- Email: [email protected]
Methodology for meat species identification: A review
Meat Science 15 (1985) 215-224
Methodology for Meat Species Identification: A Review* C. H. S. Hitchcock & A. A. Crimes Unilever Research, Colworth House, Sharnbrook, Bedford MK44 1LQ, Great Britain (Received: 29 January, 1985)
SUMMARY This review covers methods available for the recognition of the species of animal whose meat ispresent in raw materials or meat products. Isolation of species-specific proteins is cumbersome; however, isoelectric focusing followed by recognition of the pattern of protein bands is particularly effective for speciation. Immunoassay has been established as a powerful technique for the determination of food protein analytes in suitable extracts of unresolved mixtures. Simple immunodiffusion using antisera to serum proteins is sufficient to provide a check on the identity of suspect raw samples; routine control would involve extensive sampling programmes. Enzyme-linked immunosorbent assay procedures are also appropriate, but attempts to quantify individual meat species in mixtures have not been successful due to the variability of residual blood levels. Species identification in heat-processed meat products is hindered by progressive denaturation of the protein markers, leading to loss of solubility and antigenicity; the effects of heating can be minimised by a judicious choice of analyte and its solubilisation by renaturation.
INTRODUCTION The presence of meat in a sample of any food can be established by chemical analysis based on the occurrence of 3-methylhistidine in washed * This review is based on our contribution to the EEC Workshop on 'Biochemical Identification of Meat Species' held on 27th-28th November, 1984, at the Commission of the European Communities in Brussels. The Workshop was organised by Dr R. L. S. Patterson (Meat Research Institute, Langford, Bristol, Great Britain) and financed by the Commission from its budget for the co-ordination of agricultural research. 215
Meat Science 0309-1740/85/$03-30 © Elsevier Applied Science Publishers Ltd, England, 1985. Printed in Great Britain
216
C. H. S. Hitchcock, A. A. Crimes
muscle hydrolysates (Jones et al., 1982). This unusual amino acid arises in vivo by the specific methylation of a particular histidine residue in actin and in myosin in the myofibrillar protein. However, since this biological reaction is common to all species examined, it cannot be used for identification purposes; moreover, its use as a quantitative index for total meat is hindered by variable levels in certain types of manufacturing meat and offals (Jones et al., 1985). Minor species-specific proteins are appropriate for species recognition, but these do not contain such convenient component markers; they must therefore be monitored without hydrolysis, since this would destroy their distinctive features. Two very different approaches to the qualitative detection or quantitative determination of such proteins can be considered. First, a solution of proteins can be separated by electrophoresis or chromatography and the isolated analyte recognised or estimated; secondly, the specific protein can be characterised without separation using specific immunoreagents (Kurth & Shaw, 1983). In both cases, the preparation of the protein solution from a raw meat or a mixed raw product is straightforward; however, heat-processing causes denaturation, interaction and chemical reaction between sample components, which hinders extraction of a proteinaceous analyte from the partially insoluble sample. For this reason, attempts have been made to discriminate between species by analysing the simple extracts for non-protein analytes; this is unlikely to be generally useful since metabolism of these small molecules is less species-specific. The measurement of heat-stable dipeptides (anserine, carnosine, balenine) has had limited success in the discrimination of lamb and pork (Carnegie et al., 1984). Exploitation of fat analysis is illustrated by the use of linolenic acid (Payne, 1971) as a marker for horse fat; nevertheless, this marker can give false negative results. It is a characteristic of horses fed on pasture rather than concentrates (Hilditch & Williams, 1964) and is too variable for use as a reliable parameter (Verbeke & De Brabander, 1980). More complex fat analysis can distinguish lard in tallow (El Sayed & E1 Dashlauty, 1979; Verbeke & De Brabander, 1979). Indeed, the pattern of fatty acids at the 2- and 1,3positions of triglycerides can lead to inferences about the sources of fat (and therefore of meat) present in meat products containing mixtures of pork, beef, horse and chicken. Such multivariate analysis is facilitated by computerised data handling for collection and storage of a large bank of quantitative information; observed data from a particular sample can then be compared and its composition calculated as a best estimate with a given degree of confidence (Verbeke & De Brabander, 1980).
Methodology for meat species identification: a review
217
ELECTROPHORETIC METHODS Electrophoresis is a powerful research technique for the separation of proteins, usually as characteristic band patterns in a supporting gel, visualised, if necessary, by simple non-specific staining, or by enzymological or immunological methods. Separation can be governed by the use of homogeneous gels, concentration-gradient gels, pH-gradient gels or denaturants (such as urea or detergents that dissociate the tertiary protein structures); successful applications to species recognition include isoelectric focusing (IEF) of aqueous extracts of raw meat, based on the coloured myoglobin bands (Sinclair & Slattery, 1982) or on discriminating between appropriate species-specific isoenzymes after specific visualisation (King & Kurth, 1982; Slattery & Sinclair, 1983). Although it is notoriously difficult to standardise conditions of electrophoresis between laboratories, its establishment has been facilitated by the commercial availability of standard buffers, well designed apparatus, laser densitometers and computer-assisted data processing (M. Ansfield, private communication). Gels of 1 mm thickness are normally used for the electrophoresis matrix; ultrathin ( < 100/~m) gels reduce both the time required (by a factor of ten) and expenditure on chemicals (F. Grundhofer, private communication). The relatively high initial capital cost is offset in a centralised laboratory with high sample throughput. Although species-specific proteins in heat-processed samples tend to be denatured and insoluble, simple extraction can yield enough solution for electrophoretic speciation (Babiker et al., 1980/1). Efficient extraction requires the aid of solubilisation agents: for instance, cooked meats can be distinguished after extraction and electrophoresis in 8 M urea and visualisation by non-specific dye-binding (Mattey et al., 1970). Specific detection gives more reliable results, but this can involve extraction into 6 M guanidine hydrochloride, dialysis into 1% Triton X-100, isoelectric focusing in pH-gradient gels and visualisation by renatured enzyme activity, e,g. adenylate or creatine kinase, as demonstrated by King (1984). These lengthy procedures are perhaps most appropriate for definitive confirmatory tests after immunological screening, or for special discrimination between two immunologically related species. The interpretation of electrophoretic data is hampered by the visual complexity of the pherogram; the monitoring of the isoenzyme activity gives only a few bands whose resolution typically allows detection of 1 ~o adulteration of raw beef, illustrating its overall power (Kurth & Shaw, 1983).
218
C. H. S. Hitchcock, A. A. Crimes
I M M U N O L O G I C A L METHODS Immunoassays, reviewed by Hitchcock (1984), based on very specific antigen-antibody interactions, can detect and determine particular analytes in situ in complex mixtures such as biological fluids or food extracts. While there are many analytical procedures in which immunoreagents can be used, the two most appropriate for application in species recognition are the classical Ouchterlony double immunodiffusion technique and the more recent enzyme-linked immunosorbent assay (ELISA). Both are robust and suitable for screening purposes; of the two, ELISA requires less antiserum and less time; it is more objective, more sensitive and, although more complicated, it can be semiautomated. Immunodiffusion
Currently, factory and control laboratories favour simple immunodiffusion in agar gels based on commercial antisera to species-specific blood proteins. If necessary, antisera to unusual species are readily prepared. The properties of each batch must be checked before use, since antisera are difficult to standardise. An aqueous extract of soluble proteins from the meat sample is prepared and aliquots placed in wells cut in a shallow layer (2-3 mm) of agarose gel. Antisera against the species to be tested are placed in an adjacent well. The plate is then incubated to allow antibodies and antigens to diffuse together. If both are present, a visible opaque band will appear in the gel between the wells. Typically, contamination by one species in another can be detected without the use of complex equipment and results are available within 24h (Hayden, 1978; Swart & Wilks, 1982). Using an optimised procedure with commercial precipitating antiserum to species-specific blood proteins, it is possible to detect beef, horse, pig, poultry, sheep and kangaroo serum routinely, and thus infer the presence of the meat; however, it is not possible to distinguish sheep from goat and horse from donkey meat. Subjective detection limits in raw beef are: sheep, 5 ~o; pork, 10 ~ ; horse, 10 ~o; chicken/turkey, 15 ~ and kangaroo, 20 ~o; beef itself could be detected at the 2 ~o level in pork. Of a wide range of meat product ingredients examined, only one interfered: egg protein gave a positive response in the assay for poultry meat, which was based on rabbit anti-hen antiserum (A. A. Crimes, unpublished). False-positive results occasionally arise by non-specific interactions
Methodology for meat species identification: a review
219
leading to a relatively diffuse precipitate which may sometimes be recognised by its abnormal appearance and its solubility in water; it may therefore disappear on soaking the gel in saline. Other reported difficulties have been ascribed to the presence of citrate or ascorbate (Tizard et al., 1982) and mechanically deboned veal (Gleeson & Walker, 1984). Samples of raw meat can be conveniently screened in factory environments and inspection laboratories by double immunodiffusion using a supply of appropriate antisera, standards and agar gels prepared when required; even more convenient would be a standardised kit which provides for individual tests off the shelf. The on-site (overnight) rapid beef identification test (ORBIT) has been designed to verify beef species: filter paper discs are inserted into slits cut into the meat sample and then applied directly to precast gels; the antisera are provided freeze-dried onto similar discs (Mageau et al., 1984). Double immunodiffusion normally takes 24 h, but wherever speed is important, the diffusion can be assisted by the application of an electrical voltage. This procedure of Countercurrent Immunoelectrophoresis (CIE) should perhaps be regarded as accelerated immunodiffusion, to distinguish it from the more sophisticated techniques designed for separation of antigens (e.g. immunoelectrophoresis, Laurell rockets, 2Dimmunoelectrophoresis). CIE is particularly effective where speed and cost are important (T. N. Allsup, private communication). Enzyme-linked immunosorbent assay Semiquantitative determination of antigens by simple immunodiffusion is only possible by subjective comparison of results between various dilutions of sample extract and those of standards; more objective data can be derived from more sophisticated immunoprecipitation techniques. However, enzyme-linked immunosorbent assay (ELISA) is one of several powerful quantitative immunological procedures which are not monitored by precipitation; here the antibody-antigen interaction occurs in a monomolecular layer immobilised on an inert surface and is followed by means of an enzyme chemically bonded to one of the immunoreagents (Hitchcock & Crimes, 1983). ELISA is a well established technique, particularly in clinical laboratories where it provides rapid analysis of biological fluids and tissues for many antigens and antibodies both in qualitative screening and quantitative diagnostic procedures. In food
220
C. H. S. Hitchcock, A. A. Crimes
control laboratories, it has been applied with some success (Crimes et al., 1984; Olsman et al., 1984) to the determination of soya protein in meat products, which serves as a model for all food proteins (Hitchcock et al., 1981). ELISA has therefore been applied to species recognition by choosing species-specific proteins; in particular, the blood proteins tested by immunodiffusion have been demonstrated to be appropriate for uncooked samples in several laboratories using somewhat different ELISA procedures: the initial immobilised species may be the standard antigen (Griffiths & Billington, 1984), the analyte antigen in the sample itself (Kang'ethe et al., 1982; Whittaker et al., 1983) or the antibody (Neaves et al., 1983; Patterson et al., 1984). While qualitative results can be obtained in a few hours, attempts to interpret quantitative data in terms of individual meat species have not been successful (Griffiths & Billington, 1984); evidently the residual blood levels are too variable to correlate with the corresponding meat. Nevertheless, the specificity of immunological recognition implies that more suitable species-specific or organ-specific antigens may be made a future basis of both source identification and level determination of meats and offals.
Heated samples A major disadvantage in species recognition via blood proteins lies in the fact that the procedure is not applicable to the analysis of cooked or retorted meat products. For instance, using immunodiffusion, it has been shown that there is no change in the response of beef to conventional antisera, whether the meat is raw (frozen) or cooked for 1 h in a water bath at 50 °C or 60 °C. Above this temperature there is a marked decrease, and no meat can be detected after similar heating at 80 °C, 100 °C or 121 °C (A. A. Crimes, unpublished). Evidently, heating between 60°C and 80°C progressively destroys the antigenic sites on the blood proteins which respond to the antisera used. Products heated in commercial processes to nominally higher temperatures often contain a core of less severely denatured protein which can be recognised by conventional methods. However, in order to monitor proteins after heating to true temperatures beyond 80 °C, one must either choose a heat-stable antigen or one must 'renature' the heat-denatured antigen to reveal or restore its antigenicity. There are indications that both approaches are feasible, provided that appropriate antisera can be raised.
Methodology for meat species identification: a review
221
Heat-stable antigens Most laboratories currently use conventional antisera raised against mixed native blood proteins from each species under test, because these are commercially available. However, it would be more expedient to base the procedure on antibodies to a selected heat-stable species-specific protein, preferably one which is associated with meat, rather than blood. It has been suggested that the major protein currently detected is heatlabile serum albumin (Hayden, 1978), although this would open the possibility of cross-reaction between species; this danger can be minimised by using antibodies to a serum beta-globulin (Gombocz & Petuely, 1983) applicable to samples heated at 70°C-80°C. Species identification in similarly heated products is also possible via muscle proteins: particularly appropriate are troponin (Hayden, 1977) and myoglobin (Hayden, 1979) as well as crude extracts (Doberstein & Greuel, 1982). The examination of more severely heated products requires antigens which retain their immunological properties even after autoclaving at 120°C for 30min. Such heat-stable antigens have been observed in autoclaved extracts of adrenals and adrenal/kidney tissue; one of these antigens is both species-specific and widely distributed in other organs including muscle, and has been made the basis for an immunodiffusion test (Hayden, 1981). This proved to be applicable to products cooked in air at up to 110 °C for 3.5 h, although the observed core temperature in this case was less than 72°C. The speciation of sterilised meat products heated to true temperatures of 100 °C or 120 °C remains problematic.
Renaturation One of the major difficulties in the examination of severely heat-processed samples is in the preparation of a suitable extract in which the proteins are both soluble and antigenic. Hitherto, the knowledge that denaturation at about 70°C progressively destroys both solubility and antigenicity has encouraged the use of simple aqueous buffers designed to extract the soluble protein which has hopefully retained its ability to interact with antibodies to the native protein. However, when the same problem was considered in the ELISA design for soya protein in meat products by Hitchcock et al. (1981), it was demonstrated that the whole sample could be analysed after 'renaturation'. The primary structure of the proteins is
222
C. H. S. Hitchcock, A. A. Crimes
retained after dissolution in hot aqueous denaturants (e.g. urea, sodium dodecyl sulphate, or guanidine hydrochloride), containing mercaptoethanol to reduce any disulphide cross-links. This solution contains mainly the random-coil structures amenable to electrophoretic analysis. However, direct immunoanalysis is precluded by the presence of significant levels of denaturant, which would interfere with any residual antibody-antigen interaction; as the denaturant is removed (e.g. by dilution or dialysis) the polypeptides refold and some tertiary structure is re-established. While some structure may both escape heat-denaturation during processing and withstand chemical denaturation/renaturation during sample preparation, the final conformations may not always be identical with the original native structure; an analogous situation is the effect of these procedures on enzyme activity (King, 1984). Nevertheless, we have confirmed that the 'renatured' protein retains antigenicity on which a useful immunoassay of severely heated proteinaceous products can be based. Indeed, mixed meat products heated to 100°C for 30min and then 'renatured', fully retain their response to antisera against soya proteins that have not been renatured (Crimes et al., 1984; Olsman et al., 1984). Using these conventional antisera, this response falls somewhat after the meat product has been sterilised at 120°C for 30min. Nevertheless, the retention of antigenicity demonstrates that renaturation provides an important solubilisation procedure in the preparation of heat-processed foods for protein assay. Its application to the determination of kangaroo muscle proteins has been demonstrated by Manz (1983), who examined a renatured solution of heat-processed model products using the ELISA procedure for soya protein with antiserum raised against a crude extract of renatured kangaroo muscle, in which the alpha2-globulins are the active component. These preliminary experiments illustrate the encouraging results now being observed in many laboratories where food proteins are examined specifically, qualitatively and quantitatively, by immunoassay of renatured sample solutions.
CONCLUSIONS The development of methodology for the identification of meat species has made significant progress in individual laboratories over the past few years. No single method is likely to emerge in the near future as superior
Methodology for meat species identification: a review
223
to all others in every situation: several procedures reviewed here merit further investigation towards standardised designs to meet particular speciation problems. The analyst should choose the most expedient procedure, bearing in mind the circumstances expected: number of samples per week; value of rapid results; degree of confidence required; whether differentiation from a given species or identification of unknown species will be necessary; whether the sample will be raw or previously heat-processed; whether the test should cover familiar domestic species, wild species, exotic species or closely related species. Ifa single cut of meat is being examined, results from a small subsample can be taken to apply to the whole piece, and its identification is normally unequivocal. With mixed meat samples a positive result from a properly controlled test leads to the unequivocal qualitative conclusion that meat of a particular species is present. However, a negative result does not necessarily imply its total absence; such a result needs careful interpretation in the knowledge of the limits of detection of the method used. Moreover, conclusions about any batch depend on the subsample being representative of the whole, and multiple sampling may be necessary when non-homogeneous batches are being investigated. It is tempting to consider extending these qualitative methods to yield data about the quantitative levels of meat from each different species in the subsample, as calculated from the observed levels of each speciesspecific protein. The calculation would be analogous to the non-specific determination of total lean meat from observed total nitrogen, but would measure lean meat content from each species separately. Such speciesspecific quantification will only be practicable if the observed levels of species-specific protein in all sources of each meat is found to be constant.
REFERENCES Babiker, S. A., Glover, P. A. & Lawrie, R. A. (1980/1). Meat Science, 5, 473-7. Carnegie, P. R., Collins, M. C. & Ilic, M. Z. (1984). Meat Science, 10, 145 54. Crimes, A. A., Hitchcock, C. H. S. & Wood, R. (1984). J. Assoc. Publ. Analysts, 22, 59-78. Doberstein, K. H. & Greuel, E. (1982). Arch. Lebensmittelhygiene, 33, 133-4. E1 Sayed, L. & E1 Dashlauty, A. (1979). Riv. Ital. Sost. Grasse, 56, 52-8. Gombocz, V. E. & Petuely, F. (1983). Deutsche Lebensmittel-Rundschau, 79, 137-56. Gleeson, L. J. & Walker, P. R. (1984). Aust. Vet. J., 61, 165-7.
224
C. H. S. Hitchcock, A. A. Crimes
Griffiths, N. M. & Billington, M. J. (1984). J. Sci. Food Agric., 35, 909-14. Hayden, A. R. (1977). J. Food Sci., 42, 1189-92. Hayden, A. R. (1978). J. Food Sci., 43, 476-8. Hayden, A. R. (1979). J. Food Sci., 44, 494-500. Hayden, A. R. (1981). J. Food Sci., 46, 1810-13. Hilditch, T. P. & Williams, P. N. (1964). The chemical constitution of natural fats. (4th edn.). Chapman and Hall, London, 102-4. Hitchcock, C. H. S. (1984). In: Control of food quality and food analysis. (Birch, G. G. & Parker, K. J. (Eds)). Elsevier Applied Science Publishers, London and New York, 117-33. Hitchcock, C. H. S. & Crimes, A. A. (1983). Analyt. Proc., 20, 413-15. HitchcoCk, C. H. S., Bailey, F. J., Crimes, A. A., Dean, D. A. G. & Davis, P. J. (1981). J. Sci. Food Agric., 32, 157-65. Jones, D., Shorley, D. & Hitchcock, C. (1982). J. Sci. FoodAgric., 33, 677-85. Jones, A. D., Homan, A. C., Favell, D. J., Hitchcock, C. H. S., Berryman, P., Griffiths, N. M. & Billington, M. J. (1985). Meat Science, 15, 137-47. Kang'ethe, E. K., Jones, S. J. & Patterson, R. L. S. (1982). Meat Science, 7, 229-40. King, N. L. (1984). Meat Science, 11, 59-72. King, N. L. & Kurth, L. (1982). J. Food Sci., 47, 1608-12. Kurth, L. & Shaw, F. D. (1983). Food Technol. in Australia, 35, 328--31. Mageau, R. P., Cutrufelli, M. E., Schwab, B. & Johnston, R. W. (1984). J. Assoc. Off. Anal. Chem., 67, 949-54. Manz, J. (1983). Fleischw., 63, 1767-9. Mattey, M., Parsons, A. L. & Lawrie, R. A. (1970). J. Fd. Technol., 5, 41-.6. Neaves, P., Patel, P. D. & Woods, L. F. J. (1983). BFMIRA Tech. Circ. No. 777. British Food Manufacturing Industries Research Association, Randalls Road, Leatherhead, Surrey, England. Olsman, W. J., Dobbelaere, S. & Hitchcock, C. H. S. (1985). J. Sci. FoodAgric., 36, 499-507. Patterson, R. M., Whittaker, R. G. & Spencer, T. L. (1984). J. Sci. FoodAgric., 35, 1018-23. Payne, E. (1971). J. Sci. Food Agric., 22, 520 2. Sinclair, A. J. & Slattery, W. J. (1982). Aust. Vet. J., 58, 79-80. Slattery, W. J. & Sinclair, A. J. (1983). Aust. Vet. J., 60, 47-51. Swart, K. S. & Wilks, C. R. (1982). Aust. Vet. J., 59, 21-2. Tizard, I. R., Fish, N. A. & Caoili, F. (1982). J. Fd. Protection, 45, 353-5. Verbeke, R. & De Brabander, H. (1979). Vlaams Diergeneesk. Tijdschr., 48, 47-63 (In Dutch). Proc. 25th Meeting of European Meat Research Workers, Budapest, 2, 767-72. Verbeke, R. & De Brabander, H. (1980). Proc. 26th Meeting of European Meat Research Workers, Colorado Springs, 1, 150-3. Whittaker, R. G., Spencer, P. L. and Copland, J. W. (1983). J. Sci. FoodAgric., 34, 1143-8.
Methodology for Meat Species Identification: A Review* C. H. S. Hitchcock & A. A. Crimes Unilever Research, Colworth House, Sharnbrook, Bedford MK44 1LQ, Great Britain (Received: 29 January, 1985)
SUMMARY This review covers methods available for the recognition of the species of animal whose meat ispresent in raw materials or meat products. Isolation of species-specific proteins is cumbersome; however, isoelectric focusing followed by recognition of the pattern of protein bands is particularly effective for speciation. Immunoassay has been established as a powerful technique for the determination of food protein analytes in suitable extracts of unresolved mixtures. Simple immunodiffusion using antisera to serum proteins is sufficient to provide a check on the identity of suspect raw samples; routine control would involve extensive sampling programmes. Enzyme-linked immunosorbent assay procedures are also appropriate, but attempts to quantify individual meat species in mixtures have not been successful due to the variability of residual blood levels. Species identification in heat-processed meat products is hindered by progressive denaturation of the protein markers, leading to loss of solubility and antigenicity; the effects of heating can be minimised by a judicious choice of analyte and its solubilisation by renaturation.
INTRODUCTION The presence of meat in a sample of any food can be established by chemical analysis based on the occurrence of 3-methylhistidine in washed * This review is based on our contribution to the EEC Workshop on 'Biochemical Identification of Meat Species' held on 27th-28th November, 1984, at the Commission of the European Communities in Brussels. The Workshop was organised by Dr R. L. S. Patterson (Meat Research Institute, Langford, Bristol, Great Britain) and financed by the Commission from its budget for the co-ordination of agricultural research. 215
Meat Science 0309-1740/85/$03-30 © Elsevier Applied Science Publishers Ltd, England, 1985. Printed in Great Britain
216
C. H. S. Hitchcock, A. A. Crimes
muscle hydrolysates (Jones et al., 1982). This unusual amino acid arises in vivo by the specific methylation of a particular histidine residue in actin and in myosin in the myofibrillar protein. However, since this biological reaction is common to all species examined, it cannot be used for identification purposes; moreover, its use as a quantitative index for total meat is hindered by variable levels in certain types of manufacturing meat and offals (Jones et al., 1985). Minor species-specific proteins are appropriate for species recognition, but these do not contain such convenient component markers; they must therefore be monitored without hydrolysis, since this would destroy their distinctive features. Two very different approaches to the qualitative detection or quantitative determination of such proteins can be considered. First, a solution of proteins can be separated by electrophoresis or chromatography and the isolated analyte recognised or estimated; secondly, the specific protein can be characterised without separation using specific immunoreagents (Kurth & Shaw, 1983). In both cases, the preparation of the protein solution from a raw meat or a mixed raw product is straightforward; however, heat-processing causes denaturation, interaction and chemical reaction between sample components, which hinders extraction of a proteinaceous analyte from the partially insoluble sample. For this reason, attempts have been made to discriminate between species by analysing the simple extracts for non-protein analytes; this is unlikely to be generally useful since metabolism of these small molecules is less species-specific. The measurement of heat-stable dipeptides (anserine, carnosine, balenine) has had limited success in the discrimination of lamb and pork (Carnegie et al., 1984). Exploitation of fat analysis is illustrated by the use of linolenic acid (Payne, 1971) as a marker for horse fat; nevertheless, this marker can give false negative results. It is a characteristic of horses fed on pasture rather than concentrates (Hilditch & Williams, 1964) and is too variable for use as a reliable parameter (Verbeke & De Brabander, 1980). More complex fat analysis can distinguish lard in tallow (El Sayed & E1 Dashlauty, 1979; Verbeke & De Brabander, 1979). Indeed, the pattern of fatty acids at the 2- and 1,3positions of triglycerides can lead to inferences about the sources of fat (and therefore of meat) present in meat products containing mixtures of pork, beef, horse and chicken. Such multivariate analysis is facilitated by computerised data handling for collection and storage of a large bank of quantitative information; observed data from a particular sample can then be compared and its composition calculated as a best estimate with a given degree of confidence (Verbeke & De Brabander, 1980).
Methodology for meat species identification: a review
217
ELECTROPHORETIC METHODS Electrophoresis is a powerful research technique for the separation of proteins, usually as characteristic band patterns in a supporting gel, visualised, if necessary, by simple non-specific staining, or by enzymological or immunological methods. Separation can be governed by the use of homogeneous gels, concentration-gradient gels, pH-gradient gels or denaturants (such as urea or detergents that dissociate the tertiary protein structures); successful applications to species recognition include isoelectric focusing (IEF) of aqueous extracts of raw meat, based on the coloured myoglobin bands (Sinclair & Slattery, 1982) or on discriminating between appropriate species-specific isoenzymes after specific visualisation (King & Kurth, 1982; Slattery & Sinclair, 1983). Although it is notoriously difficult to standardise conditions of electrophoresis between laboratories, its establishment has been facilitated by the commercial availability of standard buffers, well designed apparatus, laser densitometers and computer-assisted data processing (M. Ansfield, private communication). Gels of 1 mm thickness are normally used for the electrophoresis matrix; ultrathin ( < 100/~m) gels reduce both the time required (by a factor of ten) and expenditure on chemicals (F. Grundhofer, private communication). The relatively high initial capital cost is offset in a centralised laboratory with high sample throughput. Although species-specific proteins in heat-processed samples tend to be denatured and insoluble, simple extraction can yield enough solution for electrophoretic speciation (Babiker et al., 1980/1). Efficient extraction requires the aid of solubilisation agents: for instance, cooked meats can be distinguished after extraction and electrophoresis in 8 M urea and visualisation by non-specific dye-binding (Mattey et al., 1970). Specific detection gives more reliable results, but this can involve extraction into 6 M guanidine hydrochloride, dialysis into 1% Triton X-100, isoelectric focusing in pH-gradient gels and visualisation by renatured enzyme activity, e,g. adenylate or creatine kinase, as demonstrated by King (1984). These lengthy procedures are perhaps most appropriate for definitive confirmatory tests after immunological screening, or for special discrimination between two immunologically related species. The interpretation of electrophoretic data is hampered by the visual complexity of the pherogram; the monitoring of the isoenzyme activity gives only a few bands whose resolution typically allows detection of 1 ~o adulteration of raw beef, illustrating its overall power (Kurth & Shaw, 1983).
218
C. H. S. Hitchcock, A. A. Crimes
I M M U N O L O G I C A L METHODS Immunoassays, reviewed by Hitchcock (1984), based on very specific antigen-antibody interactions, can detect and determine particular analytes in situ in complex mixtures such as biological fluids or food extracts. While there are many analytical procedures in which immunoreagents can be used, the two most appropriate for application in species recognition are the classical Ouchterlony double immunodiffusion technique and the more recent enzyme-linked immunosorbent assay (ELISA). Both are robust and suitable for screening purposes; of the two, ELISA requires less antiserum and less time; it is more objective, more sensitive and, although more complicated, it can be semiautomated. Immunodiffusion
Currently, factory and control laboratories favour simple immunodiffusion in agar gels based on commercial antisera to species-specific blood proteins. If necessary, antisera to unusual species are readily prepared. The properties of each batch must be checked before use, since antisera are difficult to standardise. An aqueous extract of soluble proteins from the meat sample is prepared and aliquots placed in wells cut in a shallow layer (2-3 mm) of agarose gel. Antisera against the species to be tested are placed in an adjacent well. The plate is then incubated to allow antibodies and antigens to diffuse together. If both are present, a visible opaque band will appear in the gel between the wells. Typically, contamination by one species in another can be detected without the use of complex equipment and results are available within 24h (Hayden, 1978; Swart & Wilks, 1982). Using an optimised procedure with commercial precipitating antiserum to species-specific blood proteins, it is possible to detect beef, horse, pig, poultry, sheep and kangaroo serum routinely, and thus infer the presence of the meat; however, it is not possible to distinguish sheep from goat and horse from donkey meat. Subjective detection limits in raw beef are: sheep, 5 ~o; pork, 10 ~ ; horse, 10 ~o; chicken/turkey, 15 ~ and kangaroo, 20 ~o; beef itself could be detected at the 2 ~o level in pork. Of a wide range of meat product ingredients examined, only one interfered: egg protein gave a positive response in the assay for poultry meat, which was based on rabbit anti-hen antiserum (A. A. Crimes, unpublished). False-positive results occasionally arise by non-specific interactions
Methodology for meat species identification: a review
219
leading to a relatively diffuse precipitate which may sometimes be recognised by its abnormal appearance and its solubility in water; it may therefore disappear on soaking the gel in saline. Other reported difficulties have been ascribed to the presence of citrate or ascorbate (Tizard et al., 1982) and mechanically deboned veal (Gleeson & Walker, 1984). Samples of raw meat can be conveniently screened in factory environments and inspection laboratories by double immunodiffusion using a supply of appropriate antisera, standards and agar gels prepared when required; even more convenient would be a standardised kit which provides for individual tests off the shelf. The on-site (overnight) rapid beef identification test (ORBIT) has been designed to verify beef species: filter paper discs are inserted into slits cut into the meat sample and then applied directly to precast gels; the antisera are provided freeze-dried onto similar discs (Mageau et al., 1984). Double immunodiffusion normally takes 24 h, but wherever speed is important, the diffusion can be assisted by the application of an electrical voltage. This procedure of Countercurrent Immunoelectrophoresis (CIE) should perhaps be regarded as accelerated immunodiffusion, to distinguish it from the more sophisticated techniques designed for separation of antigens (e.g. immunoelectrophoresis, Laurell rockets, 2Dimmunoelectrophoresis). CIE is particularly effective where speed and cost are important (T. N. Allsup, private communication). Enzyme-linked immunosorbent assay Semiquantitative determination of antigens by simple immunodiffusion is only possible by subjective comparison of results between various dilutions of sample extract and those of standards; more objective data can be derived from more sophisticated immunoprecipitation techniques. However, enzyme-linked immunosorbent assay (ELISA) is one of several powerful quantitative immunological procedures which are not monitored by precipitation; here the antibody-antigen interaction occurs in a monomolecular layer immobilised on an inert surface and is followed by means of an enzyme chemically bonded to one of the immunoreagents (Hitchcock & Crimes, 1983). ELISA is a well established technique, particularly in clinical laboratories where it provides rapid analysis of biological fluids and tissues for many antigens and antibodies both in qualitative screening and quantitative diagnostic procedures. In food
220
C. H. S. Hitchcock, A. A. Crimes
control laboratories, it has been applied with some success (Crimes et al., 1984; Olsman et al., 1984) to the determination of soya protein in meat products, which serves as a model for all food proteins (Hitchcock et al., 1981). ELISA has therefore been applied to species recognition by choosing species-specific proteins; in particular, the blood proteins tested by immunodiffusion have been demonstrated to be appropriate for uncooked samples in several laboratories using somewhat different ELISA procedures: the initial immobilised species may be the standard antigen (Griffiths & Billington, 1984), the analyte antigen in the sample itself (Kang'ethe et al., 1982; Whittaker et al., 1983) or the antibody (Neaves et al., 1983; Patterson et al., 1984). While qualitative results can be obtained in a few hours, attempts to interpret quantitative data in terms of individual meat species have not been successful (Griffiths & Billington, 1984); evidently the residual blood levels are too variable to correlate with the corresponding meat. Nevertheless, the specificity of immunological recognition implies that more suitable species-specific or organ-specific antigens may be made a future basis of both source identification and level determination of meats and offals.
Heated samples A major disadvantage in species recognition via blood proteins lies in the fact that the procedure is not applicable to the analysis of cooked or retorted meat products. For instance, using immunodiffusion, it has been shown that there is no change in the response of beef to conventional antisera, whether the meat is raw (frozen) or cooked for 1 h in a water bath at 50 °C or 60 °C. Above this temperature there is a marked decrease, and no meat can be detected after similar heating at 80 °C, 100 °C or 121 °C (A. A. Crimes, unpublished). Evidently, heating between 60°C and 80°C progressively destroys the antigenic sites on the blood proteins which respond to the antisera used. Products heated in commercial processes to nominally higher temperatures often contain a core of less severely denatured protein which can be recognised by conventional methods. However, in order to monitor proteins after heating to true temperatures beyond 80 °C, one must either choose a heat-stable antigen or one must 'renature' the heat-denatured antigen to reveal or restore its antigenicity. There are indications that both approaches are feasible, provided that appropriate antisera can be raised.
Methodology for meat species identification: a review
221
Heat-stable antigens Most laboratories currently use conventional antisera raised against mixed native blood proteins from each species under test, because these are commercially available. However, it would be more expedient to base the procedure on antibodies to a selected heat-stable species-specific protein, preferably one which is associated with meat, rather than blood. It has been suggested that the major protein currently detected is heatlabile serum albumin (Hayden, 1978), although this would open the possibility of cross-reaction between species; this danger can be minimised by using antibodies to a serum beta-globulin (Gombocz & Petuely, 1983) applicable to samples heated at 70°C-80°C. Species identification in similarly heated products is also possible via muscle proteins: particularly appropriate are troponin (Hayden, 1977) and myoglobin (Hayden, 1979) as well as crude extracts (Doberstein & Greuel, 1982). The examination of more severely heated products requires antigens which retain their immunological properties even after autoclaving at 120°C for 30min. Such heat-stable antigens have been observed in autoclaved extracts of adrenals and adrenal/kidney tissue; one of these antigens is both species-specific and widely distributed in other organs including muscle, and has been made the basis for an immunodiffusion test (Hayden, 1981). This proved to be applicable to products cooked in air at up to 110 °C for 3.5 h, although the observed core temperature in this case was less than 72°C. The speciation of sterilised meat products heated to true temperatures of 100 °C or 120 °C remains problematic.
Renaturation One of the major difficulties in the examination of severely heat-processed samples is in the preparation of a suitable extract in which the proteins are both soluble and antigenic. Hitherto, the knowledge that denaturation at about 70°C progressively destroys both solubility and antigenicity has encouraged the use of simple aqueous buffers designed to extract the soluble protein which has hopefully retained its ability to interact with antibodies to the native protein. However, when the same problem was considered in the ELISA design for soya protein in meat products by Hitchcock et al. (1981), it was demonstrated that the whole sample could be analysed after 'renaturation'. The primary structure of the proteins is
222
C. H. S. Hitchcock, A. A. Crimes
retained after dissolution in hot aqueous denaturants (e.g. urea, sodium dodecyl sulphate, or guanidine hydrochloride), containing mercaptoethanol to reduce any disulphide cross-links. This solution contains mainly the random-coil structures amenable to electrophoretic analysis. However, direct immunoanalysis is precluded by the presence of significant levels of denaturant, which would interfere with any residual antibody-antigen interaction; as the denaturant is removed (e.g. by dilution or dialysis) the polypeptides refold and some tertiary structure is re-established. While some structure may both escape heat-denaturation during processing and withstand chemical denaturation/renaturation during sample preparation, the final conformations may not always be identical with the original native structure; an analogous situation is the effect of these procedures on enzyme activity (King, 1984). Nevertheless, we have confirmed that the 'renatured' protein retains antigenicity on which a useful immunoassay of severely heated proteinaceous products can be based. Indeed, mixed meat products heated to 100°C for 30min and then 'renatured', fully retain their response to antisera against soya proteins that have not been renatured (Crimes et al., 1984; Olsman et al., 1984). Using these conventional antisera, this response falls somewhat after the meat product has been sterilised at 120°C for 30min. Nevertheless, the retention of antigenicity demonstrates that renaturation provides an important solubilisation procedure in the preparation of heat-processed foods for protein assay. Its application to the determination of kangaroo muscle proteins has been demonstrated by Manz (1983), who examined a renatured solution of heat-processed model products using the ELISA procedure for soya protein with antiserum raised against a crude extract of renatured kangaroo muscle, in which the alpha2-globulins are the active component. These preliminary experiments illustrate the encouraging results now being observed in many laboratories where food proteins are examined specifically, qualitatively and quantitatively, by immunoassay of renatured sample solutions.
CONCLUSIONS The development of methodology for the identification of meat species has made significant progress in individual laboratories over the past few years. No single method is likely to emerge in the near future as superior
Methodology for meat species identification: a review
223
to all others in every situation: several procedures reviewed here merit further investigation towards standardised designs to meet particular speciation problems. The analyst should choose the most expedient procedure, bearing in mind the circumstances expected: number of samples per week; value of rapid results; degree of confidence required; whether differentiation from a given species or identification of unknown species will be necessary; whether the sample will be raw or previously heat-processed; whether the test should cover familiar domestic species, wild species, exotic species or closely related species. Ifa single cut of meat is being examined, results from a small subsample can be taken to apply to the whole piece, and its identification is normally unequivocal. With mixed meat samples a positive result from a properly controlled test leads to the unequivocal qualitative conclusion that meat of a particular species is present. However, a negative result does not necessarily imply its total absence; such a result needs careful interpretation in the knowledge of the limits of detection of the method used. Moreover, conclusions about any batch depend on the subsample being representative of the whole, and multiple sampling may be necessary when non-homogeneous batches are being investigated. It is tempting to consider extending these qualitative methods to yield data about the quantitative levels of meat from each different species in the subsample, as calculated from the observed levels of each speciesspecific protein. The calculation would be analogous to the non-specific determination of total lean meat from observed total nitrogen, but would measure lean meat content from each species separately. Such speciesspecific quantification will only be practicable if the observed levels of species-specific protein in all sources of each meat is found to be constant.
REFERENCES Babiker, S. A., Glover, P. A. & Lawrie, R. A. (1980/1). Meat Science, 5, 473-7. Carnegie, P. R., Collins, M. C. & Ilic, M. Z. (1984). Meat Science, 10, 145 54. Crimes, A. A., Hitchcock, C. H. S. & Wood, R. (1984). J. Assoc. Publ. Analysts, 22, 59-78. Doberstein, K. H. & Greuel, E. (1982). Arch. Lebensmittelhygiene, 33, 133-4. E1 Sayed, L. & E1 Dashlauty, A. (1979). Riv. Ital. Sost. Grasse, 56, 52-8. Gombocz, V. E. & Petuely, F. (1983). Deutsche Lebensmittel-Rundschau, 79, 137-56. Gleeson, L. J. & Walker, P. R. (1984). Aust. Vet. J., 61, 165-7.
224
C. H. S. Hitchcock, A. A. Crimes
Griffiths, N. M. & Billington, M. J. (1984). J. Sci. Food Agric., 35, 909-14. Hayden, A. R. (1977). J. Food Sci., 42, 1189-92. Hayden, A. R. (1978). J. Food Sci., 43, 476-8. Hayden, A. R. (1979). J. Food Sci., 44, 494-500. Hayden, A. R. (1981). J. Food Sci., 46, 1810-13. Hilditch, T. P. & Williams, P. N. (1964). The chemical constitution of natural fats. (4th edn.). Chapman and Hall, London, 102-4. Hitchcock, C. H. S. (1984). In: Control of food quality and food analysis. (Birch, G. G. & Parker, K. J. (Eds)). Elsevier Applied Science Publishers, London and New York, 117-33. Hitchcock, C. H. S. & Crimes, A. A. (1983). Analyt. Proc., 20, 413-15. HitchcoCk, C. H. S., Bailey, F. J., Crimes, A. A., Dean, D. A. G. & Davis, P. J. (1981). J. Sci. Food Agric., 32, 157-65. Jones, D., Shorley, D. & Hitchcock, C. (1982). J. Sci. FoodAgric., 33, 677-85. Jones, A. D., Homan, A. C., Favell, D. J., Hitchcock, C. H. S., Berryman, P., Griffiths, N. M. & Billington, M. J. (1985). Meat Science, 15, 137-47. Kang'ethe, E. K., Jones, S. J. & Patterson, R. L. S. (1982). Meat Science, 7, 229-40. King, N. L. (1984). Meat Science, 11, 59-72. King, N. L. & Kurth, L. (1982). J. Food Sci., 47, 1608-12. Kurth, L. & Shaw, F. D. (1983). Food Technol. in Australia, 35, 328--31. Mageau, R. P., Cutrufelli, M. E., Schwab, B. & Johnston, R. W. (1984). J. Assoc. Off. Anal. Chem., 67, 949-54. Manz, J. (1983). Fleischw., 63, 1767-9. Mattey, M., Parsons, A. L. & Lawrie, R. A. (1970). J. Fd. Technol., 5, 41-.6. Neaves, P., Patel, P. D. & Woods, L. F. J. (1983). BFMIRA Tech. Circ. No. 777. British Food Manufacturing Industries Research Association, Randalls Road, Leatherhead, Surrey, England. Olsman, W. J., Dobbelaere, S. & Hitchcock, C. H. S. (1985). J. Sci. FoodAgric., 36, 499-507. Patterson, R. M., Whittaker, R. G. & Spencer, T. L. (1984). J. Sci. FoodAgric., 35, 1018-23. Payne, E. (1971). J. Sci. Food Agric., 22, 520 2. Sinclair, A. J. & Slattery, W. J. (1982). Aust. Vet. J., 58, 79-80. Slattery, W. J. & Sinclair, A. J. (1983). Aust. Vet. J., 60, 47-51. Swart, K. S. & Wilks, C. R. (1982). Aust. Vet. J., 59, 21-2. Tizard, I. R., Fish, N. A. & Caoili, F. (1982). J. Fd. Protection, 45, 353-5. Verbeke, R. & De Brabander, H. (1979). Vlaams Diergeneesk. Tijdschr., 48, 47-63 (In Dutch). Proc. 25th Meeting of European Meat Research Workers, Budapest, 2, 767-72. Verbeke, R. & De Brabander, H. (1980). Proc. 26th Meeting of European Meat Research Workers, Colorado Springs, 1, 150-3. Whittaker, R. G., Spencer, P. L. and Copland, J. W. (1983). J. Sci. FoodAgric., 34, 1143-8.