Screening and Assessment of Low-Molecular-Weight Biomarkers of

May 10, 2018 - Adulteration of Mozzarella di Bufala Campana with cow milk is a common fraud ... lutein, and β-cryptoxanthin as LMWBs of cow milk, whi...
1 downloads 0 Views 2MB Size
Download PDF
Article Cite This: J. Agric. Food Chem. 2018, 66, 5410−5417

pubs.acs.org/JAFC

Screening and Assessment of Low-Molecular-Weight Biomarkers of Milk from Cow and Water Buffalo: An Alternative Approach for the Rapid Identification of Adulterated Water Buffalo Mozzarellas Chiara Dal Bosco, Stefania Panero, Maria Assunta Navarra, Pierpaolo Tomai, Roberta Curini, and Alessandra Gentili* Department of Chemistry, Faculty of Mathematical, Physical and Natural Sciences, University of Rome “La Sapienza”, Piazzale Aldo Moro 5, 00185 Rome, Italy S Supporting Information *

ABSTRACT: Adulteration of Mozzarella di Bufala Campana with cow milk is a common fraud because of the high price and limited seasonal availability of water buffalo milk. To identify such adulteration, this work proposes a novel approach based on the use of species-specific, low-molecular-weight biomarkers (LMWBs). Liquid chromatography−tandem mass spectrometry screening analyses identified β-carotene, lutein, and β-cryptoxanthin as LMWBs of cow milk, while ergocalciferol was found only in water buffalo milk. Adulterated mozzarellas were prepared in the laboratory and analyzed for the four biomarkers. Combined quantification of β-carotene and ergocalciferol enabled the detection of cow milk with a sensitivity threshold of 5% (w/w). The method was further tested by analyzing a certificated water buffalo mozzarella and several commercial products. This approach is alternative to conventional proteomic and genomic methods and is advantageous for routine operations as a result of its simplicity, speed, and low cost. KEYWORDS: food analysis, liquid chromatography−mass spectrometry, mozzarella, milk, biomarkers, adulteration, vitamins, carotenoids

1. INTRODUCTION Mozzarella di Bufala Campana is a traditional Italian cheese with protected designation of origin (PDO). It is the third Italian PDO cheese in terms of turnover after Grana Padano and Parmigiano Reggiano and has a supply chain that employs 15 000 people in the center−south of Italy. This prestigious Italian food is manufactured in compliance with a rigorous production protocol based on exclusive use of water buffalo milk in dedicated plants of Campania, Latium, and Molise (Article 4 of the Italian Law Decrees number 91 of 24 June 2014 and number 116 of 11 August 2014, Commission Regulation EC 273/2008). Adulterations with cow milk are widespread because of its low cost and availability, especially in the summer time when the demand for water buffalo mozzarella increases. Although there are no sanitary−health implications, such fraudulent activities damage consumers and leave serious economic fallouts in the Italian dairy sector. Therefore, simple and rapid analytical methods aimed at identifying adulterations of water buffalo mozzarella are a resource for manufacturing companies as well as official institutions responsible for food quality control. To this end, several analytical methods have been developed thus far.1−16 Most of them have been focused on proteomic evaluation of milk from different animal species using electrophoretic,1−4 chromatographic,5−10 mass spectrometric,11−14 and immunological15 techniques. The reference method of the European Community consists of isoelectric focusing of γ-caseins after plasmolysis, and it is used to detect bovine proteins in cheese made from ewe, goat, or water buffalo milk.17 This method can identify percentages of added cow milk as low as 0.5−1% (w/ © 2018 American Chemical Society

w), but it is long, laborious, and suffering from difficulties of interpretation as a result of overlapped species-specific bands; moreover, integration with the complicated technique of immunoblotting is sometimes necessary.3 In the last few years, mass-spectrometry-based strategies have been developed to profile peptides and proteins (caseins and whey proteins) in milk from various mammalian species. In particular, matrixassisted laser desorption/ionization−time of flight mass spectrometry (MALDI−TOFMS) can detect protein pattern differences very effectively and with great sensitivity.12,13 Nevertheless, MALDI does not provide reliable quantitative analyses and involves time-consuming tasks of preparation and interpretation. In recent years, the polymerase chain reaction (PCR) has been used to detect cow DNA in milk and mozzarella from water buffalo with a sensitivity threshold of 0.1−1%.16 In fact, the large amount of somatic cells makes milk a source of DNA and a suitable substrate for PCR amplification; however, the imprecise count of somatic cells in milk used for cheese preparation hampers quantitative determination. Although proteomic and genomic approaches are very sensitive, the quantitative dosage of cow milk in adulterated cheeses is still a serious challenge. Moreover, even if qualitative detection is considered a solved analytical problem, routine analysis needs faster and simpler screening methods. In Received: Revised: Accepted: Published: 5410

March 10, 2018 May 2, 2018 May 10, 2018 May 10, 2018 DOI: 10.1021/acs.jafc.8b01270 J. Agric. Food Chem. 2018, 66, 5410−5417

Article

Journal of Agricultural and Food Chemistry

water, used in the extraction procedure based on the cold saponification, was further purified by passing it through a Milli-Q Plus apparatus (Millipore, Bedford, MA, U.S.A.). Screw-capped polyethylene centrifuge tubes (50 mL) were purchased from Aldrich-Fluka-Sigma Chemical. 2.2. Standard Solutions. Individual stock solutions of the biomarkers and internal standards (ISs) were prepared by dissolving their weighed amounts (Ohaus DV215CD Discovery semi-micro and analytical balance, 81/210 g capacity, 0.01/0.1 mg readability) in solvents containing 0.1% (w/v) BHT: ergocalciferol and ergocalciferol-d3 in ethanol at 1 μg/μL, β-carotene and β-carotene-d6 in chloroform at 1 μg/μL, and lutein, zeaxanthin, and β-cryptoxanthin in chloroform at 0.5 μg/μL. Individual stock solutions of all other fat-soluble micronutrients were prepared in chloroform containing 0.1% (w/v) BHT at the concentration of 0.2 μg/μL. Multi-standard working solutions of the four biomarkers were prepared from their individual solutions by diluting in methanol with 0.1% BHT to obtain concentrations suitable for several experiments. To avoid photodegradation, amber glassware was used for all preparations, which were stored at −18 °C in the dark when unused. Ergocalciferol-d3 was the IS for ergocalciferol, while β-carotene-d6 was the IS for β-carotene, lutein, and β-cryptoxanthin. 2.3. Mozzarella Samples. Cow and water buffalo milk were bought from eight different farms of central Italy (Latium and Campania). The eight samples (four of cow milk and four of water buffalo milk) were analyzed to screen and quantify LMWBs. Afterward, cow milk and water buffalo milk samples were paired randomly and mixed in different proportions [0, 1, 5, 10, 25, 50, 75, 100% (w/w) of cow milk] to prepare four series of lab-made mozzarellas. Their analyses were then performed to check LMWBs reliability and to establish both the method sensitivity and its discriminating power in recognizing adulterated products. Seven cow mozzarellas and eight PDO water buffalo mozzarellas were purchased from supermarkets and retail grocery stores of Latium and Campania. Aliquots from these samples were pooled together within each species (cow or water buffalo) to have representative matrices to be used for the method development and validation (see section 2.6). Milk samples and mozzarellas were purchased in the spring/ summer season of 2017. A certified PDO Mozzarella di Bufala Campana, kindly provided by the Italian Breeder Association, was also analyzed in November 2017. 2.4. Preparation of Lab-Made Mozzarellas. Pure and mixed cow/water buffalo mozzarellas were prepared in the laboratory according to the following procedure: 1 L of milk was acidified with citric acid until pH 5.6 was reached; 5 g of rennet was added to heated milk (35−37 °C), which was then allowed to stand at around 40 °C for about 20 min. Thereafter, the obtained curd was sliced, squeezed to remove excess liquid, and heated in a microwave oven for 1−2 min at 800 W. Finally, mozzarella was obtained by quickly kneading the hot curd. 2.5. Sample Treatment. Extraction of LMWBs from milk samples was performed according to a procedure previously developed for milk.20 This protocol based on overnight cold saponification followed by liquid−liquid extraction (LLE) was modified and optimized to treat mozzarella samples. In all cases, operations were performed in subdued light, using low actinic glass tubes and wrapping plastic tubes with aluminum foil to protect the analytes from UV light. All organic solvents used for extraction are to be intended as containing 0.1% (w/ v) BHT. Briefly, a 6 g aliquot of mozzarella, minced in very small pieces, was transferred to a 50 mL screw-capped polyethylene centrifuge tube and spiked with known amounts of ISs (50 μL of 10 ng/μL solution). After a 15 min period for equilibration at room temperature, 18 mL of ethanol and 6 mL of aqueous KOH solution (50%, w/v) were added and the tube was placed in a water bath at 25 °C overnight under continuous stirring. Following the incubation period (15 h), the digest was diluted with 8.5 mL of Milli-Q water and the analytes were extracted by two 12 mL aliquots of hexane. After each aliquot addition,

comparison to protein biomarkers, the identification of lowmolecular-weight biomarkers (LMWBs) can simplify analytical procedures, save time, and reduce costs. Over the past few years, attempts were made by analyzing lipid composition of milk from different animal species, but the procedures were long and demanding.18 More recently, carotenoids have been individuated as diagnostic molecules of cow milk.19,20 In particular, β-carotene was detected in cow milk at concentrations ranging from 10 to 300 μg/L depending upon the season, type of feeding, breed, stage of lactation, etc.19−21 On the basis of this evidence, Cerquaglia et al. have proposed a liquid chromatography−ultraviolet (LC−UV) method to identify the occurrence of cow milk in water buffalo ricotta through determination of β-carotene.5 In light of what has been discussed above, this work has had the objectives of (i) screening biomarkers of cow milk and water buffalo milk among the fat-soluble micronutrients, (ii) developing a liquid chromatography−tandem mass spectrometry (LC−MS/MS) method to quantify cow milk addition to water buffalo milk used to prepare PDO mozzarellas, and (iii) verifying the reliability of the novel approach. To these ends, several milk samples from cow and water buffalo were submitted to screening tests, performed by acquiring chromatograms in scheduled multiple reaction monitoring (SMRM) mode. SMRM is a potent algorithm that allows for acquisition of many ion currents simultaneously without compromising data quality. In this way, the targeted screening of 26 fat-soluble micronutrients could be carried out with great sensitivity. After four biomarkers were identified, SMRM was not necessary anymore and a liquid chromatography−multiple reaction monitoring (LC−MRM) method, exclusively focused on their ion currents, was edited and applied to analyze lab-made adulterated mozzarellas. The adulterated mozzarellas were prepared using calibrated mixtures of cow milk and water buffalo milk in our laboratory. This step was fundamental to verify the sensitivity threshold and feasibility of the novel LC− MRM approach in quantifying cow milk. Eventually, the method reliability was also tested by analyzing several retailed mozzarellas and a certified material provided by the Italian Breeder Association. The study has also been able to highlight the important and distinctive nutritional value of this traditional Italian PDO cheese.

2. MATERIALS AND METHODS 2.1. Chemicals and Materials. The following standards were purchased from Aldrich-Fluka-Sigma Chemical (Milan, Italy): retinol, ergocalciferol, δ-tocopherol, β-tocopherol, γ-tocopherol, cholecalciferol, α-tocopherol, menaquinone-4, phylloquinone, all-trans-lutein, alltrans-zeaxanthin, all-trans-β-cryptoxanthin, all-trans-β-carotene, ergocalciferol-d3 [ergocalciferol (6,19,19-d3)]. β-Carotene-d6 [β-carotene(19,19,19,19′,19′,19′-d6)] was obtained from Spectra 2000 Srl (Rome, Italy). Standards of α-tocotrienol, β-tocotrienol, δ-tocotrienol, and γtocotrienol were bought from LGC Standards (Middlesex, U.K.). Standards of 15-cis-phytoene, all-trans-phytoene, all-trans-phytofluene, 13-cis-β-carotene, 9-cis-β-carotene, all-trans-ζ-carotene, all-trans-γ-carotene, all-trans-lycopene, and 5-cis-lycopene were purchased from CaroteNature GmbH (Ostermundigen, Switzerland). All chemicals had a purity grade of >97%. Butylated hydroxytoluene (BHT), provided by Aldrich-Fluka-Sigma Chemical, was used as an antioxidant. Acetonitrile and methanol were of RS-Plus grade (special grade reagents); 2-propanol, hexane, and chloroform were of RS grade (elevated purity grade); and absolute ethanol was of RPE grade (analytical grade). All of these solvents and potassium hydroxide (KOH) were purchased from Carlo Erba (Milan, Italy). Distilled 5411

DOI: 10.1021/acs.jafc.8b01270 J. Agric. Food Chem. 2018, 66, 5410−5417

Article

Journal of Agricultural and Food Chemistry the mixture was vortex-mixed for 5 min and centrifuged at 6000 rpm for 10 min (model PK131R from ALC International, Cologno Monzese, Milan, Italy). Upper hexane layers (approximately 24 mL) were collected in another 50 mL Falcon tube and washed twice with 12 mL of Milli-Q water to remove the residual KOH. After each washing, the mixture was stirred for 5 min and centrifuged at 6000 rpm for 10 min. Aqueous layers were discharged, while the hexane fraction was evaporated at 37 °C under a nitrogen flow until 100 μL and then diluted to a final volume of 200 μL with a mixture of 2propanol/hexane (75:25, v/v). This solution was adopted because the unsaponified lipids, co-extracted with analytes, prevented the extract from being dried. Eventually, 40 μL were injected into the highperformance liquid chromatography−tandem mass spectrometry (HPLC−MS/MS) system. 2.6. LC−MS/MS. Liquid chromatography was performed on a micro HPLC series 200 (PerkinElmer, Norwalk, CT, U.S.A.) equipped with an autosampler, a vacuum degasser, and a column chiller. Analytes were separated on a ProntoSIL C30 column (4.6 × 250 mm, 3 μm) from Bischoff Chromatography (Leonberg, Germany), protected by a guard C30 column (4.0 × 10 mm, 5 μm), under non aqueous-reversed phase (NARP) conditions at 19 °C. Elution was carried out using the following gradient of methanol (phase A) and 2propanol/hexane (50:50, v/v; phase B): 0−1 min, 0% B; 1−15 min, 0−75% B; 15−15.1 min, 75−99.5% B; and 15.1−30.1 min, 99.5% B. The flow rate was 1 mL/min and was entirely introduced into the MS detector. Phase B was also used to wash the autosampler injection device. Analytes were detected by a 4000 Qtrap (AB SCIEX, Foster City, CA, U.S.A.) mass spectrometer equipped with an atmospheric pressure chemical ionization (APCI) probe on Turbo V source. APCI detection was in positive ionization, setting a needle current (NC) of 3 μA and a probe temperature of 450 °C. High-purity nitrogen was used as curtain (40 psi) and collision (4 mTorr) gas, whereas air was used as nebulizer (55 psi) and makeup (30 psi) gas. The Q1 and Q3 mass analyzers were calibrated by infusing a polypropylene glycol solution at 10 μL/min. A full width at half maximum (fwhm) of 0.7 ± 0.1 unit, corresponding to a unit mass resolution, was established and kept in each mass-resolving quadrupole. APCI−Q1−full scan spectra and product ion scan spectra of analytes were acquired by working in flow injection analysis (1−10 ng injected, 1 mL/min flow rate). A targeted screening was carried out in SMRM mode, selecting one MRM transition per analyte. The SMRM algorithm was used with a MRM detection window of 90 s in the retention window characteristic of each analyte (tr of ±1.5 min) and a target scan time of 0.3 s. MRM is a well-known scan mode used for its excellent sensitivity, selectivity, and speed. The SMRM algorithm is an advanced option that intelligently uses information on retention times to automatically optimize the dwell time of each transition and total cycle time. This software tool is useful for maintaining high data quality when the number of compounds/ion currents to be acquired is considerable. Quantitative analysis of the four LMWBs was carried out by selecting two MRM ion currents per analyte. The identity of each biomarker in the matrix was confirmed by matching the retention time and ion ratio (i.e., the relative abundance of the two selected MRM transitions) with the values obtained for the authentic standard in the solvent. Tables 1 and 2 list LC−MS/MS parameters used for (1) targeted screening of LMWBs in cow and water buffalo milk and (2) quantitative analysis of the identified biomarkers in mozzarella samples. Analyst 1.6.2 software (AB Sciex) was used for acquisition and elaboration of LC−MS/MS data. 2.7. Method Validation. The HPLC−MRM method was validated in the matrix using pooled samples of cow and water buffalo mozzarellas (see section 2.3). The evaluated parameters were recovery, precision, sensitivity, linear dynamic range, linearity, limit of detection (LOD), and limit of quantitation (LOQ). After preliminary evaluation of biomarker concentrations, recoveries were calculated on five 6 g replicates from each pool spiked with the ISs and analytes (β-carotene, lutein, β-cryptoxanthin, and ergocalcifer-

Table 1. LC−SMRM Parameters Used for the Targeted Screening of Fat-Soluble Micronutrients in Cow and Buffalo Milk analyte 1 2 3+4 5 6 7+8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

retention time,a average ± SD (min)

MRM transition (m/z)

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

269.1/93.1 397.4/137.2 411.5/151.2 425.3/165.2 402.4/177.2 416.3/151.1 397.3/379.3 385.3/367.3 430.2/165.1 445.3/187.1 551.5/175.0 569.6/477.2 451.5/187.1 553.5/119.1 545.5/69.0 545.5/69.0 543.4/81.0 537.0/177.0 537.5/177.2 537.0/177.0 541.7/69.0 537.5/119.1 537.5/119.1 537.2/119.0

retinol δ-tocotrienol β- + γ-tocotrienol α-tocotrienol δ-tocopherol β- + γ-tocopherol ergocalciferol cholecalciferol α-tocopherol menaquinone-4 all-trans-lutein all-trans-zeaxanthin phylloquinone all-trans-β-cryptoxanthin 15-cis-phytoene all-trans-phytoene all-trans-phytofluene 13-cis-β-carotene all-trans-β-carotene 9-cis-β-carotene all-trans-ζ-carotene all-trans-γ-carotene all-trans-lycopene 5-cis-lycopene

4.7 6.5 7.1 7.8 7.9 8.5 8.7 8.7 9.2 9.6 9.8 10.5 11.9 12.8 15.0 15.2 15.5 16.2 16.8 17.5 17.5 19.3 25.8 26.4

0.3 0.3 0.4 0.3 0.3 0.3 0.4 0.4 0.4 0.3 0.4 0.6 0.3 0.4 0.4 0.4 0.4 0.4 0.5 0.5 0.5 0.5 0.6 0.6

a

The retention times are reported as an arithmetic average of 10 replicates.

Table 2. LC−MRM Parameters Used for the Quantitative Analysis of the Identified LMWBs in Cow and Buffalo Milk and Mozzarella

analyte

retention time,a average ± SD (min)

ergocalciferol

8.7 ± 0.4

all-trans-lutein

9.8 ± 0.4

all-trans-β-cryptoxanthin

12.8 ± 0.4

all-trans-β-carotene

16.8 ± 0.5

internal standards ergocalciferol-d3 β-carotene-d6

8.5 ± 0.4 16.7 ± 0.5

MRM transitionb (m/z)

ion ratioc (%), average ± SD (min)

397.3/107.1 397.3/379.3 551.5/135.2 551.5/175.0 553.5/135.1 553.5/119.1 537.5/119.1 537.5/177.2

94 ± 7 85 ± 9 60 ± 10 78 ± 9

400.3/382.4 543.6/180.4

a

The retention times are reported as an arithmetic average of 10 replicates. bThe first line reports the least intense MRM transition (qualifier, q), and the second line reports the most intense MRM transition (quantifier, Q). cThe ion ratio (relative abundance) between the two MRM transitions is calculated as the percentage intensity ratio of Iq/IQ; the results are reported as an arithmetic average of 10 replicates. ol) at levels 2−3 times higher than the endogenous analyte concentrations. All aliquots were extracted according to what was described in section 2.5, and peak areas were compared to those of another 6 g aliquot that was spiked post-extraction with the same nominal amount of standards and ISs. Intraday precision was 5412

DOI: 10.1021/acs.jafc.8b01270 J. Agric. Food Chem. 2018, 66, 5410−5417

Article

Journal of Agricultural and Food Chemistry Table 3. Recovery, Precision, LOD, and LOQ of the Target Analytes in Cow and Water Buffalo Mozzarellasa precision (%) recoveryb (%)

intraday

interday

LOD (ng/g)

LOQ (ng/g)

analyte

cow

buffalo

cow

buffalo

cow

buffalo

cow

buffalo

cow

buffalo

β-carotene β-cryptoxanthin lutein ergocalciferol

77 91 52 83

91 93 50 77

4 12 20 4

3 20 20 6

4 19 18 5

6 20 20 7

6.5 1.9 8.0 9.9

6.4 4.1 8.7 14

22 6.3 27 33

21 14 29 47

a

See section 2.7 in the Materials and Methods for the details. bThe listed values are representative of absolute recoveries; when corrected for the ISs, relative recoveries ranged from 70 to 100% for all analytes.

Table 4. Linear Regression Parameters for the Analyte Quantification in Cow and Water Buffalo Mozzarellasa R2

calibration curve analyte β-carotene β-cryptoxanthin lutein ergocalciferol a

cow y y y y

= = = =

buffalo

0.004x − 0.998 0.006x − 0.768 0.0009x − 0.095 0.038x + 0.869

y y y y

= = = =

0.003x − 0.007 0.004x − 0.008 0.0004x − 0.009 0.016x + 0.584

cow

buffalo

0.983 0.981 0.862 0.984

0.980 0.995 0.850 0.986

See section 2.7 in the Materials and Methods for the details.

calculated as the relative standard deviation (RSD) of mean recovery, while RSD of recoveries obtained from 10 replicates performed within 2 weeks was representative of interday precision. Internal calibration was performed by analyzing seven aliquots from each pool (C0−C6), six of which (C1−C6) were spiked pre-extraction with increasing concentrations of standards and with the same concentration of ISs (50 μL, 10 ng/μL; see Table S1 of the Supporting Information). Before a calibration curve of a biomarker was constructed, the peak area detected in the C0 aliquot was subtracted from the peak areas of C1−C6 calibrators. Then, the relative peak area (Aanalyte/AIS) was plotted against the fortification level (ng/g). LOD and LOQ were extrapolated as the concentrations able to exceed 3 and 10 times the noise level, respectively. 2.8. Statistical Analysis. Linear regression, means, and standard deviations (SDs) were calculated using Microsoft Excel 2010. Data analysis has been supported by significance test for small size samples, performed using OriginPro8: t test at 99% confidence level has been applied to evaluate if the observed differences between cow and water buffalo mozzarellas were statistically significant or not. Before doing this, the Shapiro−Wilk test was used to check that data, previously purged of suspect values according to Grubbs’ outlier test at 99% confidence level, were normally distributed.

maintaining the desired cycle time for the best signal-to-noise ratio (S/N). As a result, targeted LC−SMRM allowed monitoring of many MRM transitions in a single run without compromising data quality. Analyses were performed on four samples of water buffalo milk and four samples of cow milk from different geographical areas of Latium and Campania. Carotenoids were exclusively detected in cow milk, while ergocalciferol was detected only in water buffalo milk. Figure S1 of the Supporting Information depicts their structures. Figures S2 and S3 of the Supporting Information show representative LC−MRM profiles of the LMWBs, emphasizing differences between cow milk and water buffalo milk. 3.2. Quantitative Analysis of LMWBs Identified in Mozzarellas from Cow and Water Buffalo. The second step consisted of examining mozzarellas prepared in our lab (32 mozzarellas in all) from the milk samples previously analyzed. To this end, a LC−MRM method was set up to be only focused on the ion currents of the identified biomarkers (the SMRM acquisition mode was not necessary because of the limited number of MRM transitions to be monitored). The HPLC− MRM method was validated as described in section 2.7. Tables 3 and 4 list validation parameters for the two kinds of mozzarella. Because saponification conditions were optimized to maximize the extraction of ergocalciferol and β-carotene, absolute recoveries of lutein were around 50% (Table 3). Overall, considering that calibration curves were constructed by fortifying sample aliquots pre-extraction, the determination coefficients were quite satisfactory (Table 4). As an example, Figure S4 of the Supporting Information illustrates calibration curves for β-carotene extracted from (a) water buffalo and (b) cow mozzarellas. From the values of recovery, precision, and coefficient of determination (R2) (Tables 3 and 4), it is possible to infer that β-carotene-d6 is not a good IS for lutein. On the other hand, all validation parameters were optimal for the two most abundant biomarkers, i.e., β-carotene and ergocalciferol. When the validated LC−MRM method was applied for the characterization of the lab-made mozzarellas, it unraveled that concentrations of biomarkers were 3−5 times higher than the levels found in milk used for their preparation. Table 5 compares concentration levels of these micronutrients in milk

3. RESULTS AND DISCUSSION 3.1. Screening of LMWBs in Cow and Water Buffalo Milk. In our previous study dealing with the comprehensive LC−MS profiling of fat-soluble micronutrients in milk from different animal species, β-carotene, β-cryptoxanthin, and lutein were found in cow milk but not in water buffalo milk;20 lutein was also detected in ewe milk. The exclusive occurrence of βcarotene in cow milk was previously observed also by other researchers.5,19,21 In this study, the authentic standards of 26 fat-soluble micronutrients (see Table 1) were used to develop a LC− SMRM method to screen LMWBs of cow milk and water buffalo milk with increased sensitivity. In fact, in comparison to MRM mode, the SMRM algorithm acquires selected MRM transition(s) only in the retention time window of an analyte. Thus, at any one point in time, the number of ion currents to be simultaneously monitored is significantly reduced, leading to higher duty cycles for each analyte. The software computes maximum dwell times for co-eluting compounds while still 5413

DOI: 10.1021/acs.jafc.8b01270 J. Agric. Food Chem. 2018, 66, 5410−5417

Article

Journal of Agricultural and Food Chemistry

concentration was significantly lower than that observed in 100% cow mozzarellas. On the other hand, ergocalciferol was exclusively detected in water-buffalo-based milk products. Figures 1 and 2 illustrate LC−MRM chromatograms of the four biomarkers in cow and water buffalo mozzarellas. Therefore, in accordance with the literature, β-carotene was not detected in water buffalo milk but occurred at concentrations between LOD and LOQ in the final products (Table 5). This outcome was not a result of an accidental contamination because extreme care had been taken during the preliminary preparation phase. On the contrary, it is possible that water buffalo milk contains β-carotene at an endogenous level that is below LOD of the LC−MRM method.20 Thus, this molecule becomes detectable in mozzarella because the preparation process causes pre-concentration of lipids and fatsoluble compounds in finished products. This also means that pure water buffalo mozzarella may not be distinguishable from the mixed mozzarellas merely on the detection or not of βcarotene, as supposed by Cerquaglia et al.5 As a matter of fact,

Table 5. Comparison of Biomarker Concentrations in Cow and Buffalo Milk and Mozzarellas (Two Replicates per Each of the Four Samples)a concentration (ng/g) milk

mozzarella

analyte

cow

buffalo

cow

buffalo

β-carotene β-cryptoxanthin lutein ergocalciferol

160 ± 90 3.5 ± 0.9 12 ± 8