The sandwich ELISA described in this work uses polyclonal antibodies raised in goats against BC. To minimise non-specific interactions, the immunoglobulins were recovered from the crude antisera by ammonium sulphate precipitation. The antibodies against the BCs were further purified by affinity chromatography to remove those antibodies capable of cross-reacting with the caseins from ovine and caprine milk. The ammonium sulphate precipitated antibodies were circulated once through columns containing immobilised ovine, caprine and BCs, followed by elution of the anti-BC from the bovine column.

To be used in the sandwich ELISA developed in this work aliquots of the affinity-recovered anti-BC were conjugated to biotin. Extensive checkerboard titrations were performed to determine the dilutions of antigens and antibodies required to optimise the assay. Optimum conditions for the sandwich ELISA were obtained using the capture anti-BC and the biotinylated anti-BC diluted 1 : 500, and the ExtrAvidin-peroxidase conjugate diluted 1 : 3000.

Using the double-antibody sandwich ELISA devel- oped in this work, it was observed (Fig 1) that the anti-

BC effectively discriminated bovine milk from the ovine and caprine milk in a range of milk dilutions from 1 : 100 to 1:6400. From these results it is clear that the immunoadsorption step removed most cross-reacting antibodies and enabled positive identification of bovine milk using this sandwich ELISA procedure.

The sandwich ELISA was further used to detect the substitution of ovine milk by bovine milk in prepared milk mixtures. The milk samples were diluted 1 :400, and Fig 2 shows the absorbance values from two independent batches of ovine milk containing 0, 0.5, 1, 2.5, 5, 10, 15, 25,50 and 100 YO of bovine milk. For these milk samples, absorbance was related to bovine milk content (B) in the range 0.5-1 5 YO by the following equation :

A,,, = 0.965 + 0.458 In (YO B)

2.8{

Antigen dilution (I/Agx1O2)

Fig 1. Sandwich ELISA results for the identification of (0) bovine, (A) ovine, and (w) caprine milk using the capture anti-

BC and detection biotinylated antibodies diluted 1 : 500.

2'8 1

" 0 5 10 15 x Bovine milk in ovine milk

Fig 2. Sandwich ELISA response to (A) ovine milk sub- stitution by (.) bovine milk. Bars represent standard deviation

of triplicate assays from two independent milk mixtures.

178 E Rodriguez et a1

E4 t

0 5 10 15 20 25

X Bovine cheese in ovine cheese

Fig 3. Sandwich ELISA response to (A) ovine cheese substitution by (0) bovine cheese. Bars represent standard deviation of triplicate assays from two independent cheese

mixtures.

15 0 ' '

0 5 10

% Bovine milk in ovine milk

Fig 4. Sandwich ELISA response to (A) ovine milk sub- stitution by (.) unheated, (0) pasteurised, (A) UHT, and (0)

sterilised bovine milk.

with a correlation coefficient of r = 0.975. This re- lationship could be used as a reference curve for evaluating the addition of bovine milk to ovine milk. However, since the reference curve is likely to differ slightly under the conditions given on different days of work and in different laboratories, it is advisable to run a reference curve with known percentages of bovine milk in ovine milk along with the unknown samples. The coefficient of variation of absorbances for a standard sample of bovine milk allocated over two entire 96-well

plates was less than 5 % for intra-assay and inter-assay results obtained by trained operators in the same laboratory.

The sandwich ELISA described here was further used to detect and quantify ovine milk cheese substitution by bovine milk cheese. Optimal conditions for the assay were obtained using the capture anti-BC antibodies diluted 1 :500, the cheese samples diluted 1 : 10, the biotinylated anti-BC diluted 1 : 1000 and the ExtrAvidin- peroxidase conjugate diluted 1 : 3000. Figure 3 shows the absorbance values from two independent batches of ovine milk cheese prepared to include 0, 1,2.5, 5, 10, 15, 25, 50 and 100 YO of bovine milk cheese. The absorbance values increased until substitution of ovine cheese by bovine cheese reached 25 %. In these cheese batches, absorbance was related to bovine cheese content (B) in the range 0.5-25 % by the following equation :

A,,, = 0.7 18 + 0.384 In B with a correlation coefficient of r = 0.953.

A control experiment was also performed to measure the response of the anti-BC against milk samples subjected to different heat treatments. Figure 4 shows that batches of ovine milk containing 0, 0.5, 1,2.5, 5, 10, 15, 25, 50 and 100% of either commercially UHT or sterilised bovine milk gave much lower absorbance responses than those obtained with bovine raw milk. However, mixtures containing pasteurised bovine milk showed only a minimal decrease in their absorbance values as compared to those with bovine raw milk.

DISCUSSION

Substitution of ovine milk by bovine milk constitutes a fraudulent practice in the milk industry due to the difference in price between them. This substitution may become a serious problem in cheese manufacture, as the composition of the milk influences the rennet clotting time, rate of curd formation and coagulum strength of the final product. In a previous report (Rodriguez et a1 1990), the development of an indirect ELISA for the detection of bovine milk in ovine milk and cheese was described. However, the indirect ELISA was not sensitive enough for the quantitation of small percentages of bovine milk (1-10 %) in ovine milk or cheese. Following previous work (Garcia et a1 1991) for the detection of bovine milk in ovine milk using an immunoserum against whey proteins and a sandwich ELISA format, the authors have applied a similar ELISA procedure for ovine milk identification based on the use of an immunoserum against bovine milk caseins as antigens.

The immunopurification of anti-BC on columns containing immobilised ovine, caprine and bovine caseins was found adequate to remove cross-reacting antibodies and to recover those anti-BC that enable confident identification of bovine milk (Fig 1) from ovine and caprine milks. Although immunoadsorption chromat-

Sandwich ELISA identi$cation of cow’s milk 179

ography is time-consuming and requires technical ex- pertise to obtain monospecific antibodies, it is considered an excellent procedure to remove cross-reacting anti- bodies.

To be used as the immunorecognition reagent in the sandwich ELISA developed in this work, aliquots of the affinity recovered anti-BC were conjugated to biotin. Since biotin is a small molecule, its binding to the probe antibody lacks the deleterious effects associated with linking an enzyme directly to the antibody (Wilchek and Bayer 1988). In the sandwich ELISA described here, the biotinylated anti-BC were detected with ExtrAvidin- peroxidase.

The sandwich ELISA developed facilitates the de- tection and quantification of bovine milk in ovine milk and cheese, detecting a 0.5-1 5 YO of bovine milk in ovine milk and a 05-25 YO of bovine cheese in ovine cheese. It should be noted (Figs 2 and 3), that a large difference in the absorbance values was achieved when a 1-10% of ovine milk or cheese was substituted by bovine milk or cheese in the prepared mixtures. In dairy analyses this resolution is very important because a problem currently concerning the dairy industry is the detection of small amounts of bovine milk in ovine and caprine milk.

Using an indirect ELISA technique (Rodriguez et a1 1990) it was observed that mixtures of ovine milk with commercially pasteurised, UHT and sterilised bovine milk gave essentially similar absorbance values to those prepared with bovine raw milk. However, with the sandwich ELISA developed in this work, mixtures of ovine milk with commercially UHT and sterilised bovine milk gave lower immunological responses than expected. The response with the commercially pasteurised bovine milk was slightly lower than that obtained with the raw bovine milk (Fig 3). The differences in the two ELISA procedures in their sensitivity to detect heated bovine milk in the milk mixtures, may be due to the different way in which the BCs are presented to the antibodies. In the indirect ELISA, the heated caseins directly adsorbed to the wells may still present a large number of conserved epitopes to the anti-BC. However, in the sandwich ELISA procedure, the capture anti-BC may mask a number of conserved epitopes that are no longer presented to the immunorecognition biotinylated anti- bodies.

The sandwich ELISA developed in this work for bovine milk identification in ovine milk and cheese is more sensitive than some reported chromatographic (Kaiser and Krause 1985; Iverson and Sheppard 1989), electrophoretic (Pierre and Portman 1970; Ramos et a1 1977; Addeo et a1 1981; Amigo et a1 1986) and immunological methods (Levieux 1977; Gombocz et a1 1981; Radford e t a1 1981; Garcia et a1 1989, 1990; Rodriguez et a1 1990, 1991), being in the range of the sensitivity reported by others (Bernhauer et a1 1983; Aranda et a1 1988; Krause et all988; Addeo e t all990, Garcia et a1 1991, Sauer et a1 1991). This ELISA test

using anti-BC was accurate, easy to perform, sensitive and specific. It could also be suitable for commercial production of stable kits readily available for species verification of milk and cheese in large or small scale inspection programmes.

ACKNOWLEDGEMENTS

This work was supported by grant no AL188-0111 from the Comision Interministerial de Ciencia y Tecnologia (CICYT). ER and TG were recipients of Fellowships from the Ministerio de Educacion y Ciencia, Spain. The authors are grateful to the Cooperativa Castellana de Ganaderos, Campo Real, Madrid, Spain, for their generous gift of the raw milk samples.

REFERENCES

Addeo F, Trie-Cuot P, Chianese L, Ameno M 1981 Addition of cow’s milk to buffaloes’ milk. 111. Electrophoretic analysis of caseins. Sci Tec Lattiero-Casearia 32 95-108.

Addeo F, Moio L, Chianese L, Nota G 1989 Evaluation of bovine and water buffalo milk in mixtures of liquid milk and mozzarella cheese by gel isoelectric focusing. Ital J Food Sci

Addeo F, Moio L, Chianese L, Stingo C 1990 Improved procedure for detecting bovine and ovine milk mixtures in cheese by isoelectric focusing of para-rc-casein. Milchwissen- schaft 45 221-224.

Allen J C 1990 The value of immunoassays to food analysis. In: Development and Application of Immunoassay for Food Analysis, ed Rittenburg J H. Elsevier, London, UK, pp

Amigo L, Santa-Maria G, Gonzalez del Llano D, Ramos M 1986 Polyacrylamide gel electrophoresis of whey proteins in cheeses made from milk of different species. In: XXIII International Dairy Congress, The Hague, p 28.

Amigo L M , Ibaiiez I, Santa-Maria G, Ramos M 1989 Comparison of an electrophoretic and an immunological method for the determination of goat and cow milks in cheese. Milchwissenschaft 44 215-218.

Aranda P, Oria R, Calvo M 1988 Detection of cows’ milk in ewes’ milk and cheese by an immunodotting method. J Dairy Res 55 121-124.

Aschaffenburg R, Dance J E 1968 Detection of cow’s milk in goat’s milk by gel electrophoresis. J Dairy Res 35 383-384.

Bernhauer H, Baudner S, Gunther H 0 1983 Immunological detection of proteins of cow milk in sheep or goat milk and cheese with a specific immunoglobulin. Z Lebensm Unters Forsch 117 8-10.

Bonnard C, Papermaster D S, Kraehenbul J P 1984 The streptavidin-biotin bridge technique : application in light and electron microscope immunocytochemistry. In : Immuno- labeling for Electron Microscopy, ed Polak J M & Varndell I M. Elsevier, London, UK, pp 95-98.

Furtado M M 1983 Detection of cow milk in goat milk by polyacrylamide gel electrophoresis. J Dairy Sci 66,

Garcia T, Martin R, Rodriguez E, Hernandez P E, Sanz B 1989 Development of a cow’s milk identification test (COMIT) for field use. J Dairy Res 56 691498.

Garcia T, Martin R, Rodriguez E, Morales P, Hernandez P E, Sanz B 1990 Detection of bovine milk in ovine milk by an

3 71-80.

59-80.

1822-1 824.

180 E R o d r k u e z et a1

indirect enzyme-linked immunosorbent assay. J Dairy Sci 73

Garcia T, Martin R, Rodriguez E, Azcona J I , Sanz B, Hernandez P E 1991 Detection of bovine milk in ovine milk by a sandwich enzyme-linked immunosorbent assay (ELISA). J h a Prot 54 366369.

Gombozc E, Hellwig E, Petuely F 1981 Immunological estimation of bovine casein in sheep cheese products. Z Lebensm Unters Forsch 172 178-181.

Haasnot W, Venema D P, Elenbaas H L 1986 Determination of cow milk in the milk and cheese of ewes and goats by fast protein liquid chromatography. Milchwissenschaft 41

Harlow E, Lane D 1988 Antibodies. A Laboratory Manual. Cold Spring Harbor Laboratory, New York, USA.

Iverson J L, Sheppard A J 1989 Detection of adulteration in cow, goat and sheep cheeses utilizing gas-liquid chromat- ographic fatty acid-data. J Dairy Sci 72 1707-1712.

Kaiser K P, Krause I 1985 Analysis of proteins in foods by means of electrophoretic and chromatographic methods. 2 Lebensm Unters Forsch 180 181-201.

Krause I, Elbertzhagen H, Berner I, Kostermeyer H 1988 Isoelectric focusing and crossed-immunoelectrophoresis : two supplementary methods for fast and reliable detection of cows milk present in ewe’s or goat’s milk and cheese. Fresenius Zeitschrft fuer Analytische Chemie 330 46M67.

Levieux D 1977 New technique for detecting adulteration of goat’s and ewe’s milk. Dossiers de L‘Elevage 2 3746.

Levieux D 1980 The development of a rapid and sensitive method based on hemagglutination inhibition, for the measurement of cow milk in goat milk. Ann Rech Ver 11

1489-1493.

642-645.

151-156.

Ouchterlony 0 1948 In vitro method for testing the toxin producing capacity of diphteria bacteria. Acta Pathol Microbiol Scand 25 186191.

Pierre A, Portmann A 1970 Detection of cows’ milk added to goats’ milk using polyacrylamide-gel electrophoresis. Ap- plication to cheese. Ann Technol Agric 19 107-130.

Prager M J 1989 Differential characteristics of fatty acids in cheese from milk of various animal species by capillary gas chromatography. JAOAC 72 418-421.

Radford D V, Tchan Y T, McPhillips J 1981 Detection of cow’s milk in goat’s milk by immunoelectrophoresis. Aust J Dairy Technol 36 144-146.

Ramos M, Martinez-Castro J, Juarez M 1977 Detection of cow’s milk in Manchego cheese. J Dairy Sci 60 87C877.

Rispoli S, Saugues R 1989 Isoelectricfocusing in poly- acrylamide gels of the whey proteins from cheeses made with mixtures of cows’ and goats’ milk. Application to the detection of cows’ milk in goats’ cheese. Le Lair 69 21 1-222.

Rodriguez E, Martin R, Garcia T, Hernandez P E, Sanz B 1990 Detection of cows’ milk in ewes’ milk and cheese by an indirect enzyme-linked immunosorbent assay (ELISA). J Dairy Res 57 197-205.

Rodriguez E, Martin R. Garcia T, Azcona J I, Sanz B, Hernandez P E 1991 Indirect ELISA for detection of goats’ milk in ewes’ milk and cheese. Znt J Food Sci Technol 26 457465.

Sauer Von S, Dietrich R, Schneider E, Terplan G 1991 Ein enzymimmunologischer schnelltest zum nachweis von kuhmilch in schaf und ziegenmilch. Archiu fur Lebensmittel- hygiene 42 133-1 60.

Wilchek M, Bayer E A 1988 The avidin-biotin complex in bioanalytical applications. Anal Biochem 171 1-32.