Food Chemistry 169 (2015) 241–245

Contents lists available at ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Analytical Methods

An improved UPLC method for the detection of undeclared horse meataddition by using myoglobin as molecular marker

http://dx.doi.org/10.1016/j.foodchem.2014.07.1260308-8146/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +39 0823 274535; fax: +39 0823 274571.E-mail address: antimo.dimaro@unina2.it (A. Di Maro).

1 These authors contributed equally to this research and should be consideredco-first authors.

Antonella M.A. Di Giuseppe a,1, Nicola Giarretta b,1, Martina Lippert b, Valeria Severino a,Antimo Di Maro a,⇑a Department of Environmental, Biological and Pharmaceutical Sciences and Technologies, Second University of Naples, Via Vivaldi 43, I-81100 Caserta, Italyb Freelance, Via Ragozzino 6, I-81037 Capua, Italy

a r t i c l e i n f o a b s t r a c t

Article history:Received 24 October 2013Received in revised form 25 June 2014Accepted 29 July 2014Available online 9 August 2014

Keyword:Food fraudHorse meatMyoglobinSpecies identificationUPLC

In 2013, following the scandal of the presence of undeclared horse meat in various processed beef prod-ucts across the Europe, several researches have been undertaken for the safety of consumer health. In thisframework, an improved UPLC separation method has been developed to detect the presence of horsemyoglobin in raw meat samples. The separation of both horse and beef myoglobins was achieved in onlyseven minutes. The methodology was improved by preparing mixtures with different compositionpercentages of horse and beef meat. By using myoglobin as marker, low amounts (0.50 mg/0.50 g,w/w; �0.1%) of horse meat can be detected and quantified in minced raw meat samples with high repro-ducibility and sensitivity, thus offering a valid alternative to conventional PCR techniques.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Nowadays, consumers demand clear and reliable informationabout meat and meat products they consume. In Europe, recently,the undeclared meat addition and/or substitution with other typesof meat species are common frauds in the meat industry. Nonethe-less, in 2013 Europe has been involved in the scandals of the pres-ence of undeclared horse meat in foods containing raw beef meat(AAVV, 2013; Brown, 2013), for which several food companieshad to withdraw fraudulent products. Although the eating of horsemeat has not a direct food safety issue, the scandal revealed abreakdown in the traceability of the food supply chain, perceivedalso by consumers (Condon, 2013). Furthermore, due to religiousprohibitions, observant Jews consider sinful to eat horse meatand many other animals, therefore the presence of traces of horsemeat in beef is unacceptable (Satter, 2013). Anyhow, in severalcases, the undeclared horse meat presence was not due to manu-facturers of processed beef products but to providers of raw meat(Nau, 2013).

In this framework, European Food Safety Authority hasremarked on the need for increased controls, underlining the

urgency of fast and accurate verification methods able to detectthe presence of different species in raw meat samples (Stoyke,Hamann, Radeck, & Gowik, 2013; Walker, Burns, & Burns, 2013).Therefore, for species identification several analytical methodolo-gies have been developed (Ballin, 2010; Hsieh, 2006). In particular,the methods improved for the detection and quantification ofhorse meat in meat samples are based on: (i) enzymatic assays;(ii) fatty acid and triglyceride analysis; (iii) chromatographic sepa-ration, mass spectrometry, immuno-, gel-, or isoelectric focusingelectrophoresis (IEF) of soluble proteins; and (iv) enzyme linkedimmunoassay (ELISA). Moreover, in the mid-2000’s Food StandardsAgency funded the development of real-time PCR assays for thespecific detection in commercial products (Walker et al., 2013).

In a previously work, our group has proposed the myoglobin(Mb) as a powerful molecular marker to evaluate the presence ofundeclared meat addition in raw beef burgers by using anionicultra-performance liquid chromatography (UPLC). Furthermore,the method was validated to verify undeclared presence of porkmeat in minced beef meat (sensitivity 5%, 50 mg/1 g) (Giarretta,Di Giuseppe, Lippert, Parente, & Di Maro, 2013). Mb was chosenbeing easily detectable because of its high absorbance coefficientat 409 nm and being the principal heme-protein in sarcoplasm(Mancini & Hunt, 2005; Ordway & Garry, 2004).

In the present work, we develop a fast anionic UPLC-visiblemethod for selectively detect the presence of horse meat in rawbeefs, with a sensitivity of �0.1% (0.5 mg/0.50 g).

Table 1UPLC elution program applied for the fast separation of horse and beef Mbs.

Time (min) Flow rate (mL/min) Solvent A (%) Solvent B (%) Curve

0 0.4 80 20 Initial7 0.4 70 30 68 0.4 30 70 69 0.4 30 70 6

10 0.4 80 20 612 0.4 80 20 6

Fig. 1. (a) SDS–PAGE analysis of purified horse and beef myoglobins (lanes 1 and 2,respectively; 3 lg), soluble extracted proteins from horse and beef meat (lanes 3and 4, respectively). M, protein markers. (b) Chromatographic profile of horse andbeef standard myoglobins analysed by using the UPLC system.

242 A.M.A. Di Giuseppe et al. / Food Chemistry 169 (2015) 241–245

2. Materials and methods

2.1. Chemicals and buffers

All chemicals used were of analytical reagent grade. SDS–PAGEreference proteins, sodium nitrite, glycine and sodium hydroxidewere purchased from Sigma–Aldrich Srl (Milan, Italy). PD-10desalting columns were supplied by GE Healthcare (Milan, Italy).Bicinchoninic acid (BCA) kit was purchased from Pierce (Rockford,IL, USA).

The following buffers were used: buffer A, 50 mM sodiumphosphate (Na/P), pH 7.2, containing 50 lM EDTA; buffer B,20 mM glycine�NaOH, pH 9.2; buffer C, 20 mM glycine�NaOH, pH9.2, containing 0.2 M NaCl. The solutions were prepared freshlywith milli-Q water and filtered through a 0.22 lm filter by usinga glass filtration unit (Millipore, Billerica, MA, USA).

2.2. Animal meat

Myocardium of horse (Equus caballus L.) and beef (Bos taurus L.)were obtained from local slaughterhouse and kept frozen at �80 �Cuntil use. Meat samples (thigh muscles and minced meat) of horseand beef were obtained from local hypermarkets and kept frozen at�20 �C until use.

2.3. Extraction and purification of standard myoglobins

Mbs were isolated from horse and beef as previously described(Balestrieri, Colonna, & Irace, 1973; Dosi et al., 2006). For determi-nation of Mb content from both horse and bovine muscle inminced meat, analytical separation methods were used, as previ-ously reported (Dosi et al., 2006). The amount is expressed inmg/g of meat. Samples were analysed in triplicate for each sourceand mean values presented.

2.4. Analytical procedures

Mbs homogeneity was determined first by SDS–PAGE, as previ-ously reported (Dosi et al., 2012), and then by RP-HPLC (Di Maroet al., 2009). Protein concentration was determined with thebicinchoninic acid (BCA) kit, following the manufacturer’s instruc-tions, using bovine serum albumin (BSA) as standard (Smith et al.,1985).

2.5. Preparation of horse and beef meat mixtures

Horse and beef meats were manually mixed in a known weightratio and analysed as reported in Section 2.6. Before mixing, meatwas defatted and triturated in order to obtain a homogeneoussample.

2.6. Sample preparation for UPLC analysis

In order to transform any oxymyoglobin and deoxymyoglobinform to metmyoglobin, meat samples and purified standard horseand beef Mbs were subjected to oxidation by using sodium nitrite(Di Iorio, 1981). This chemical treatment is necessary because Mbsforms show physico-chemical properties leading to different chro-matographic elution times in the anionic chromatography(Yamazaki, Yokota, & Shikama, 1964). Briefly, aliquots (0.5 g) ofhorse and beef meat were homogenised in 1.0 mL of buffer A andthen centrifuged at 64,000g for 30 min, at 4 �C. Then, the superna-tant (�1 mL), containing sarcoplasmic proteins and principallymyoglobin, was incubated with 40 mg (40 mg/mL or 0.579 M) ofsodium nitrite for 10 min at 25 �C. Finally, treated samples were

quickly desalted on a PD-10 column, equilibrated and eluted withbuffer B. The same procedure was performed for the standard. Themetmyoglobin content in treated samples was calculated using aspectrophotometric procedure (Krzywicki, 1982). All the experi-ments were performed in triplicate (n = 3).

2.7. Apparatus and instruments

The LC system used for method development and validationconsisted of an Acquity UPLC with PDA detector, and a separationmodule with a Protein-Pak Hi Res Q column [4.6 � 100 mm, 5 lm(Waters SpA)], at 40 �C operating temperature. Empower software(Waters SpA) was used for data treatment.

2.8. Separation of metmyoglobins by anion exchange chromatographyusing the UPLC system

The analysis was carried out using a quasi-linear gradient(Table 1), made up with buffers B and C (see Section 2.1). ElutedMbs were monitored at 409 nm. The injection volume was 25 lL(when necessary, buffer B was used for sample dilution). Separa-tion of all the myoglobins was achieved with a total run time of7 min. The retention times of horse and beef myoglobins were�2.92 and �4.20 min, respectively. The resolution value (Rs) was4.9 between horse and beef Mbs, higher than 2.0.

Fig. 4. Chromatographic profiles of premixed raw meat samples with 5.0% (a), 1.0%(b), 0.5% (c) and 0.1% (d) horse and 95%, 99%, 99.5% and 99.9% beef meat,respectively, after oxidation treatment and UPLC analysis.

A.M.A. Di Giuseppe et al. / Food Chemistry 169 (2015) 241–245 243

3. Results and discussion

3.1. Purification of standard myoglobins

Standard Mbs were purified from horse (E. caballus) and beef (B.taurus) as previously reported (Dosi et al., 2006, 2012). The purityof Mbs was confirmed by SDS–PAGE analysis (Fig. 1a) and by RP-HPLC (data not shown). The purified Mbs were used as standards.

3.2. Improvement of the UPLC separation method for horse and beefmyoglobins

In our previous work we have optimised a robust and suitableUPLC method for the separation of ostrich (Struthio camelus austral-is), chicken (Gallus gallus domesticus), horse (E. caballus), pig(Sus scrofa domesticus), beef (B. taurus) and water buffalo (Bubalusbubalis) Mbs (Giarretta et al., 2013). Considering the recent scandalof beef products adulterated with horse meat, a revised fast UPLCmethod was developed, allowing the separation of horse and beefmyoglobins in only 7 min (Fig. 1b). This method was achieved usingthe Protein-Pak Hi Res Q column, eluted with quasi-linear gradients(see Methods, Section 2.8). Mbs elution was monitored at 409 nm,because of the higher extinction coefficient. Then, a preliminaryanalysis was carried out with 50% pre-mixture of horse and beef meatsample. Fig. 2a shows the chromatographic profile obtained using 50%pre-mixture raw extract sample, while Fig. 2b shows the 50% pre-mixture raw extract sample after sodium nitrite treatment, necessaryto transform any oxymyoglobin and deoxymyoglobin form present

Fig. 2. Chromatographic profiles of premixed raw burgers with 50% horse and 50% beef meat, without (a) and with (b) oxidation treatment. The upper panels show theabsorbance spectra of the peaks, numbered as reported in the chromatographic profile.

Fig. 3. Standard calibration curve and linearity plots for horse (a) and beef (b) myoglobin quantitation. In the insert, the equation and R2 values of both calibrations are alsoreported.

Table 2Amounts (lg) of horse and beef myoglobin recovered from the analyses of raw beef meats containing different percentages of horse meat. Data are means of triplicate analyses(n = 3).

Beef meat (%) Horse meat (%) Beef meat (g) Horse meat (g) aRecovered beef Mb (lg) aRecovered horse Mb (lg)

100 0 0.500 0.000 22.733 ± 0.984 050 50 0.250 0.250 10.353 ± 0.543 12.806 ± 0.44290 10 0.450 0.050 17.554 ± 0.738 2.583 ± 0.0995 5.0 0.475 0.025 20.276 ± 0.932 1.413 ± 0.05999 1.0 0.495 0.005 20.596 ± 0.791 0.262 ± 0.00999.5 0.5 0.4975 0.0025 21.336 ± 0.989 0.146 ± 0.00799.9 0.1 0.4995 0.0005 22.596 ± 1.153 0.031 ± 0.0010 100 0.000 0.500 0 26.940 ± 1.357

a lg of horse and beef myoglobins were calculated by using the areas of standard horse and beef myoglobins, respectively.

244 A.M.A. Di Giuseppe et al. / Food Chemistry 169 (2015) 241–245

in meat to metmyoglobin. Data show the clear separation betweenhorse and beef Mbs in raw meat samples treatment and therefore,the potential use of these Mbs as marker.

However, the improved short-time UPLC method was used toseparate water buffalo, beef, pig, horse, chicken and ostrich myo-globins. The obtained chromatographic profile is shown inFig. 1S. In only 9 min we have obtained a good separation of allthe six myoglobins. Thus the data proves the robustness and adapt-ability of the developed method, which may certainly still beimproved and/or adapted to specific non-declared raw meataddition.

3.3. Calibration curves and linearity

Calibration curves for horse or beef Mbs were obtained by ana-lysing in triplicate 3.675 lg, 0.735 lg, 0.490 lg, 0.367 lg or0.184 lg and 31.25 lg, 3.125 lg, 0.625 lg or 0.312 lg of standardhorse or beef Mbs, respectively (Fig. 3a and b). The obtained curveequation, calculated with the linear regression method, was uti-lised for the quantitation. The response of the detector was linearin the analysed range for both horse and beef standard Mbs, withan identical correlation coefficient (R2 = 0.999).

3.4. Limits of detection and quantitation

For the determination of the limit of detection (LOD) and thelimit of quantitation (LOQ) a specific calibration curve was usedwith samples containing the horse or beef myoglobin in the rangeof detection limit (DL) and quantitation limit (QL). The LOD forhorse myoglobin was 0.3 lg/mL, and the LOQ was 1.0 lg/mL,whereas for beef myoglobin the LOD was 0.2 lg/mL, while theLOQ was 0.6 lg/mL. LOD and LOQ values data are very low, thusconfirming that anion chromatography coupled with the visibledetector, is a selective detection system to determine myoglobins.

3.5. Raw minced meat analyses

Raw burgers with different percentages of horse and beefminced meats were prepared in triplicate. The relative content ofhorse and beef Mbs was determined by using the UPLC elution con-ditions as previously described. Before UPLC analysis, meat sam-ples (0.5 g of bulk mixed minced meat) were subjected to theextraction and oxidation procedure (Giarretta et al., 2013). InFig. 4a–d are shown representative chromatographic profiles ofpremixed raw burgers at 5.0%, 1.0%, 0.5% and 0.1% of horse and95.0%, 99.0%, 99.5% and 99.9% of beef minced meat, respectively.Results obtained for raw beef burger analyses are summarised inTable 2. The presence of horse and beef Mbs was detected andaccurately quantitated in all the pre-mixed samples (0.1–50%). Inparticular, in the sample with 0.1% horse meat the presence ofhorse Mb is clearly detectable allowing its quantitation

(0.031 lg). Moreover, horse and beef Mb amount in raw mixedsamples was 4.56 ± 0.31 mg/g and 3.16 ± 0.17 mg/g, respectively,with a mean recovery value of about 50%. Furthermore, followingUPLC analyses, Mbs showed an average %CV (variation coefficient)of 1.97% and 0.59% for horse and beef, respectively. Actually, thedetection of undeclared horse meat addition is performed by usingPCR techniques, with a detection limit of �1% and �0.05% in thecase of real time PCR (TaqMan). Our UPLC method, with an exper-imentally determined detection limit of 0.1% (w/w), is thereforecomparable with the common commercial kit used. Moreoverour methodology is very fast (only 7 min for horse and beef myo-globins separation and quantification) and not expensive in com-parison with the traditional PCR techniques.

Nevertheless, as previously reported, myoglobin content inhorse and beef tissues is variable considering muscle type andage (Cornforth & Jayasingh, 2004; Keeton & Eddy, 2004), thus fur-ther studies are needed to verify the lower detection limit on mix-tures, prepared with different meat cuts with known age.

4. Conclusion

This study reports the development of a fast method (7 min) forthe separation and quantification of horse myoglobin in raw beefmeat samples. The market analysis reveals that the food industrymainly relies on PCR techniques to detect horse meat in beef. Inparticular, real time PCR (TaqMan) is commonly used with a detec-tion limit of �0.05% (w/w) for horse meat. The method presentedin this paper showing an experimentally determined limit of detec-tion of 0.1% (w/w) of horse meat, comparable to the common com-mercial kit used.

Undoubtedly, the approach described in this study should beconsidered as a very low-cost and anti-fraud test, being rapidlytransferred to control laboratories and thus providing a powerfultool to prove the presence at horse meat in raw meat productsand to protect customers from fraudulence.

Acknowledgements

This study was made possible by care and abnegation of all par-ticipants, despite the absence of dedicated funds and chronic diffi-culties afflicting the Italian scientific community. We acknowledgethe Waters Italia (Milan) and, in particular, Dr. Andrea Perissi(Application Specialist) for their assistance.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.foodchem.2014.07.126.

A.M.A. Di Giuseppe et al. / Food Chemistry 169 (2015) 241–245 245

References

AAVV (2013). Horse meat: The real scandal (Editorial). New Scientist, 217, 3.Balestrieri, C., Colonna, G., & Irace, G. (1973). The skeletal muscle myoglobin of the

water buffalo (Bos bubalus L.). Comparative Biochemistry and Physiology Part B:Biochemistry and Molecular Biology, 46(4), 667–672.

Ballin, N. Z. (2010). Authentication of meat and meat products. Meat Science, 86(3),577–587.

Brown, B. (2013). Findus beef lasagne contained up to 100% horsemeat, FSA says (7February). In BBC News.

Condon, D. (2013). Horse meat health risks ‘not a concern’ (21 February). In IrishHealth. <irishhealth.com>.

Cornforth, D. P., & Jayasingh, P. (2004). Chemical and physical characteristics ofmeat: Colour and pigment. In W. K. Jensen, C. Devine, & M. Dikeman (Eds.).Encyclopedia of meat sciences (Vol. 1, pp. 249–256). Oxford, UK: Elsevier ScienceLtd.

Di Iorio, E. E. (1981). Preparation of derivatives of ferrous and ferric hemoglobin.Methods in Enzymology, 76, 57–72.

Di Maro, A., Chambery, A., Carafa, V., Costantini, S., Colonna, G., & Parente, A. (2009).Structural characterization and comparative modeling of PD-Ls 1-3, type 1ribosome-inactivating proteins from summer leaves of Phytolacca dioica L.Biochimie, 91(3), 352–363.

Dosi, R., Carusone, A., Chambery, A., Severino, V., Parente, A., & Di Maro, A. (2012).Rapid primary structure determination of myoglobins by a complementaryapproach based on mass spectrometry and Edman degradation. Food Chemistry,133(4), 1646–1652.

Dosi, R., Di Maro, A., Chambery, A., Colonna, G., Costantini, S., Geraci, G., et al. (2006).Characterization and kinetics studies of water buffalo (Bubalus bubalis)myoglobin. Comparative Biochemistry and Physiology Part B: Biochemistry andMolecular Biology, 145(2), 230–238.

Giarretta, N., Di Giuseppe, A. M. A., Lippert, M., Parente, A., & Di Maro, A. (2013).Myoglobin as marker in meat adulteration: A UPLC method for determining thepresence of pork meat in raw beef burger. Food Chemistry, 141, 1814–1820.

Hsieh, Y. H. P. (2006). Meat species identification. In Y. H. Hui (Ed.). Handbook of FoodScience, Technology, and Engineering (Vol. 1, pp. 1–19). Boca Raton: CRC Press.

Keeton, J. T., & Eddy, S. (2004). Chemical and physical characteristics of meat: Chemicalcomposition. In W. K. Jensen, C. Devine, & M. Dikeman (Eds.). Encyclopedia of meatsciences (Vol. 1, pp. 210–218). Oxford, UK: Elsevier Science Ltd.

Krzywicki, K. (1982). The determination of haem pigments in meat. Meat Science,7(1), 29–36.

Mancini, R. A., & Hunt, M. C. (2005). Current research in meat color. Meat Science,71(1), 100–121.

Nau, J.-Y. (2013). Horse meat: Early lessons from a ‘‘scandal’’. Revue Medicale Suisse,9, 532–533.

Ordway, G. A., & Garry, D. J. (2004). Myoglobin: An essential hemoprotein in striatedmuscle. Journal of Experimental Biology, 207(Pt 20), 3441–3446.

Satter, R. (2013). Horsemeat found in burger in Ireland (15 January). In HuffingtonPost. <http://www.huffingtonpost.com/>.

Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gartner, F. H., Provenzano,M. D., et al. (1985). Measurement of protein using bicinchoninic acid. AnalyticalBiochemistry, 150(1), 76–85.

Stoyke, M., Hamann, F., Radeck, W., & Gowik, P. (2013). Tasks of the NRLs withrespect to crisis management: Contribution of the EURL/NRL at the BVL asexemplified by the horse meat scandal. Journal fur Verbraucherschutz undLebensmittelsicherheit, 8, 91–100.

Walker, M. J., Burns, M., & Burns, D. T. (2013). Horse meat in beef products-speciessubstitution 2013. Journal of the Association of Public Analysts, 41, 67–106.

Yamazaki, I., Yokota, K. N., & Shikama, K. (1964). Preparation of crystallineoxymyoglobin from horse heart. The Journal of Biological Chemistry, 239(12),4151–4153.

Di Giuseppe et al., Title: An improved UPLC method for the detection of undeclared horse meat

addition by using myoglobin as molecular marker.

Figure 1S

Fig. 1S. Elution profile of ostrich, chicken, horse, pig, beef and water buffalo myoglobins with

improved short-time UPLC method.