CTAB DNA Extraction Method Optimization for Quantification of Lard Content in Cheese and Butter Using Real-Time PCR

ABSTRACT Porcine adulteration in dairy products for economic purposes has been tackled mostly by protein- and lipid-based porcine detection methods. In this study, a DNA-based approach was used to determine the presence of porcine DNA in dairy products, especially butter and cheese. DNA was extracted by optimizing the conventional CTAB extraction method followed by an assessment of DNA fragmentation by real-time PCR assays targeting different sizes of amplicon. Butter and cheese DNA samples were analyzed using both endogenous and porcine-specific real-time PCR assays and the lard content was estimated based on the linear regression analysis of a series of different percentages of lard-adulterated butter and cheese DNA samples. Even though the optimization steps did not improve the DNA yield, the successful amplification of the 200 bp amplicon depicted high DNA quality. Approximately 1% lard was able to be quantified exhibiting high potential of the assay in quantifying porcine content in lard-adulterated highly processed food products.


Introduction
As milk is considered one of the expensive raw materials that can be modified to produce numerous dairy products, some manufacturers have attempted to substitute it with other cheaper ingredients.There have been a lot of cases pertaining to milk adulteration, for instance, milk substitution with water resulting in diluted milk with less nutritional value as well as the addition of starch, melamine, sugar, urea, detergents, hydrogen peroxide, food colors, chlorine, antibiotics, milk powder, non-milk proteins and fat, low-value milk, preservatives, whey, pesticides, and neutralizers causing hazardous health effects (Silva and Rocha 2020).Other dairy products have also been subjected to adulteration where cow milk was added in the production of commercial PDO buffalo Mozzarella cheese (Cutarelli et al. 2021) and milk fat was partially substituted with cheaper food materials such as hydrogenated oil, potato puree, and banana pulp; pure hydrogenated oil (Gemechu et al. 2021), lard (Nurrulhidayah et al. 2015) and palm oil (Platov, Metlenkin, and Rubtzov 2019) in butter production.Dairy product adulteration issues are not only against consumer rights but also deceive some religious beliefs like Muslims and Jews by mixing sources from prohibited animals (in this case pigs) in food products.Food manufacturers have been using lard as an adulterant in their products as it is less expensive and readily available and the issue has also been discovered in dairy products (Rohman and Windarsih 2020).Furthermore, the high-quality milk fat in butter which is considered the most expensive fat has been substituted with cheaper animal fat like lard or vegetable oil without correct labeling for economic purpose (Platov, Metlenkin, and Rubtzov 2019).Thus, it is crucial to be able to distinguish milk fat from fat originated from either animal or plant species to ensure the authenticity of the dairy products available in the market.
Adulteration of milk between different animal species often requires particular analyses as milk from different animal species cannot be distinguished based on visual inspection, for example, textures, and appearance.Milk protein has been frequently used to identify animal species through the common enzyme-linked immunosorbent assay (ELISA) method and the rapid lateral flow immunoassay for detection of animal species in dairy products (Mafra, Honrado, and Amaral 2022).Analytical methods like Fourier transform infrared (FTIR) (Nurrulhidayah et al. 2015;Platov, Metlenkin, and Rubtzov 2019) and Raman spectroscopy (Taylan et al. 2020) have also been implemented in differentiating butter from other animal fats and vegetable oils, respectively.Even though both methods offer high specificity and simplicity, there are some limitations.Detection of the protein component in dairy products is low or almost impossible when the muscle proteins are denatured or changed during processing.On the other hand, lipid-based methods are less reliable as both types and amounts of fats can be extensively modified during the cooking process (Ali et al. 2011).In order to counter the limitations, the DNA-based method has been extensively used particularly polymerase chain reaction (PCR).In comparison with protein molecule, DNA molecule exhibits more stability which makes them able to survive high temperature and pressure during food processing as well as provides more sensitive detection (Pirondini et al. 2010).
Good-quality DNA is required for successful DNA-based method detection where a range of methods have been developed to extract good-quality DNA depending on the type of sample matrices (Pirondini et al. 2010;Sajali et al. 2018).The most common conventional CTAB extraction method has been frequently used (He et al. 2007;Pirondini et al. 2010) but never been optimized to maximize the yield of DNA extracted from butter and cheese samples.Increased DNA yield could possibly improve the chance of adulteration detection in such samples.In addition, there are not many studies reported on detection and quantification of adulterants in butter and cheese samples.Latest study has employed real-time PCR with EvaGreen intercalating dye to quantitate cattle DNA in buffalo cheese samples at as low as 0.1% (Giglioti et al. 2022) and PCR-RFLP to detect cow's butter in buffalo's butter at a detection limit of 5% (Abdelfatah, El-Araby, and Mohamed 2015).There has been no report on quantification of adulterants in butter sample using DNA-based methods.
This study attempted to increase DNA yield from butter and cheese samples using the conventional CTAB DNA extraction method by optimizing the amount of starting material, the inclusion of mechanical lysis, and incubation temperature upon DNA precipitation.The quality of the DNA extracted from the samples was assessed using EvaGreen real-time PCR assays targeting different sizes of bovine-specific target DNA sequences.Using the optimized CTAB method for butter and cheese, DNA was extracted from laboratoryprepared lard-adulterated as well as commercial butter and cheese samples which was subjected to both endogenous and porcine-specific real-time PCR assays for detection and quantification of lard.

Raw meat and lard
The fresh, raw meat samples of different animal species (pig, cattle, chicken, goat, buffalo, and deer) and lard were purchased from local wet markets in Sri Kembangan, Selangor, Malaysia.The identity of animal species was confirmed by the veterinary experts from the Department of Animal Sciences at Universiti Putra Malaysia (UPM).Samples were stored frozen at −20°C until use to avoid enzymatic degradation of DNA.

Laboratory-prepared samples
Rendered lard.Rendered lard was prepared by rendering the adipose tissues from various parts of slaughtered pigs based on the method reported previously by Marikkar et al. (2001).Initially, the lard must be cleaned properly and all blood and meat were removed to produce pure, rendered lard.Next, it was baked in an oven for 2 h at 90-100°C.The melted lard was then filtered through a muslin cloth and added with anhydrous sodium sulfate to remove residual moisture.Finally, the extract was then filtered through Whatman® No. 2 filter paper and stored at 4°C.
Butter and lard-adulterated butter.Fresh cow's milk was purchased from the Sales Centre, Universiti Putra Malaysia (UPM).By referring to the method described by Nurrulhidayah et al. (2015), the cream was first separated from fresh whole milk using a disc bowl centrifuge (Armfield, Ringwood, and England) and stirred using a mixer on ice.During this step, solid butter and buttermilk were obtained.Next, the butter was rinsed with ice water to separate it from buttermilk.The extracted butter was stored at 4°C until further use.Lard-adulterated butter was then prepared by spiking 0.01%, 0.1%, 1% 10%, 20%, 50%, and 100% (w/w) of rendered lard into laboratoryprepared butter and stored at 4°C until further use.
Cheese and lard-adulterated cheese.A cheesemaking kit containing cheese cultures, rennet, cheesecloth, and manual for cheese processing, was purchased from New England Cheesemaking Supply Company, Inc (South Deerfield, MA, USA).The cheesemaking process was carried out in a water bath with a controllable temperature.According to the manual, firstly, 5 L of fresh cow's milk was heated up to 30°C followed by the addition of 100 mL of prepared mother culture and left untouched for 45-60 min to allow the ripening of the milk.Diluted calf rennet was then added into the ripened milk and left untouched for 45 min at 32°C temperature to ensure good formation of curd.The curd was then chopped to small pieces and poured off through a mold, causing the separation of cheese and whey.To ensure complete removal of whey, the cheese was rinsed with plain water and stored at 4°C until further use.Lard-adulterated cheese was then prepared by spiking 0.01%, 0.1%, 1% 10%, 20%, 50%, and 100% (w/w) of rendered lard into laboratory-prepared cheese before being stored at 4°C until further use.

Commercial butter and cheese samples
Commercial butter and cheese produced by 10 different manufacturers were purchased from several local supermarkets in Selangor, Negeri Sembilan, and Kuala Lumpur, Malaysia.Both butter and cheese product samples originated from cattle and did not contain any lard or pig derivative.All samples were stored at 4°C until further use.

Optimization of CTAB method
Total DNA extraction was carried out from 0.2 g of the sample using the CTAB method as reported by He et al. (2007).Initially, 1 mL of CTAB extraction buffer [1.4 M NaCl, 2% CTAB, 100 mM Tris, 20 mM EDTA, pH 8.0] and 1 µL of proteinase K (100 µg/mL) were added into the sample which was then incubated for 1 h at 65°C with shaking.Later, 1 mL chloroform was added into the lysed sample followed by centrifugation for 20 min at 12,000 × g.Then, 2 sample volumes of CTAB precipitation butter (40 mM NaCl and 0.5% CTAB) were added to the supernatant and incubated at room temperature for 1 h.The solution was centrifuged for 20 min at 15,000 × g.The pellet obtained was dissolved in 350 µL of 1.2 M NaCl and added with an equal volume of chloroform.After centrifugation at 15,000 × g for 20 min, the pellet was recovered and added with 350 µL of 2-propanol and 1 µL of glycogen (Sigma Aldrich) and incubated overnight followed by another round of centrifugation to precipitate the DNA.Finally, the DNA pellet was washed with 70% ethanol and eluted in 50 µL of nuclease-free water.All centrifugation steps were carried out at room temperature.

Varying the amount of starting material
Different amounts of starting material ranging from 0.2 to 1 g of sample were used to investigate its effects on DNA yield.However, the ratio of the CTAB extraction buffer to samples was made constant at 5 mL buffer per gram of sample.

Inclusion of mechanical lysis
The sample was subjected to additional mechanical lysis for at least 2-3 min until the sample was completely dissolved in CTAB extraction buffer by using TissueRuptor (Qiagen, Hilden, Germany) before sample incubation at 65°C.

Reduction of incubation temperature upon DNA precipitation
During the DNA precipitation step using 2-propanol and glycogen, the incubation temperature was reduced to 4°C and −20°C to study its effect on DNA yield.

Determination of DNA quantity
The concentration of the DNA extracted from butter and cheese was quantified using the Quant-iT PicoGreen Assay Kit (Molecular Probes, Eugene, OR, USA) according to the manufacturer's protocol.For standard, 1 µg/mL of lambda DNA standard was prepared and serially diluted in 10-fold dilution in 1X TE buffer.100 µL of each serial dilution of the standard was used in the assay.For samples, the DNA samples were diluted in 1X TE buffer at a dilution factor of 20 to a total volume of 100 µL.A volume of 100 µL of DNA standards and samples were then transferred into a Fluotrac microplate (Greiner Bio-One, GmbH, Frickenhausen, Germany) followed by the addition of 100 µL of 1X PicoGreen dye into each standard and sample.After 5 min of incubation, the fluorescence was measured using a microplate reader (Infinite M200; Tecan Group Ltd., Mannedorf, Switzerland) where the samples were excited at 480 nm and the fluorescence intensity was measured at 520 nm.Each DNA standard and sample were measured 3 times and the fluorescence data of the standards was used to construct the standard curve which was used to calculate the DNA concentration of the samples.

Real-time PCR
Endogenous and species-specific EvaGreen real-time PCR assays were used to evaluate DNA intactness and the presence of inhibitors as well as to detect both porcine and bovine species in butter and cheese samples (Table 1).DNA fragmentation was evaluated using three sets of bovine-specific primers that amplify different sizes of target sequences ranging from below 100 bp to around 200 bp.
A linear plot for lard quantification was constructed by assaying 0.01%, 0.1%, 1% 10%, 20%, 50%, and 100% (w/w) lard-adulterated butter and cheese sample with porcine-specific assay.The Cq values were then plotted against log 10 of lard percentage followed linear regression analysis.The linear equation was then used to extrapolate the Cq value obtained for porcine-positive sample.
Species detection in 10 butter and cheese samples produced by different manufacturers was analyzed using two sets of primers including porcine-specific primers amplifying 89 bp amplicon and endogenous primers to amplify all animal DNA present in the sample.Along with the samples, lard, and pork DNA were also analyzed as positive controls of the assay.
Real-time PCR was done in a 20 μL reaction containing 1X SSofast EvaGreen Supermix (Bio-Rad Laboratories, Hercules, CA, USA), 400 nM of each forward and reverse primer, and 5 ng of butter DNA or 1 ng of cheese DNA.The assay was performed on Mastercycler ep-realplex (Eppendorf AG, Hamburg, Germany) according to the following thermal cycling condition; pre-denaturation at 95°C for 5 min followed by 40 cycles of denaturation at 95°C for 15 seconds and annealing at a specific temperature for each primer set (Table 1) for 1 min.Each sample was run in triplicates.

Statistical analysis
DNA concentration data obtained was analyzed using one-way ANOVA (Minitab 16, Minitab Inc., State College, PA, USA).The analysis aimed to assess the impact of variations in starting material quantity, additional mechanical lysis and reduced incubation temperature on DNA precipitation.In addition, linear regression analysis was performed to determine the PCR efficiency of real-time PCR assays based on the formula, E = [10(−1/slope)-1] (Ali et al. 2011) and to obtain linear equation for lard quantification.

Optimization of CTAB method
In general, the application of the standard CTAB extraction method resulted in extracting over 100 times more DNA from the cheese sample compared to the butter sample.This substantial disparity can be primarily attributed to the prevalence of microbial DNA in the cheese.
The challenge of obtaining a low DNA yield from butter samples is a common issue encountered during DNA extraction from extensively processed products, leading to limited DNA availability for subsequent PCR analysis.Consequently, various optimization steps were explored for the CTAB-based DNA extraction method.These included adjusting the quantity of starting material, introducing mechanical lysis, and modifying temperature conditions during DNA precipitation to enhance the overall DNA yield.

The effect of different amount of starting material
Originally, 0.2 g of samples were used in the DNA extraction method as reported by He et al. (2015).In this study, the amount of sample was increased to 1 and 2 g to increase the yield of DNA extracted from butter and cheese samples.However, according to Figure 1, this increment had a negative effect on the DNA yield as maximum values of 37.82 ± 0.37 ng/g and 3127.52 ± 47.6 ng/g were obtained from 0.2 g of butter and cheese respectively.One-way ANOVA analysis of the DNA yields for both samples revealed that they were significantly reduced down to 50% and 70% (p-value ≤0.05) when the initial amount of sample used was increased up to 1 g and 2 g respectively.As highly processed foods like butter and cheese are composed of complex molecules, the increased amount of starting material results in extra components such as food stabilizers, emulsifiers and food coloring that exist in the sample matrix.These components are regarded as contaminants that can possibly bind to DNA molecules causing difficulties in extracting more DNA from the samples.Thus, it is recommended to use smaller scale DNA extraction method as it does not affect the efficiency of PCR detection and saves both cost and time (Demeke et al. 2014).

The effect of inclusion of mechanical lysis
Additional mechanical lysis using TissueRuptor (Qiagen, Hilden, Germany) was carried out before sample incubation for lysis.Theoretically, homogenization of the sample contributes to more efficient sample lysis that will release the DNA trapped in the matrix.Unfortunately, according to one-way ANOVA results, there was no significant difference in the DNA yield extracted from the cheese sample upon additional mechanical lysis (p-value >0.05).In contrast, the yield of DNA extracted from the butter sample was significantly decreased with additional mechanical lysis (Fig. 1).Reduced DNA yields might be caused by incomplete disruption of cells which leads to failure in releasing nucleic acids from the sample (TissueRuptor Handbook).Different matrix and composition of butter and cheese also gave different effects of mechanical lysis on the DNA yield where different period of lysis was needed to lyse the sample completely using this instrument.However, prolonged use of TissueRuptor is not recommended to avoid DNA degradation which will affect both quantity and quality of the DNA (Mohamad et al. 2016).

The effect of reduced incubation temperature upon DNA precipitation
On the other hand, the effect of sample incubation temperature during DNA precipitation step was also evaluated upon individual CTAB extraction of the DNA from butter and cheese samples.One-way ANOVA analysis on the DNA yields obtained when the samples were incubated overnight at 4°C and −20°C against the original temperature gave p-value ≤0.05 indicating poor DNA precipitation at a chilled temperature.This result complements the findings by Li et al. (2020) where precipitation of long DNA sequences is not improved at lower temperature.Ethanol viscosity increases with decreasing temperature inhibiting the movement of DNA aggregates, especially the small ones.Moreover, Zumbo (2012) added that more salts will begin to co-precipitate with DNA molecules as solubility decreases at chilled temperatures.Therefore, incubation of the solution before DNA precipitation at a chilled temperature is counter-productive and not recommended.

Real-time PCR analysis
Prior to detection in butter and cheese samples, all sets of primers (Table 1) were tested for their specificity and sensitivity in addition to the BLAST analysis where it gave 100% confirmation of the respective species (data not shown).The specificity test was performed against six animal species including pig, cattle, chicken, buffalo, sheep, and deer where only porcine and bovine DNA was detected by the porcine-and bovine-specific assays, respectively, while all animal species can be detected by the universal assay (supplementary data).The function of universal assay used in this research was to confirm the amplifying ability of the DNA samples tested and prove the absence of inhibitor presents in the samples that can prohibit the real-time PCR.On the other hand, the sensitivity test results have shown that the LOD of all realtime PCR assays used were 0.1 pg DNA.Moreover, the sensitivity of the porcine-specific assay was also tested in pork/beef meat binary mixtures where the LOD obtained was 0.1% (w/w) pork (data not shown).Very low LOD obtained ensures the capability of real-time PCR assay in detecting species in complex meat mixtures which solves the problem involving the production of multiple DNA fragments that cause complex subsequent result analysis when using the PCR-RFLP technique (Abdelfatah, El-Araby, and Mohamed 2015).
It is important to determine the efficiency of a particular PCR assays for species detection.A good real-time PCR efficiency generally varies between 90% and 110% with high linearity (R 2 = 0.99) (Ali et al. 2011).From the linear regression analysis, the PCR efficiencies of three bovine-specific assays ranges from 93% to 99% (supplementary data) while for endogenous and porcinespecific assays, the efficiencies are 92.7% and 103.2% (Fig. 2) which were fairly high.

Determination of the quality of DNA extracted from butter and cheese
As shown in Table 2, all sizes of target sequences tested were amplifiable depicting the availability of as large as around 200 bp fragment in butter and cheese DNA samples.Although larger DNA fragments were detected later than the smaller ones, they could still survive the harsh conditions during the manufacturing process of butter and cheese.Hence, detection in butter and cheese samples carried out with primers targeting as large as 200 bp fragment and real-time PCR assays are usually designed to target DNA sequence of 150 bp or smaller (Kang 2019).

Species detection in laboratory-prepared lard-adulterated samples
Upon analysis using DNA samples from both lard-adulterated butter and cheese samples, successful amplification by the endogenous PCR assays indicated good-quality DNA had been extracted from the samples.The Cq values obtained ranged from 19.38 ± 0.02 to 26.49 ± 0.77 for lard-adulterated butter samples and from 19.30 ± 0.16 to 22.52 ± 0.15 for lard-adulterated cheese samples.In addition, no significant difference in the Cq values has been observed from the ANOVA results at a 99% confidence level which eliminated the need for Cq value normalization of the porcine-specific assay.
On the other hand, porcine DNA was successfully amplified by the porcinespecific assay in the positive control containing 100% (w/w) lard at an average Cq value of 26.65 ± 0.88 showing that all primers, PCR reagents, and the realtime PCR instrument worked properly.The LOQ of the assay obtained was 0.1% (w/w) lard for both butter-and cheese-adulterated samples with Cq values ranging from 30.59 ± 0.24 to 37.54 ± 0.07 and 29.25 ± 0.16 to 34.24 ± 0.24 at 50 and 0.1% (w/w) lard respectively.Unfortunately, based on the linear regression analysis, the efficiency of the porcine-specific assay has been greatly increased up to 105.8% and 176.8% when using lard-adulterated butter and cheese samples, respectively (Fig. 2).The efficiency of the assay was still acceptable for butter samples but it was totally out of range for cheese samples which might be caused by co-amplification of nonspecific products such as primer-dimers (Bustin et al. 2009).In comparison, Ali et al. (2011) obtained a much lower LOQ of 0.01% (w/ w) with good efficiency of 93.8% from a porcine-specific assay amplifying 109 bp target sequence of mitochondrial cyt b gene sequence using pork-adulterated beef burger.These results were clearly better but burger formulation comprises meat-based product and does not involve highly processed product which is subjected to harsh processing conditions and consists of more complex food matrix such as lard, butter, and cheese.A high content of PCR-inhibitors in the DNA samples might also result in lower PCR efficiency as lard is produced from the cream of milk which contains a high amount of lipids (Pirondini et al. 2010).The amount of inhibitor also tended to increase when significantly more DNA was extracted especially from the cheese sample which resulted in delayed Cq value, affecting the PCR efficiency ultimately.The delayed Cq value for the butter sample might also be due to the lower amount of mitochondrial DNA in somatic cells of animal fat in butter compared to the muscle cells in meat (Zhang et al. 2007).Other than that, inconsistent efficiency of the DNA extraction process from highly processed product samples could also contribute to reduced real-time PCR assay efficiency.

Species detection in commercial butter and cheese samples
Successful amplification by the endogenous PCR assay indicated goodquality DNA has been extracted from all the samples tested with Cq values between 28 to 32 and 23 to 26 for butter and cheese samples, respectively (Table 3).Earlier detection in cheese samples might be due to the amplification of bacterial DNA that is extracted together with the animal DNA (Pirondini et al. 2010).This also explained the significantly higher yield of  DNA extracted from such type of sample as compared to that of butter samples.
On the other hand, it was unanticipated to observe positive detection of porcine DNA in one of the cheese samples tested (product no.7) with an average Cq value of 29.45 ± 0.11.By interpolating the Cq value using the linear regression analysis of lard adulterated cheese (Fig. 2b), this result was quite significant as the resulting sample could be estimated to contain around 1% lard roughly.This was a surprising finding as the product was labeled to originate from cattle and contained no porcine derivative.However, further evaluation of all of the products manufactured in the same as well as different batches is required to confirm the adulteration of lard or other porcine derivatives in the product.

Conclusion
The original CTAB DNA extraction method applied in this study has been proven to produce maximum DNA yield from butter and cheese samples as none of the optimization steps taken led to the improvement of DNA yield from such samples.Using this method, around 38 ng and 3.8 mg of DNA can be extracted from 1 g of butter and cheese samples, respectively.The quality of the extracted DNA samples was also good indicated by successful amplification of 200 bp fragment.On the other hand, the porcine-specific real-time PCR assay used in this study is highly efficient in amplifying porcine DNA from lard-adulterated butter and cheese samples where it is able to quantify down to 0.1% (w/w).Thus, it has a high suitability as a screening tool for porcine adulteration in dairy products to detect the presence of lard or pork DNA and possibly to estimate the lard or pork content in highly processed food products.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Figure 1 .
Figure 1.Comparison of DNA yields obtained from different optimization steps for a) butter and b) cheese samples.

Figure 2 .
Figure2.Linear regression of (a) endogenous and porcine-specific assay using meat DNA and (b) porcine-specific assay using lard-adulterated butter and cheese DNA.

Table 1 .
List of primers used in this study.

Table 2 .
Cq values obtained from laboratory-prepared butter and cheese DNA samples using bovine-specific real-time PCR assays targeting 67, 116 and 217 bp amplicons.

Table 3 .
Cq values obtained from porcine-specific and endogenous real-time PCR assays of commercial butter and cheese samples produced by different manufacturers.