November 12, 2016 | Author: Bhoyarlonsy | Category: N/A
Food Control 18 (2007) 1149–1158 www.elsevier.com/locate/foodcont
A TaqMan real-time PCR system for the identification and quantification of bovine DNA in meats, milks and cheeses Chun-Lai Zhang *, Mark R. Fowler, Nigel W. Scott, Graham Lawson, Adrian Slater
*
Forensics Research Unit/Systems Biology Research Laboratory, Faculty of Health and Life Sciences, De Montfort University, Hawthorn Building, The Gateway, Leicester, LE1 9BH, UK Received 22 January 2006; received in revised form 9 July 2006; accepted 16 July 2006
Abstract Accurate quantitative assays are required for enforcing food labelling procedures and preventing food ingredient contamination, misdescription and fraud. Simplex and duplex TaqMan real-time PCR systems have been tested for the identification and quantification of DNA in meat, milk and cheese. DNA was isolated from meat and cheese using a standard CTAB protocol and from milk using a Promega Wizard Magnetic kit and purified by Qiagen silicon spin columns. High quality DNA isolated from beef mince was used for standard curve construction in the TaqMan real-time PCR assay using a bovine-specific primer pair for the mitochondrial cytb gene and a FAM-labelled mammalian-specific cytb probe. The real-time PCR assay can quantitatively detect as little as 35 pg bovine DNA and showed no cross-reaction with ovine, caprine or porcine DNA. The system has been successfully used to measure bovine DNA in fresh and processed meat, milk and cheese, and will prove useful for bovine species identification and quantitative authentication of animal-derived products. 2006 Elsevier Ltd. All rights reserved. Keywords: Real-time PCR; Bovine DNA; Quantitative detection; Meat; Milk; Cheese
1. Introduction Food safety, quality and composition have become the subjects of increasing public concern. Consumers have been given more choices with regard to food composition and dietary requirements via food labels. A number of people are allergic to specific molecules in meats, milks or cheeses. Healthy diet followers tend to prefer chicken instead of beef, pork or lamb, due to its low dietary fat content. Various religious groups avoid specific meats such as beef or pork; whilst vegetarians choose not to consume any meat. Each constituency has an interest in ensuring the authenticity of the foods that they consume. *
Corresponding authors. E-mail addresses:
[email protected] (C.-L. Zhang),
[email protected] (A. Slater). 0956-7135/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodcont.2006.07.018
The fraudulent misdescription of food contents on product labels is a widespread problem, particularly with high added-value products commanding a premium price (Woolfe & Primrose, 2004). There can be intentional or unintentional contamination in the production chain or during processing. Proving conclusively that adulteration or contamination has occurred requires the detection and quantification of food constituents. This can be difficult because the materials replaced are often biochemically very similar and food matrices are extremely complex and variable. Lipid, protein and DNA based methods have been established for food identification. Lipid analysis is only applicable for gross measurement of animal-derived fats (Lumley, 1996; Saeed, Ali, Abdul Rahman, & Sawaya, 1989). Protein-based methods such as high performance liquid chromatography (HPLC) (Espinoza, Kirms,
1150
C.-L. Zhang et al. / Food Control 18 (2007) 1149–1158
& Filipek, 1996), enzyme-linked immunosorbent assays (ELISA) (Chen & Hsieh, 2000) or isoelectric focusing protein profiles (Skarpeid, Kvaal, & Hildrum, 1998) are effective mainly for unprocessed food and are unable to differentiate species such as lamb and goat or chicken and turkey. Both require complicated procedures and it has proved difficult to accurately quantify the analytes in a short time (Mayer, 2005). A variety of DNA-based methods including polymerase chain reaction (PCR) amplification, RFLP mapping and microarray gene chip assay have now been successfully adapted for the detection of food substitution (Peter, Bru¨nen-Nieweler, Cammann, & Bo¨rchers, 2004; Woolfe & Primrose, 2004). DNA based methods have been well received because of the relative stability of the DNA molecule under extreme conditions and its efficient amplification by PCR (Lanzilao, Burgalassi, Fancelli, Settimelli, & Fani, 2005; Matsunaga et al., 1999; Mayer, Hofflein, Luthy, & Candrian, 1995; Sun & Lin, 2003; Zhang, Zheng, Zhou, Ouyang, & Li, 1999). The limitations of standard PCR assays include the insensitivity and lack of quantitation of end-point analysis, and the dependence on a low throughput technique (agarose gel electrophoresis) for analysis of the products. Real-time PCR assay has provided sensitive and safe solutions by monitoring PCR products continuously using fluorescent markers (Heid, Stevens, Livak, & Williams, 1996; Holland, Abramson, Watson, & Gelfand, 1991). Several real-time PCR methodologies have been developed for meat species identification and quantification using low copy nuclear genes (Laube et al., 2003), high copy genomic short interspersed elements (Walker et al., 2003) or mitochondrial genes (Dooley, Paine, Garrett, & Brown, 2004; Hird et al., 2004; Lo´pez-Andreo, Lugo, Garrido-Pertierra, Prieto, & Puyet, 2005). The real-time system is applicable to processed or mixed meats as well (Lo´pez-Andreo et al., 2005). Cows’ milk is more widely available and cheaper than that of goat and water buffalo. Cows’ milk is also processed in large quantities to produce a range of dairy produce including a wide variety of cheeses. On the other hand, specialist cheeses such as Greek Feta cheese made from goats’ and sheep’s milk and Italian mozzarella di bufala campana cheese made from water buffalo milk, both registered by European law with the Protected Designation of Origin (PDO), have become widely accepted throughout the EU and command a premium price because of their production from more costly milks than cows’ milk. It is also more difficult to stretch out mozzarella cheese prepared from buffalo milk and to spin it mechanically due to the different rheologic characteristics of buffalo milk casein compared to cows’ milk casein. Adulteration of goat and water buffalo milk and their products by cows’ milk is therefore economically attractive. Their products are traditionally tested for adulteration by immunological and/or electrophoretic methods (Amigo, Ramos, Calhau, & Barbosa, 1992; Cerquaglia & Avellini, 2004; Hurley, Ireland, Coleman, & Williams, 2004; Levieux & Venien, 1994; Mayer, 2005;
Mimmo & Pagani, 1998). Identification of milk by PCR amplification of DNA is based on the presence of mammalian somatic cells in the milk (Herman, 2001). Several simplex PCR procedures have been developed for species identification in milk, cheese and yogurt based on primers designed to amplify a number of mitochondrial genes: cytochrome b (cytb) (Bania, Ugorski, Polanowsk, & Adamczyk, 2001; Di Pinto, Conversano, Forte, Novello, & Tantillo, 2004; Herman, 2001), D-loop region (Maudet & Taberlet, 2001), 12S ribosomal RNA gene (Lo´pezCalleja et al., 2004; Lo´pez-Calleja et al., 2005), cytochrome oxidase II (Mayer, 2005), cytochrome oxidase I (Feligini et al., 2005), and nuclear encoded genes e.g. coat colour MC1R (Maudet & Taberlet, 2002). Procedures have also been developed for duplex PCR based on cow and buffalospecific cytb primers (Bottero, Civera, Anastasio, Turi, & Rosati, 2002; Rea et al., 2001), bovine and ovine 12S and 16S rRNA genes (Mafra, Ferreira, Faria, & Oliveira, 2004), and multiplex PCR based on bovine, ovine and caprine 12S and 16S rRNA genes (Bottero et al., 2003). Procedures based on polymerase chain reaction-restriction fragment length polymorphism have also been developed using cytb primers to differentiate mozzarella cheese made from water buffalo milk and from less expensive bovine milk and also Feta cheeses made from bovine, ovine, and caprine milk (Branciari, Nijman, Plas, Di Antonio, & Lenstra, 2000) and ovine yogurt (Stefos et al., 2004). However, most of these procedures are not applicable for accurate quantitative measurement and have the disadvantages of conventional PCR discussed above. This study showed that TaqMan real-time PCR is applicable to the authentication of milk and cheese. 2. Materials and methods 2.1. Sample preparation Fresh and processed meats, milks and cheeses were purchased from several national food retailers and/or producers (randomly coded as A–F) in Leicester, UK (Table 1). Meats and cheeses were cut into small pieces with a hand knife. DNA was isolated using a standard cetyltrimethylammonium bromide (CTAB) method (Murray & Thompson, 1980; Zhang et al., 2001). Briefly, about 1.8 g of chopped tissues were mixed with 5 ml of extraction buffer and incubated at 65 C for 2 h. The above mixture was extracted twice with an equal volume of chloroform and an equal volume of isopropanol was added to the aqueous fraction. After centrifugation the precipitate was washed with ethanol. The dried pellet was dissolved in 0.5 ml sterilised Millipore MQ water. DNA from milks was isolated by the Promega Wizard Magnetic kit following the manufacturer’s instructions as only a small amount of DNA was extracted from fresh whole milk using the CTAB method. Firstly, 3 ml of milk was centrifuged at 13,000 rpm for 10 min. The sediment portions were collected and mixed with 0.4 ml lysis buffer
C.-L. Zhang et al. / Food Control 18 (2007) 1149–1158
1151
Table 1 Commercial food products used in general and real-time PCR assays Products
Retailers and/or producers
Descriptions
Beef mince Lamb mince Pork mince Chicken portions Turkey mince Corned beef Steak pie Tinned oxtail soup Fresh whole cows’ milk Fresh whole cows’ milk Fresh goats’ milk Cheddar cheese Cheese onion and chive quiche Somerset goats cheese Welsh goats cheese Italian mozzarella cheese ball Italian mozzarella cheese
E B A A B B B C/F B C B/D B A B A B A
Fresh minced beef Fresh minced lamb Fresh minced pork Frozen chicken portion meat Fresh minced turkey Processed beef meat Cooked beef in pie Prepared soup with vegetables and oxtail Standardized pasteurised homogenised cows’ milk Standardized pasteurised homogenised cows’ milk Fresh pasteurised homogenised goats’ milk Contains milk Processed quiche, containing milk, cheese and egg Contains milk Contains goats’ milk Made in Italy, contains milk Made in Italy, contains milk
A and 4 ll of RNaseA. This step avoided further processing of the cream and skimmed milk fractions which contain lower concentrations of DNA than the sediment (Poms, Glossl, & Foissy, 2001). Further DNA extraction procedures followed the manufacture’s instructions. The isolated DNA was eluted from the magnetic beads with 50 ll of MQ water. All DNA preparations were purified through a Qiagen silicon spin column and dissolved in MQ water. Initial DNA analysis was carried out using agarose gel electrophoresis (1% SeaKem LE agarose gel containing ethidium bromide). DNA aliquots (12 ll for meats, 15 ll for milks and 20 ll for cheeses) were loaded into each well and electrophoresed at 75 V/cm for 1.5 h. Electrophoresis gels were visualized by UV transillumination, photographed and processed using GeneSnap software. DNA amount and quality were determined with a Unicam Hekios spectrophotometer at 260 nm and 280 nm. DNA concentrations of various samples were calculated and further DNA samples were obtained by dilution with MQ water or mixed with lamb and pork, or chicken DNA (see Section 3 for details). For serial dilution, 3 ll of DNA solution was transferred into 27 ll MQ water and mixed thoroughly. A water buffalo milk DNA sample (40 ng/ll) and a goat blood DNA sample (350 ng/ll) were provided by Dr. Isabel Gonza´lez, Universidad Complutense de Madrid, Spain. 2.2. PCR primers and probes PCR primers and probes were based on those described by Dooley et al. (2004). Bovine-specific cytochrome b gene (cytb) primers were Bovcytbf: CGG AGT AAT CCT TCT GCT CAC AGT, Bovcytbr: GGA TTG CTG ATA AGA GGT TGG TG to amplify a 116 bp fragment. Ovinespecific cytb gene primers were Ovicytbf: GAG TAA TCC TCC TAT TTG CGA CA, Ovicytbr: AGG TTT
GTG CCA ATA TAT GGA ATT to amplify a 133 bp fragment. The universal mammalian-specific cytb gene probe (mammalcytprobe) was TGA GGA CAA ATA TCA TCA TTC TGA GGA GCW ARG TYA. A fluorescent dye, 6-carboxyfluorescein (FAM) was attached to the 5 0 end of the probe. The quencher moiety, 6-carboxytetramethylrhodamine (TAMRA), was added to the 3 0 end of the probe. The pair of chicken-specific cytb primers was: Chicytbf: AGC AAT TCC CTA CAT TGG ACA CA, Chicytbr: GAT GAT AGT AAT ACC TGC GAT TGC A to amplify a 133 bp fragment. A universal poultry-specific cytb gene probe was ACA ACC CAA CCC TTA CCC GAT TCT TC. TET, 6-carboxy-4,7,2 0 ,7 0 -tetrachlorofluorescein, a fluorescent dye, was attached to the 5 0 end of the probe. The quencher moiety TAMRA was added to the 3 0 end of the probe. Turkey and pork-specific cytb primer pairs were the same as those described by Dooley et al. (2004). The PCR primers were synthesized by Invitrogen and the TaqMan probes were synthesized by Applied Biosystems. The alignment of selected animal cytb sequences was performed using the ClustalW programme (Chenna et al., 2003) and shown in Fig. 1. 2.3. Conventional PCR protocol MJ Research PTC-200 or Techne A thermal cyclers were used in this study. Simple PCR reaction mixtures (50 ll) comprised MQ water 38 ll, 10· buffer 5 ll, 25 mM MgCl2 solution 2 ll, 10 mM each dNTPs 0.5 ll, 5 unit/ll Promega Taq polymerase 0.5 ll, primers (25 lM) 1.5 ll each and DNA (32.5–37.5 ng/ll) 1 ll. MQ water (1 ll) instead of DNA was used as a negative control. PCR cycling parameters were 95 C 4 min, followed by 35 cycles of 94 C 30 s, 60 C 1 min, 72 C 1 min, and final extension at 72 C 10 min. Duplex PCR reaction mixtures were performed using the same PCR conditions, except that 0.5–2 ll of each primer (25 lM) was used per 50 ll reaction (see Section 3.4).
1152
C.-L. Zhang et al. / Food Control 18 (2007) 1149–1158
Fig. 1. Alignment of selected animal cytb sequences using the ClustalW programme indicating the position of the mammalcytprobe, and bovine and ovine-specific cytb primers. Bovcytbf and complementary sequence to Bovcytbr are boxed and shown in red colour. Ovicytbf and complementary sequence to Ovicytbr are boxed and shown in green colour. The cytb sequences shown are: OvisX56284 for lamb (Ovis aries) X56284, CapraX56289 for goat (Capra hircus) X56289, BubalusD82894 for water buffalo (Bubalus bubalus) D82894, BosD34635 for beef (Bos taurus) D34635, SusX56295 for pork (Sus scrofa) X56295, GallusL08376 for chicken (Gallus gallus) L08376 and MeleagL08381 for turkey (Meleagris gallopavo) L08381. * indicates identical nucleotide base. (For interpretation of the references in color in this figure legend, the reader is referred to the web version of this article.)
2.4. Real-time PCR protocol A MJ Research real-time PCR system, Chromo4 Continuous Fluorescence Detector, was used for these experiments. Real-time PCR cycling parameters were optimized based on the Chromo4 Fluorescence Detector operation manual as: 50 C 2 min, 95 C 10 min, followed by 55 cycles of 95 C 15 s, 60 C 45 s, reading products, 72 C 1 min, then 72 C 10 min and finally, melting point analysis. This programme is controlled by Opticon Monitor 3 software. Experiments were repeated three to five times. Real-time PCR reaction mixtures (25 ll) using the TaqMan master mix kit (Applied Biosystems) were: MQ water 10 ll, 2· TaqMan master mix 12.5 ll, primers (25 lM) 0.5 ll each, probe (10 lM) 0.5 ll, DNA (minute amount up to 625 ng/ll) 1 ll. The 2· TaqMan master mix is optimized for 5 0 nuclease assay using TaqMan probes and contains ROX as a passive reference dye, 2 units of AmpliTaq Gold DNA Polymerase, 0.4 units of AmpErase uracil DNA glycosylase (UNG), 400 lM dATP, dCTP, dGTP with 800 lM dUTP and 6 mM MgCl2. Duplex PCR reaction mixtures were the same as above with the inclusion of additional primers (0.5 ll each of bovine-specific cytb primers and 1 ll each of chicken-specific cytb primers) and probes (0.5 ll each of mammalian and poultry-specific cytb probe).
General Promega Taq DNA polymerase and buffersbased PCR mixture included MQ water 12.75 ll, 10· reaction buffer 2.5 ll, 25 mM MgCl2 solution 5 ll, 10 mM each dNTPs 1 ll, 5 u/ll Taq polymerase 0.25 ll, primers (25 lM) 1 ll each, probe (10 lM) 0.5 ll, DNA solution 1 ll. MQ water (1 ll) instead of DNA was used as a negative control. 2.5. Gel electrophoresis of PCR products Gel electrophoresis with SeaKem LE agarose (1.7% for general PCR, 3% for duplex PCR) gel containing ethidium bromide was used to separate the PCR products. PCR (13 ll) products was loaded onto each well. Electrophoretic profiles reflecting different molecular sizes of PCR products were visualized and analysed as described in Section 2.1. 2.6. Data analysis Primary real-time PCR data were analysed by the Opticon Monitor 3 software and the threshold cycle (Ct) was calculated. Replicate standard curves of Ct value (Y) vs log10[DNA amount] (X) were analysed using Microsoft Excel software and a linear regression equation of the Ct value plotted against the log10[DNA amount] was calculated.
C.-L. Zhang et al. / Food Control 18 (2007) 1149–1158
3. Results and discussion 3.1. Extraction of DNA from fresh and processed foods DNA was isolated from fresh meats, processed meats and cheeses using the standard CTAB method, which was very effective for solid products, and analysed by agarose gel electrophoresis (Fig. 2). DNA extracted from fresh beef mince was of relatively good quality, as determined by an A260/ A280 ratio of 1.73 (Table 2) and lack of degradation (Fig 2), and high yield (357.8 lg/g fresh tissue), whilst DNA obtained from cooked, processed cows’ meats was of lower quality (A260/A280 ratio of 1.35–1.69), highly degraded (very little high molecular weight DNA was observed in oxtail
Fig. 2. Gel electrophoresis of total DNA extracted from meats, milks and cheeses. Lane 1, fresh beef mince DNA; lane 2, fresh lamb mince DNA; lane 3, corned beef DNA; lane 4, steak pie DNA; lane 5, tinned oxtail soup DNA; lane 6, C cows’ milk DNA; lane 7, B cows’ milk DNA; lane 8, goats’ milk DNA; lane 9, Welsh goats cheese DNA; lane 10, B Mozzarella cheese DNA; lane 11, cheese quiche DNA. Lane M1, HindIII digested k DNA; lane M2, 1 kb DNA ladder.
1153
soup) and low yield (33–322 lg/g fresh weight). DNA isolated from cheeses and cooked dairy products (quiche) was also degraded with a yield of 34–349.6 lg/g fresh weight (Table 3). Using the Wizard Magnetic kit (a mini-prep system) small amount (2.4–3 lg/ml) of DNA was extracted from pasteurised cows’ milks. This was not degraded (Fig. 2) and had A260/A280 ratios of 1.44–2.0. The DNA yields of various products were calculated from the A260 values and are included in Tables 2 and 3 for reference though these values obtained for highly degraded DNA or samples with a low A260/A280 ratio may not be reliable. 3.2. Specificity of bovine-specific and ovine-specific cytb primers In conventional PCR analysis, the pair of bovine-specific cytb primers amplified bovine DNA isolated from beef meat and produced a 116 bp fragment using 1 mM MgCl2 (lane 1 in Fig. 3a). No PCR product was obtained using the bovine-specific primers with DNA extracted from lamb, pork, goat, turkey and chicken tissue and water buffalo milk. The pork, turkey and chicken DNA were all PCR positive with corresponding species-specific primers, indicating that the negative reaction was not the result of poor quality DNA or PCR inhibitory contaminants (data not shown). The ovine-specific cytb primers amplified the lamb mince DNA and produced a 133 bp fragment (lane 10 in Fig. 3a), but did not amplify DNA from beef, pork turkey and water buffalo samples. The ovine cytb primers did, however, amplify a 133 bp fragment from goat DNA samples at a low efficiency. Bottero et al. (2003) reported that ovine-specific cytb gene primer amplified caprine DNAs from a few breeds.
Table 2 Ct values and their derived target DNA amount for DNA extracted from fresh or frozen and processed meats and mixed DNA samples measured by realtime PCR with bovine-specific cytb primers and a FAM-labelled mammalian-specific cytb probe DNA samples
DNA yield (lg/g fresh tissue)
A260/A280
Input DNA amount (ng)
Ct valuea
Target DNA amount (ng) derived from Ct value
Efficiency (target DNA/input DNA · 100)%
Beef DNA Beef DNA Beef DNA Beef DNA Beef DNA Beef DNA Beef DNA Lamb DNA Pork DNA Chicken DNA Turkey DNA Beef DNA mixed with lamb and pork DNA Beef DNA mixed with lamb and pork DNA Corned beef Steak pie Tinned oxtail soup
357.8 357.8 357.8 357.8 357.8 357.8 357.8 498.2 673.0 1104.4 2157.2 N/A
1.73 ± 0.17 1.73 ± 0.17 1.73 ± 0.17 1.73 ± 0.17 1.73 ± 0.17 1.73 ± 0.17 1.73 ± 0.17 1.81 ± 0.14 2.12 ± 0.05 1.90 ± 0.25 2.11 ± 0.01 N/A
18.73 ± 1.90 21.12 ± 1.07 24.71 ± 2.42 27.77 ± 0.95 32.1 ± 0.83 37.03 ± 3.35 55 54.87 44.23 50.4 51.99 25.81 ± 1.11
350 35 3.5 0.35 0.035 0.0035 N/D N/D N/D N/D N/D 1.93
N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 275.71
N/A
N/A
350 35 3.5 0.35 0.035 0.0035 0.00035 37.5 32.5 70 80 0.7 + 37.5 (lamb) + 32.5 (pork) 0.07 + 37.5 (lamb) + 32.5 (pork) 500 200 60
30.67 ± 1.08
0.088
125.71
17.94 ± 2.46 19.30 ± 1.17 35.21 ± 7.01
287.45 121.12 0.0049
57.49 60.56 0.0082
a
32.7 189.2 322.0
1.69 ± 0.08 1.56 ± 0.22 1.35 ± 0.15
Data (average ± SE) represent three repeats. N/D, not detectable. N/A, not applicable.
1154
C.-L. Zhang et al. / Food Control 18 (2007) 1149–1158
Table 3 Ct values and their derived target DNA amount for DNA isolated from milks and cheeses measured by real-time PCR with bovine-specific cytb primers and a FAM-labelled mammalian-specific cytb probe DNA samples
DNA yield (lg/g tissue)
A260/A280
Input DNA amount (ng)
Ct valuea
Target DNA amount (ng) derived from Ct value
Efficiency (target DNA/input DNA · 100)%
B cows’ milkb C cows’ milkb Cheddar cheese Cheese quiche B Italian mozzarella cheese ball A Italian mozzarella cheese Fresh goats’ milkb Somerset goats cheese Welsh goats cheese
3.0 2.4 180.0 258.8 34.0 35.0 0.75 349.6 162.7
1.44 ± 0.23 2.0 1.66 1.27 1.48 ± 0.18 1.63 ± 0.37 2.00 1.86 ± 0.09 1.73 ± 0.06
25 20 5 50 15 150 15.6 200 625
25.50 ± 0.49 26.05 ± 0.70 28.13 ± 1.76 29.68 ± 1.02 31.86 ± 2.44 24.12 ± 2.39 55 55 55
2.36 1.66 0.443 0.17 0.041 5.66 N/D N/D N/D
9.44 8.30 8.86 0.34 0.27 3.77 N/A N/A N/A
a b
Data (average ± SE) represent three repeats. DNA amount indicted by lg/ml of milk. N/D, not detectable.
Fig. 3. PCR amplification of DNA isolated from meats, milks and cheeses. (a) PCR amplification of DNA with bovine and ovine-specific cytb primers. Lanes 1–7, PCR amplification of bovine cytb gene with DNA from fresh beef mince, water buffalo milk, lamb mince, goat blood, pork mince, chicken meat and turkey mince respectively; lanes 8–11, PCR amplification using ovine cytb gene primers with DNA from fresh beef mince, water buffalo milk, lamb mince and goat blood, respectively. (b) PCR amplification of DNA using bovine-specific cytb primers. Lane 1, beef mince DNA; lane 2, B cows’ milk DNA; lane 3, Cheddar cheese DNA; lane 4, B Italian mozzarella cheese DNA; lane 5, Somerset goats cheese DNA. Arrows indicate the molecular size of expected PCR products: unfilled arrow for the 116 bp bovine sequence and filled arrow for a 133 bp sequence.
Here we observed non-specific amplification of goat DNA with ovine cytb primers at a low efficiency (lane 11 in Fig. 3a). Alignment of the ovine primers with the caprine cytb sequence present in the Genbank database (Accession number X56289) shows that there are four base differences at the 3 0 end of the primers (Fig. 1), indicating that they are unlikely to amplify caprine DNA efficiently. However, a different caprine cytb sequence (AB110597), was identified with greater similarity at 3 0 end to the ovine cytb primers (data not shown). Thus, the ovine primers may be able to amplify certain goat varieties, depending upon the cytb DNA sequence. Lo´pez-Calleja et al. (2005) also reported amplification of buffalo DNA of a few breeds with a bovine-specific 12S primer. For the development of a robust quantitative animal products detection system it is necessary to check those primers with DNA from many breeds of cow, sheep and goat. 3.3. PCR amplification of DNA from milk and cheese samples The pair of bovine-specific cytb primers was able to amplify DNA from cows’ milk and Cheddar cheese (lane
2&3 in Fig. 3b). Positive PCR results were also obtained with the Italian mozzarella cheese samples indicating the presence of bovine DNA in those cheeses (lane 4 in Fig. 3b). Positive PCR results were obtained with goats cheeses DNA using ovine cytb primers whilst negative PCR results were obtained using the bovine-specific primers, indicating that the negative reactions were not the results of poor quality DNA or PCR inhibitory contaminants (lane 5 in Fig. 3b and data not shown). This confirms previous findings that there are sufficient somatic cells present in mammalian milk to enable the isolation of DNA of suitable quantity and quality for subsequent PCR amplification (Herman, 2001). 3.4. Duplex PCR Duplex PCR for bovine and caprine DNA detection was developed using the bovine and ovine-specific cytb primers (Fig. 4a). For equal amplification of bovine and ovine cytb sequences, 1 ll each of forward and reverse bovine or ovine-specific primers were optimal for the duplex PCR. No preferential DNA amplification was observed. A further duplex PCR using bovine and chicken-specific cytb
C.-L. Zhang et al. / Food Control 18 (2007) 1149–1158
1155
optimal for equal amplification of beef and chicken DNAs. These parameters were used in the real-time duplex PCR for simultaneous detection of beef and chicken, or of complex foods containing eggs and milk. 3.5. Real-time PCR detection of DNA in meat products
Fig. 4. Duplex PCR amplification of mixed meats DNA with bovine and ovine or chicken-specific cytb primers. (a) Duplex PCR amplification using primers for bovine and ovine cytb genes with fresh beef/lamb mince DNA (15 ng/15 ng). Lane 1, 1 ll ovine primers; lane 2, 0.5 ll bovine primers/1 ll ovine primers; lane 3, 1 ll bovine primers/1 ll ovine primers; lane 4, 1 ll bovine primers. The unfilled and filled arrows indicate the expected 116 bp bovine/133 bp ovine PCR products, respectively. (b) Duplex PCR amplification using primers for bovine and chicken cytb genes with beef/ chicken mince DNA (30 ng/35 ng). Lane 1, 1 ll chicken primers; lane 2, 0.3 ll bovine primers/1 ll chicken primers; lane 3, 0.5 ll bovine primers/ 1 ll chicken primers; lane 4, 1 ll each of bovine/chicken primers. The unfilled and filled arrows indicate the expected 116 bp bovine/133 bp chicken PCR products, respectively. Lane M, 100 bp ladder molecular marker.
primers was optimized for amplification of beef and chicken meats DNA (Fig. 4b). 0.5 ll each of bovine cytb primers and 1 ll each of chicken cytb primers were found
Serial dilutions of fresh beef mince DNA with sterilised MQ water were tested in real-time PCR assays (Fig. 5a; Table 2). The lowest Ct value (18.7) was obtained with 350 ng of beef DNA (Table 2). The lowest quantifiable level of bovine DNA was found to be 35 pg with Ct values of about 32. The Ct values with high DNA concentrations are in agreement with those reported (Dooley et al., 2004) using the same primers. However, the sensitivity of this assay is lower than 0.5 pg reported by Walker et al. (2003) using primers and probe for a high copy genomic short interspersed element. The mammalian-specific cytb probe used here was aligned with various cytb genes (Fig. 1). It was observed that the published sequence (Dooley et al., 2004) was not an exact match for the aligned sequences though a reasonable signal was obtained in this study. A modified universal mammalian-specific cytb probe, TGA GGA CAA ATA TCA TTY TGA GGR GCW ACR GTY A, is being tested for enhanced sensitivity and wide applications, e.g. for the identification of water buffalo and goat DNA.
Fig. 5. Real-time PCR graph obtained with bovine DNA using bovine-specific cytb primers and a mammalian-specific cytb probe. (a) Quantitation graphs of DNA isolated from fresh beef mince. The amount of DNA from 350 ng to 0.035 ng was used and was shown in red, green, blue, yellow and pink colour, respectively (from left to right). Data from one typical experiment were shown here. (b) A generalised standard curve for real-time quantitation of bovine DNA. Ct values of fresh beef mince DNA (350–0.035 ng) in three repeated experiments were used for the construction of standard curve. (For interpretation of the references in color in this figure legend, the reader is referred to the web version of this article.)
1156
C.-L. Zhang et al. / Food Control 18 (2007) 1149–1158
The real-time PCR data were less affected when beef DNA samples were mixed with lamb and pork DNAs. The cross-reactions with other species, i.e. pork, lamb, goat, chicken and turkey were minimal (Tables 2 and 3). Pork DNA (32.5 ng) had a Ct value of 44 which is well below the Ct value of 37 produced by 3.5 pg bovine DNA. A standard curve was constructed by plotting Ct values against the log10[calculated DNA amount] (350 ng to 35 pg) (Fig. 5b). The linear regression equation, Y = 3.34X + 26.69, R2 = 0.924 (significant at P = 0.05) was used to calculate DNA amount from a range of DNA samples isolated from processed meats and mixed foods (Table 2). DNA isolated from a number of processed beef products was tested. Corned beef DNA (500 ng) and steak pie DNA (200 ng) produced Ct values of 17.9 and 19.3, respectively, corresponding to equivalent amounts of fresh meat DNA of 287 and 121 ng. This indicates that despite the extensive degradation of DNA in these products (Fig. 2), amplification occurred at around 60% efficiency compared to fresh meat DNA. On the other hand, the oxtail soup sample (60 ng) had a low Ct value of 35, at the limits of sensitivity of the calibration curve, corresponding to about an equivalent of 0.005 ng fresh meat DNA. It is not yet clear whether this reflects the extensive degradation of DNA in this sample, the presence of PCR contaminants, or the low proportion of bovine DNA in the product. The bovine cytb real-time PCR system has been developed as bovine products have been the subject of examination. This TaqMan system could be updated to detect dual or multiple forensic subjects in a duplex or multiplex realtime PCR by using other compatible dyes (e.g. TET, VIC) to label other species-specific probes, i.e. poultry, water buffalo or goat-specific DNA probes. Duplex real-time PCR using bovine and chicken-specific cytb primer pairs and FAM-labelled mammalian and TET-labelled poultryspecific cytb probes was carried out based on the optimized duplex PCR described in Section 3.4. A DNA mixture (fresh beef DNA 30 ng, and chicken DNA 35 ng) produced Ct values of 22.22 ± 0.66 and 25.15 ± 1.90 which are comparable to data obtained using bovine and chicken single species assays respectively developed in this laboratory. For development of further duplex or multiplex real-time PCR assays, primers and probes could be designed based on those reported: e.g. buffalo-specific mitochondrial cytb genes (Bottero et al., 2002; Rea et al., 2001), ovine and caprine mitochondrial 12S and 16S rRNA genes (Bottero et al., 2003; Mafra et al., 2004). It has been demonstrated that the bovine cytb realtime PCR system is useful for the authentication of fresh and processed meats. It may be valuable in the quality assurance of meat products. The amount of DNA and the size of DNA molecule in a food product are related to the proportion of the meat and the method of preparation (Frezza et al., 2003; Matsunaga et al., 1999; Mayer et al., 1995). The amount and quality of DNA extracted from meat may also be affected by the species, breed, tissue, feeding conditions and processing proce-
dures (Li et al., 2006; Zhao et al., 2005; Zhou et al., 2001). One advantage of real-time PCR is that the amplicon is typically 100 bp, whereas conventional PCR amplicons are several-fold longer. The application of real-time PCR to the authentication of processed meat products in which the DNA may be highly degraded is therefore promising. 3.6. Real-time PCR detection of DNA in milk and cheese products The quantitative detection of bovine DNA in milks and cheeses was possible using the real-time PCR assay developed for beef meat. The yield of DNA from milk was low, but the DNA appeared to be high molecular weight. The amplification pattern of DNA isolated from cows’ milk monitored by real-time PCR was comparable to those obtained with beef meat DNA (data not shown). However, the target DNA amount derived from Ct values for DNA samples of cows’ milk were only 8–9% of the amount of milk DNA in the assay (Table 3). The quantity and quality of DNA in cheese is thought to be low because the processing of cheeses involves high temperature, microorganism and chemical treatments. It is known that there are large numbers of microorganisms involved in the cheese fermentation and ripening process (for example, Rademaker, Hoolwerf, Wagendorp, & te Giffel, 2006). DNA from Cheddar cheese and Italian mozzarella cheese samples gave positive reactions, though with Ct values equivalent to considerably lower amounts of fresh meat DNA (Table 3). A positive identification of bovine DNA in the cheese quiche (containing cows’ milk, Cheddar cheese and hens eggs) was also made (Table 3). Using the duplex real-time PCR assay, a further concentrated cheese quiche DNA sample produced Ct values of 28.11 ± 1.11 (bovine results) and 34.87 ± 2.38 (chicken results). No reaction was recorded with goats’ milk and goats’ cheese DNA, using the bovine cytb primers in realtime PCR. The amount of target DNA estimated by Ct value in the milks and cheeses used in this study was only a small fraction of the equivalent amount of fresh meat DNA. This could indicate that the DNA was of low quality, either due to degradation caused by food processing (in the case of cheese) or the presence of PCR inhibitors which were not sufficiently removed in the purification step. Another possible reason is that the mtDNA amount may be lower in somatic cells in milks than cells in meats, and that there is a considerable amount of microbial DNA in cheese samples. Testing those possibilities and optimizing procedures for DNA extraction and purification towards a higher accuracy for DNA quantification in milks and cheeses are in progress in this laboratory. Rea et al. (2001) reported that the amount of DNA recoverable from milk and cheese was directly related to the somatic cell content of the raw milk, and also to the strength of the technique used to process the product, as
C.-L. Zhang et al. / Food Control 18 (2007) 1149–1158
this can influence the yield, integrity and extractability of the DNA. Lo´pez-Calleja et al. (2005) analysed pure (100%) bovine milk including raw, pasteurised and sterilised samples and found heat-treatment halved the number of somatic cells though PCR amplification was less affected. It was thought that the detection of low amounts of cows’ milk adulteration in processed cheese will be difficult. We did not analyse samples with known percentages of cow component in mixed goat or water buffalo milks or cheeses as final DNA amount put into each PCR reaction determines the real-time PCR result. Our results indicate that it is possible to detect bovine sequences in meat mixtures by real-time PCR. They also show that bovine DNA can be detected in cows’ milk and cheeses. The negative results obtained with goats’ milk and cheeses using bovine-specific primers, compared to the positive results obtained in conventional PCR with ovinespecific primers suggest that it will be possible to use the real-time procedure to detect adulteration of goat milk products with cows’ milk. It has been reported that more than half of Mozzarella di buffalo (POD) cheeses were contaminated or adulterated with cows’ milk (Bottero et al., 2002; Di Pinto et al., 2004). The procedures described here were able to isolate DNA from mozzarella style cheese and obtain a positive reaction with the bovine DNA real-time assay. The same primers were negative against water buffalo milk DNA. This realtime assay will be able to detect the adulteration of water buffalo mozzarella with cows’ milk. 3.7. Further consideration of real-time PCR system for bovine DNA detection The uptake of real-time PCR system by the food industry depends on its technical advantages and relatively low cost. The most expensive chemicals in the real-time PCR assay are TaqMan probe and the universal PCR master mix. It was found that general PCR reagents were compatible with TaqMan probes (data not shown) but the sensitivity was reduced. This may indicate the importance of UNG in the universal PCR master mix (Longo, Berninger, & Hartley, 1990). Real-time PCR system based on SYBR Green chemistry can be a cheap alternative to TaqMan system. SYBR Green directly binds to double stranded DNA and facilitated detection of PCR products. However, this chemical is thought not to be as sensitive as TaqMan chemistry and does not permit the performance of multiple real-time PCR assays. 4. Conclusions The TaqMan bovine cytb real-time PCR system for the identification of meats, milks and cheeses is sensitive, quick and safe. Its capability to quantify low levels of bovine DNA (35 pg) will meet the standard required by many authentication measurements. A duplex real-time PCR sys-
1157
tem based on bovine and chicken-specific cytb primers and probes has also been used to measure DNA amounts in meats, milk and egg products. Acknowledgements We are grateful to Dr. Isabel Gonza´lez, Universidad Complutense de Madrid, Spain for providing DNA samples extracted from authentic water buffalo milk and goat blood samples, to Drs. J. Hall, E. Taylor and R. Allsopp for their help. This work was supported by the UK Higher Education Innovation Fund (HEIF II). References Amigo, L., Ramos, M., Calhau, L., & Barbosa, M. (1992). Comparison of electrophoresis, isoelectric focusing and immunodiffusion in determinations of cow’s and goat’s milk in Serra da Estrella cheeses. Lait, 72, 95–101. Bania, J., Ugorski, M., Polanowsk, A., & Adamczyk, E. (2001). Application of polymerase chain reaction for detection of goats’ milk adulteration by milk of cow. Journal of Dairy Research, 68, 333–336. Bottero, M. T., Civera, T., Anastasio, A., Turi, R. M., & Rosati, S. (2002). Identification of cow’s milk in ‘‘buffalo’’ cheese by duplex polymerase chain reaction. Journal of Food Protection, 65, 362–366. Bottero, M. T., Civera, T., Nucera, D., Rosati, S., Sacchi, P., & Turi, R. M. (2003). A multiplex polymerase chain reaction for the identification of cow’s, goat’s and sheep’s milk in dairy products. International Dairy Journal, 13, 277–282. Branciari, R., Nijman, I. J., Plas, M. E., Di Antonio, E., & Lenstra, J. A. (2000). Species origin of milk in Italian mozzarella and Greek feta cheese. Journal of Food Protection, 63, 408–411. Cerquaglia, O., & Avellini, P. (2004). A rapid c-casein isoelectrofocusing method for detecting and quantifying bovine milk used in cheese making: Application to sheep cheese. Italian Journal of Food Science, 16, 447–455. Chen, F. C., & Hsieh, Y. H. (2000). Detection of pork in heat-processed meat products by monoclonal antibody-based ELISA. Journal of AOAC International, 83, 79–85. Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T. J., Higgins, D. G., et al. (2003). Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Research, 31, 3497–3500. Di Pinto, A., Conversano, M. C., Forte, V. T., Novello, L., & Tantillo, G. M. (2004). Detection of cow milk in buffalo ‘‘mozzarella’’ by polymerase chain reaction (PCR) assay. Journal of Food Quality, 27, 428–435. Dooley, J. J., Paine, K. E., Garrett, S. D., & Brown, H. M. (2004). Detection of meat species using TaqMan real-time PCR assays. Meat Science, 68, 431–438. Espinoza, E. O., Kirms, M. A., & Filipek, M. S. (1996). Identification and quantitation of source from hemoglobin of blood and blood mixtures by high performance liquid chromatography. Journal of Forensic Science, 41, 804–811. Feligini, M., Bonizzi, I., Curik, V. C., Parma, P., Greppi, G. F., & Enne, G. (2005). Detection of adulteration in Italian mozzarella cheese using mitochondrial DNA templates as biomarkers. Food Technology and Biotechnology, 43, 91–95. Frezza, D., Favaro, M., Vaccari, G., Von-Holst, C., Giambra, V., Anklam, E., et al. (2003). A competitive polymerase chain reactionbased approach for the identification and semiquantification of mitochondrial DNA in differently heat-treated bovine meat and bone meal. Journal of Food Protection, 66, 103–109. Heid, C. A., Stevens, J., Livak, K. J., & Williams, P. M. (1996). Real time quantitative PCR. Genome Research, 6, 986–994.
1158
C.-L. Zhang et al. / Food Control 18 (2007) 1149–1158
Herman, L. (2001). Determination of the animal origin of raw food by species-specific PCR. Journal of Dairy Research, 68, 429–436. Hird, H., Goodier, R., Schneede, K., Boltz, C., Chisholm, J., Lloyd, J., et al. (2004). Truncation of oligonucleotide primers confers specificity on real time-PCR assays for food authentication. Food Additives and Contaminants, 21, 1035–1040. Holland, P. M., Abramson, R. D., Watson, R., & Gelfand, D. H. (1991). Detection of specific polymerase chain reaction product by utilization the 5 0 to 3 0 exonuclease activity of Thermus aquaticus. Proceedings of the National Academy of Sciences of United States of America, 88, 7276–7280. Hurley, I. P., Ireland, H. E., Coleman, R. C., & Williams, J. H. H. (2004). Application of immunological methods for the detection of species adulteration in dairy products. International Journal of Food Science and Technology, 39, 873–878. Lanzilao, I., Burgalassi, F., Fancelli, S., Settimelli, M., & Fani, R. (2005). Polymerase chain reaction-restriction fragment length polymorphism analysis of mitochondrial cytb gene from species of dairy interest. Journal of AOAC International, 88, 128–135. Laube, I., Spiegelberg, A., Butschke, A., Zagon, J., Schauzu, M., Kroh, L., et al. (2003). Methods for the detection of beef and pork in foods using real-time polymerase chain reaction. International Journal of Food Science and Technology, 38, 111–118. Levieux, D., & Venien, A. (1994). Rapid, sensitive two-site ELISA for detection of cows’ milk in goats’ or ewes’ milk using monoclonal antibodies. Journal of Dairy Research, 61, 91–99. Li, C. B., Chen, Y. J., Xu, X. L., Huang, M., Hu, T. J., & Zhou, G. H. (2006). Effects of low-voltage electrical stimulation and rapid chilling on meat quality characteristics of Chinese Yellow crossbred bulls. Meat Science, 72, 9–17. Longo, M. C., Berninger, M. S., & Hartley, J. L. (1990). Use of uracil DNA glycosylase to control carry-over contamination in polymerase chain reactions. Gene, 93, 125–128. Lo´pez-Andreo, M., Lugo, L., Garrido-Pertierra, A., Prieto, M. I., & Puyet, A. (2005). Identification and quantitation of species in complex DNA mixtures by real-time polymerase chain reaction. Analytical Biochemistry, 339, 73–82. Lo´pez-Calleja, I., Gonza´lez, I., Fajardo, V., Rodrı´guez, M. A., Herna´ndez, P. E., Garcı´a, T., et al. (2004). Rapid detection of cow’s milk in sheep’s and goat’s milk by a species-specific PCR technique. Journal of Dairy Science, 87, 2839–2845. Lo´pez-Calleja, I., Gonza´lez Alonso, I., Fajardo, V., Rodrı´guez, M. A., Herna´ndez, P. E., Garcı´a, T., et al. (2005). PCR detection of cows’ milk in water buffalo milk and mozzarella cheese. International Dairy Journal, 15, 1122–1129. Lumley, I. D. (1996). Authenticity of meat ant meat products. In P. R. Ashurst & M. J. Dennis (Eds.), Food authentication (pp. 108–139). Blackie Academic & Professional. Mafra, I., Ferreira, I. M. P. L. V. O., Faria, M. A., & Oliveira, B. P. P. (2004). A novel approach to the quantification of bovine milk in ovine cheeses using a duplex polymerase chain reaction method. Journal of Agricultural and Food Chemistry, 52, 4943–4947. Matsunaga, T., Chikuni, K., Tanabe, R., Muroya, S., Shibata, K., Yamada, J., et al. (1999). A quick and simple method for the identification of meat species and meat products by PCR assay. Meat Science, 51, 143–148. Maudet, C., & Taberlet, P. (2001). Detection of cows’ milk in goats’ cheeses inferred from mitochondrial DNA polymorphism. Journal of Dairy Research, 68, 229–235. Maudet, C., & Taberlet, P. (2002). Holstein’s milk detection in cheeses inferred from melanocortin receptor 1 (MC1R) gene polymorphism. Journal of Dairy Science, 85, 707–715.
Mayer, H. K. (2005). Milk species identification in cheese varieties using electrophoretic, chromatographic and PCR techniques. International Dairy Journal, 15, 595–604. Mayer, R., Hofflein, C., Luthy, J., & Candrian, U. (1995). Polymerase chain reaction-restriction fragment length polymorphism analysis: A simple method for species identification in food. Journal of AOAC International, 78, 1542–1551. Mimmo, P., & Pagani, S. (1998). Development of an ELISA for the detection of caprine as1-casein in milk. Milchwissenschaft, 53, 363–367. Murray, M. G., & Thompson, W. F. (1980). Rapid isolation of high molecular weight plant DNA. Nucleic Acids Research, 8, 4321–4325. Peter, C., Bru¨nen-Nieweler, C., Cammann, K., & Bo¨rchers, T. (2004). Differentiation of animal species in food by oligonucleotide microarray hybridization. European Food Research and Technology, 219, 286–293. Poms, R. E., Glossl, J., & Foissy, H. (2001). Increased sensitivity for detection of specific target DNA in milk by concentration in milk fat. European Food Research and Technology, 213, 361–365. Rademaker, J. L. W., Hoolwerf, J. D., Wagendorp, A. A., & te Giffel, M. C. (2006). Assessment of microbial population dynamics during yoghurt and hard cheese fermentation and ripening by DNA population fingerprinting. International Dairy Journal, 16, 457–466. Rea, S., Chikuni, K., Branciari, R., Sukasi Sangamayya, R., Ranucci, D., & Avellini, P. (2001). Use of duplex polymerase chain reaction (duplex PCR) technique to identify bovine and water buffalo milk used in making mozzarella cheese. Journal of Dairy Research, 68, 689–698. Saeed, T., Ali, S. G., Abdul Rahman, H. A., & Sawaya, W. N. (1989). Detection of pork lard as adulterants in processed meat: Liquid chromatographic analysis of derivatized triglycerides. Journal of the Association of Official Analytical Chemists, 72, 921–925. Skarpeid, H. J., Kvaal, K., & Hildrum, K. I. (1998). Identification of animal species in ground meat mixtures by multivariate analysis of isoelectric focusing protein profiles. Electrophoresis, 19, 3103–3109. Stefos, G., Argyrokastritis, A., Bizelis, I., Moatsou, G., Anifantakis, E., & Rogdakis, E. (2004). Detection of bovine mitochondrial DNA specific sequences in Feta cheese and ovine yogurt by PCR-RFLP. Milchwissenschaft, 59, 509–511. Sun, Y.-L., & Lin, C.-S. (2003). Establishment and application of a fluorescent polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method for identifying porcine, caprine, and bovine meats. Journal of Agricultural and Food Chemistry, 51, 1771–1776. Walker, J. A., Hughes, D. A., Anders, B. A., Shewale, J., Sinha, S. K., & Batzer, M. A. (2003). Quantitative intra-short interspersed element PCR for species-specific DNA identification. Analytical Biochemistry, 316, 259–269. Woolfe, M., & Primrose, S. (2004). Food forensics: Using DNA technology to combat misdescription and fraud. Trends in Biotechnology, 22, 222–226. Zhang, C. L., Chen, D. F., McCormac, A. C., Scott, N. W., Elliott, M. C., & Slater, A. (2001). Use of the GFP reporter as a vital marker for Agrobacterium-mediated transformation of sugar beet (Beta vulgaris L.). Molecular Biotechnology, 17, 109–117. Zhang, G., Zheng, M., Zhou, Z., Ouyang, H., & Li, Q. (1999). Establishment and application of a polymerase chain reaction for the identification of beef. Meat Science, 51, 233–236. Zhao, G. M., Zhou, G. H., Wang, Y. L., Xu, X. L., Huan, Y. J., & Wu, J. Q. (2005). Time-related changes in cathepsin B and L activities during processing of Jinhua ham as a function of pH, salt and temperature. Meat Science, 70, 381–388. Zhou, G. H., Liu, L., Xiu, X. L., Jian, H. M., Wang, L. Z., Sun, B. Z., et al. (2001). Productivity and carcass characteristics of pure and crossbred Chinese Yellow Cattle. Meat Science, 58, 359–362.