Screening and Assessment of Low-Molecular-Weight Biomarkers of

Subscriber access provided by UNIVERSITY OF SASKATCHEWAN LIBRARY

New Analytical Methods

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 J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.8b01270 • Publication Date (Web): 10 May 2018 Downloaded from http://pubs.acs.org on May 10, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

Journal of Agricultural and Food Chemistry

1

Screening and Assessment of Low Molecular Weight Biomarkers of Milk from Cow and

2

Water Buffalo: an Alternative Approach for the Rapid Identification of Adulterated Water

3

Buffalo Mozzarellas.

4

Chiara Dal Bosco, Stefania Panero, Maria Assunta Navarra, Pierpaolo Tomai, Roberta Curini,

5

Alessandra Gentili*

6

Department of Chemistry, Faculty of Mathematical, Physical and Natural Sciences, University of

7

Rome “La Sapienza“, P.le Aldo Moro 5, 00185, Rome, Italy.

8 9 10 11 12 13

*Corresponding author: Fax number: + 39-06-490631.

14

E-mail address: [email protected] (A. Gentili)

15 16 17 18

1 ACS Paragon Plus Environment


Page 2 of 27

19

ABSTRACT

20

Adulteration of Mozzarella di Bufala Campana with cow milk is a common fraud because of high

21

price and limited seasonal availability of water buffalo milk. In order to identify such adulteration,

22

this work proposes a novel approach based on the use of low molecular weight biomarkers

23

(LMWBs) species-specific. Liquid chromatography-tandem mass spectrometry (LC-MS/MS)

24

screening analyses identified β-carotene, lutein and β-cryptoxanthin as LMWBs of cow milk, while

25

ergocalciferol was found only in water buffalo milk. Adulterated mozzarellas were prepared in

26

laboratory and analyzed for the four biomarkers. Combined quantification of β-carotene and

27

ergocalciferol enabled the detection of cow milk with a sensitivity threshold of 5% (w/w). The

28

method was further tested by analyzing a certificated water buffalo mozzarella and several

29

commercial products. This approach is alternative to conventional proteomic and genomic methods

30

and it is advantageous for routine operations due to its simplicity, speed and low cost.

31 32

KEYWORDS

33

Food analysis; Liquid chromatography-mass spectrometry; mozzarella; milk; biomarkers;

34

adulteration; vitamins; carotenoids.

35 36


Page 3 of 27


37

INTRODUCTION

38

Mozzarella di Bufala Campana is a traditional Italian cheese with Protected Designation of Origin

39

(PDO). It is the third Italian PDO cheese in terms of turnover after Grana Padano and Parmigiano

40

Reggiano and has a supply chain that employs 15,000 people in the center-south of Italy. This

41

prestigious Italian food is manufactured in compliance with a rigorous production protocol based on

42

exclusive use of water buffalo milk in dedicated plants of Campania, Latium and Molise (Art. 4 of

43

the Italian Law Decrees n.91 of 24 June 2014 and n.116 of 11 August 2014; Commission

44

Regulation EC 273/2008). Adulterations with cow milk are widespread because of its low cost and

45

availability, especially in summer time when demand for water buffalo mozzarella increases.

46

Although there are not sanitary-health implications, such fraudulent activities damage consumers

47

and leave serious economic fallouts in the Italian dairy sector. Therefore, simple and rapid

48

analytical methods aimed at identifying adulterations of water buffalo mozzarella are a resource for

49

manufacturing companies as well as official institutions responsible for food quality control. To this

50

end, several analytical methods have been developed so far. 1-16 Most of them have been focused on

51

proteomic evaluation of milk from different animal species by using electrophoretic,

52

chromatographic,5-10 mass-spectrometric11-14 and immunological15 techniques. The European

53

Community’s reference method consists in isoelectric focusing of -caseins after plasmolysis and it

54

is used to detect bovine proteins in cheese made from ewe, goat or water buffalo milk.17 This

55

method can identify percentages of added cow milk as low as 0.5-1% (w/w), but it is long,

56

laborious, and suffering from difficulties of interpretation due to overlapped species-specific bands;

57

moreover, integration with the complicated technique of immunoblotting is sometimes necessary.3

58

In the last years, mass spectrometry-based strategies have been developed to profile peptides and

59

proteins (caseins and whey proteins) in milk from various mammalian species. In particular,

60

MALDI-TOFMS (Matrix Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry)

61

can detect protein pattern differences very effectively and with great sensitivity.12,13 Nevertheless,

62

MALDI does not provide reliable quantitative analyses and involves time-consuming tasks of 3 ACS Paragon Plus Environment

1-4


63

preparation and interpretation. In recent years, PCR has been used to detect cow DNA in milk and

64

mozzarella from water buffalo with a sensitivity threshold of 0.1%-1%.16 In fact, the large amount

65

of somatic cells makes milk a source of DNA and a suitable substrate for PCR amplification;

66

however, the imprecise count of somatic cells in milk used for cheese preparation hampers

67

quantitative determination.

68

Although proteomic and genomic approaches are very sensitive, the quantitative dosage of cow

69

milk in adulterated cheeses is still a serious challenge. Moreover, even if qualitative detection is

70

considered a solved analytical problem, routine analysis needs faster and simpler screening

71

methods. Compared to protein biomarkers, the identification of LMWBs can simplify analytical

72

procedures, save time and reduce cost. Over the past few years, attempts were made by analyzing

73

lipid composition of milk from different animal species, but the procedures were long and

74

demanding.18 More recently, carotenoids have been individuated as diagnostic molecules of cow

75

milk.19,20 In particular, β-carotene was detected in cow milk at concentrations ranging from 10 g/L

76

to 300 g/L depending on season, type of feeding, breed, stage of lactation, etc.19-21 On the basis of

77

this evidence, Cerquaglia et al. have proposed a LC-UV method to identify occurrence of cow milk

78

in water buffalo ricotta through determination of β-carotene.5

79

In the light of what has been discussed above, this work has had the objectives of: i) screening

80

biomarkers of cow milk and water buffalo milk among the fat-soluble micronutrients; ii) developing

81

a LC-MS/MS method to quantify cow milk addition to water buffalo milk used to prepare PDO

82

mozzarellas; iii) verifying the reliability of the novel approach. To these ends, several milk samples

83

from cow and water buffalo were submitted to screening tests, performed by acquiring

84

chromatograms in scheduled-multiple reaction monitoring (SMRM) mode. SMRM is a potent

85

algorithm that allows acquiring many ion currents simultaneously without compromising data

86

quality. In this way, the targeted screening of 26 fat-soluble micronutrients could be carried out

87

with great sensitivity. After identifying four biomarkers, S-MRM was not necessary anymore and a

88

LC-MRM method, exclusively focused on their ion currents, was edited and applied to analyze lab4 ACS Paragon Plus Environment

Page 4 of 27

Page 5 of 27


89

made adulterated mozzarellas. The last ones were prepared by using calibrated mixtures of cow

90

milk and water buffalo milk in our laboratory. This step was fundamental to verify sensitivity

91

threshold and feasibility of the novel LC-MRM approach in quantifying cow milk. Eventually, the

92

method reliability was also tested by analyzing several retailed mozzarellas and a certified material

93

provided by the Italian Breeder Association. The study has also been able to highlight the important

94

and distinctive nutritional value of this traditional Italian PDO cheese.

95 96

MATERIALS AND METHODS

97

Chemicals and Materials. The following standards were purchased from Aldrich-Fluka-Sigma

98

Chemical (Milan, Italy): retinol, ergocalciferol, δ-tocopherol, -tocopherol, γ-tocopherol,

99

cholecalciferol, α-tocopherol, menaquinone-4, phylloquinone, all-trans-lutein, all-trans-zeaxanthin,

100

all-trans-β-cryptoxanthin, all-trans-β-carotene, ergocalciferol-d3 [ergocalciferol (6,19,19-d3)]. ߚ-

101

Carotene-d6 [ߚ-carotene-(19,19,19,19’,19’,19’-d6)] was obtained from Spectra 2000 Srl (Rome,

102

Italy). Standards of α-tocotrienol, -tocotrienol, δ-tocotrienol and γ-tocotrienol were bought from

103

LGC Standards (Middlesex, UK). Standards of 15-cis-phytoene, all-trans-phytoene, all-trans-

104

phytofluene, 13-cis-β-carotene, 9-cis-β-carotene, all-trans-ζ-carotene, all-trans-γ-carotene, all-

105

trans-lycopene and 5-cis-lycopene were purchased from CaroteNature GmbH (Ostermundigen,

106

Switzerland). All chemicals had a purity grade of >97%.

107 108

Butylated hydroxytoluene (BHT), provided by Aldrich-Fluka-Sigma Chemical, was used as an antioxidant.

109

Acetonitrile and methanol were of RS-Plus grade (special grade reagents); 2-propanol, hexane,

110

and chloroform were of RS grade (elevated purity grade); absolute ethanol was of RPE grade

111

(analytical grade). All of these solvents and potassium hydroxide (KOH) were purchased from

112

Carlo Erba (Milan, Italy). Distilled water, used in the extraction procedure based on the cold



Page 6 of 27

113

saponification, was further purified by passing it through a Milli-Q Plus apparatus (Millipore,

114

Bedford, MA USA).

115 116

Screw capped polyethylene centrifuge tube (50-mL) were purchased from Aldrich-Fluka-Sigma Chemical.

117

Standard Solutions. Individual stock solutions of the biomarkers and internal standards (ISs)

118

were prepared by dissolving their weighed amounts (OHAUS DV215CD Discovery Semi-Micro

119

and Analytical Balance 81g/210g capacity, 0.01mg/0.1mg readability) in solvents containing 0.1%

120

(w/v) BHT:

121

- ergocalciferol and ergocalciferol-d3 in ethanol at 1 μg/μL;

122

- β-carotene and ߚ-carotene-d6 in chloroform at 1 μg/μL;

123

- lutein, zeaxanthin and β-cryptoxanthin in chloroform at 0.5 μg/μL.

124 125

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.

126

Multistandard working solutions of the four biomarkers were prepared from their individual

127

solutions by diluting in methanol with 0.1% BHT to obtain concentrations suitable for several

128

experiments. In order to avoid photo-degradation, amber glassware was used for all preparations,

129

which were stored at -18°C in the dark when unused. Ergocalciferol-d3 was the IS for

130

ergocalciferol, while β-carotene-d6 was the IS for β-carotene, lutein, and β-cryptoxanthin.

131

Mozzarella Samples. Cow and water buffalo milks were bought from eight different farms of

132

Central Italy (Latium and Campania). The eight samples (four of cow milk and four of water

133

buffalo milk) were analyzed to screen and quantify LMWBs. Afterwards, cow milk and water

134

buffalo milk samples were paired randomly and mixed in different proportions (0%, 1%, 5%, 10%,

135

25%, 50%, 75%, 100% w/w of cow milk) to prepare four series of lab-made mozzarellas. Their

136

analyses were then performed to check LMWBs reliability and to establish both the method

137

sensitivity and its discriminating power in recognizing adulterated products.


Page 7 of 27


138

Seven cow mozzarellas and eight PDO water buffalo mozzarellas were purchased from

139

supermarkets and retail grocery stores of Latium and Campania. Aliquots from these samples were

140

pooled together within each species (cow or water buffalo) in order to have representative matrices

141

to be used for the method development and validation (see paragraph 2.6).

142

Milk samples and mozzarellas were purchased in the spring/summer season 2017.

143

A certified PDO Mozzarella di Bufala Campana, kindly provided by the Italian Breeder

144

Association, was also analyzed in November 2017.

145

Preparation of Lab-made Mozzarellas. Pure and mixed cow/water buffalo mozzarellas were

146

prepared in laboratory according to the following procedure: 1L of milk was acidified with citric

147

acid until pH 5.6 was reached; 5 g of rennet was added to heated milk (35-37°C), which was then

148

allowed to stand at around 40°C for about 20 minutes. Thereafter, the obtained curd was sliced,

149

squeezed in order to remove excess liquid and heated in microwave oven for 1-2 minutes at 800 W.

150

Finally mozzarella was obtained by quickly kneading the hot curd.

151

Sample Treatment. Extraction of LMWBs from milk samples was performed according to a

152

procedure previously developed for milk.20 This protocol, based on overnight cold saponification

153

followed by liquid-liquid extraction (LLE), was modified and optimized to treat mozzarella

154

samples. In all cases, operations were performed in subdued light, using low actinic glass tubes and

155

wrapping plastic tubes with aluminum foils to protect the analytes from UV light. All organic

156

solvents used for extraction are to be intended as containing 0.1% (w/v) of BHT.

157

Briefly, a 6-g aliquot of mozzarella, minced in very small pieces, was transferred into a 50-mL

158

screw capped polyethylene centrifuge tube and spiked with known amounts of ISs (50 µL of 10

159

ng/µL solution). After a 15-minute period for equilibration at room temperature, 18 mL of ethanol

160

and 6 mL of aqueous KOH solution (50 % w/v) were added and the tube was placed in a water bath

161

at 25°C overnight under continuous stirring. Following the incubation period (15 h), the digest was

162

diluted with 8.5 mL of Milli-Q water and the analytes were extracted by two 12-mL aliquots of

163

hexane. After each aliquot addition, the mixture was vortex-mixed for 5 min and centrifuged at 7 ACS Paragon Plus Environment


Page 8 of 27

164

6000 rpm for 10 min (model PK131R from A.L.C. International, Cologno Monzese, Milan, Italy).

165

Upper hexane layers (approximately 24 mL) were collected in another 50-mL Falcon tube and

166

washed twice with 12 mL of Milli-Q water to remove the residual KOH. After each washing, the

167

mixture was stirred for 5 min and centrifuged at 6000 rpm for 10 min. Aqueous layers were

168

discharged, while the hexane fraction was evaporated at 37°C under a nitrogen flow till 100 µL and

169

then diluted to a final volume of 200 µL with a mixture 2-propanol:hexane (75:25, v/v). This

170

solution was adopted because the unsaponified lipids, co-extracted with analytes, prevented the

171

extract from being dried. Eventually, 40 L were injected into the HPLC-MS/MS system.

172

Liquid Chromatography-Tandem Mass Spectrometry. Liquid chromatography was

173

performed on a micro HPLC series 200 (Perkin-Elmer, Norwalk, CT, USA) equipped with an

174

autosampler, a vacuum degasser and a column chiller.

175

Analytes were separated on a ProntoSIL C30 column (4.6 × 250 mm; 3 μm) from Bischoff

176

Chromatography (Leonberg, Germany), protected by a guard C30column (4.0 x 10 mm; 5 m),

177

under non aqueous-reversed phase (NARP) conditions at 19°C. Elution was carried out using the

178

following gradient of methanol (phase A) and 2-propanol:hexane (50:50, v/v; phase B): 0-1 min 0%

179

B, 1-15 min 0-75% B, 15-15.1 min 75-99.5% B, 15.1-30.1 min 99.5% B. Flow rate was 1 mL/min

180

and was entirely introduced into the MS detector. Phase B was also used to wash the autosampler

181

injection device.

182

Analytes were detected by a 4000 Qtrap® (AB SCIEX, Foster City, CA, USA) mass

183

spectrometer equipped with an APCI probe on Turbo V source. APCI detection was in positive

184

ionization, setting a needle current (NC) of 3 A and a probe temperature of 450 °C. High-purity

185

nitrogen was used as curtain (40 psi) and collision gas (4 mTorr), whereas air was used as nebulizer

186

(55 psi) and make-up gas (30 psi).

187

The Q1 and Q3 mass-analyzers were calibrated by infusing a polypropylene glycol solution at 10

188

L/min. A full width at half maximum (FWHM) of 0.7  0.1 u, corresponding to a unit mass 8 ACS Paragon Plus Environment

Page 9 of 27


189

resolution, was established and kept in each mass-resolving quadrupole. APCI-Q1-full scan spectra

190

and product ion scan spectra of analytes were acquired by working in flow injection analysis (1-10

191

ng injected; 1 mL/min of flow rate).

192

A targeted screening was carried out in SMRM mode, selecting one MRM transition per analyte.

193

The Scheduled MRM™ algorithm was used with a MRM detection window of 90 s in the retention

194

window characteristic of each analyte (tr ± 1.5 min) and a target scan time of 0.3 s. MRM is a well-

195

known scan mode used for its excellent sensitivity, selectivity, and speed. The Scheduled MRM™

196

algorithm is an advanced option that intelligently uses information of retention times to

197

automatically optimize dwell time of each transition and total cycle time. This software tool is

198

useful for maintaining a high data quality when the number of compounds/ion currents to be

199

acquired is considerable.

200

Quantitative analysis of the four LMWBs was carried out by selecting two MRM ion currents

201

per analyte. Identity of each biomarker in matrix was confirmed by matching retention time and ion

202

ratio (i.e. the relative abundance of the two selected MRM transitions) with the values obtained for

203

the authentic standard in solvent. Tables 1 and 2 list LC-MS/MS parameters used for: 1) targeted

204

screening of LMWBs in cow and water buffalo milk and 2) quantitative analysis of the identified

205

biomarkers in mozzarella samples.

206 207

Analyst® 1.6.2 Software (AB Sciex) was used for acquisition and elaboration of LC–MS/MS data.

208

Method Validation. The HPLC-MRM method was validated in matrix using pooled samples of

209

cow and water buffalo mozzarellas (see “Mozzarella Samples”). The evaluated parameters were:

210

recovery, precision, sensitivity, linear dynamic range, linearity, limit of detection (LOD) and limit

211

of quantitation (LOQ).

212

After preliminary evaluation of biomarker concentrations, recoveries were calculated on five 6-g

213

replicates from each pool spiked with the ISs and analytes (-carotene, lutein, -cryptoxanthin,

214

ergocalciferol) at levels 2-3 times higher than the endogenous analyte concentrations. All aliquots 9 ACS Paragon Plus Environment


Page 10 of 27

215

were extracted according to what described in “Sample Treatment” and peak areas were compared

216

with those of another 6-g aliquot that was spiked post-extraction with the same nominal amount of

217

standards and ISs. Intraday precision was calculated as the relative standard deviation (RSD) of

218

mean recovery, while RSD of recoveries obtained from ten replicates performed within two weeks

219

was representative of interday precision. Internal calibration was performed by analyzing seven

220

aliquots from each pool (C0-C6), six of which (C1-C6) were spiked pre-extraction with increasing

221

concentrations of standards and with the same concentration of ISs (50 µL, 10 ng/μL) (see Table

222

S1). Before constructing calibration curve of a biomarker, peak area detected in C0 aliquot was

223

subtracted from the peak areas of C1-C6calibrators. Then, relative peak area (Aanalyte/AIS) was

224

plotted against the fortification level (ng/g). Limits of detection (LOD) and limits of quantification

225

(LOQ) were extrapolated as the concentration able to exceed 3 and 10 times the noise level,

226

respectively.

227

Statistical Analysis. Linear regression, means and standard deviations were calculated using

228

Microsoft Excel 2010. Data analysis have been supported by significance test for small size

229

samples, performed by using OriginPro8: t-test at 99% confidence level has been applied to

230

evaluate if the observed differences between cow and water buffalo mozzarellas were statistically

231

significant or not. Before doing this, Shapiro-Wilk test was used to check that data, previously

232

purged of suspect values according to Grubbs’ outlier test at 99% confidence level, were normally

233

distributed.

234 235

RESULTS AND DISCUSSION

236

Screening of LMWBs in Cow and Water Buffalo Milk. In our previous study dealing with the

237

comprehensive LC-MS profiling of fat-soluble micronutrients in milk from different animal species,

238

β-carotene, β-cryptoxanthin and lutein were found in cow milk but not in water buffalo milk;20

239

lutein was also detected in ewe milk. The exclusive occurrence of β-carotene in cow milk was

240

previously observed also by other researchers.5,19,21 10 ACS Paragon Plus Environment

Page 11 of 27


241

In this study, the authentic standards of 26 fat-soluble micronutrients (see Table 1) were used to

242

develop a LC-S-MRM method so to screen LMWBs of cow milk and water buffalo milk with

243

increased sensitivity. In fact, compared to MRM mode, Scheduled MRM™ algorithm acquires

244

selected MRM transition(s) only in the retention time window of an analyte. Thus at any one point

245

in time, number of ion currents to be simultaneously monitored are significantly reduced, leading to

246

higher duty cycles for each analyte. The software computes maximum dwell times for co-eluting

247

compounds while still maintaining the desired cycle time for best S/N. As a result, targeted-LC-

248

SMRM allowed monitoring many MRM transitions in a single run without compromising data

249

quality. Analyses were performed on 4 samples of water buffalo milk and 4 samples of cow milk

250

from different geographical areas of Latium and Campania. Carotenoids were exclusively detected

251

in cow milk, while ergocalciferol was detected only in water buffalo milk. Figure S1 in the

252

Supporting Information depicts their structures. Figures S2 and S3 show representative LC-MRM

253

profiles of the LMWBs, emphasizing differences between cow milk and water buffalo milk.

254

Quantitative Analysis of LMWBs Identified in Mozzarellas from Cow and Water Buffalo.

255

The second step consisted in examining mozzarellas prepared in our lab (32 mozzarellas in all)

256

from the milk samples previously analyzed. To this end, a LC-MRM method was set up to be only

257

focused on the ion currents of the identified biomarkers (the SMRM acquisition mode was not

258

necessary because of the limited number of MRM transitions to be monitored). The HPLC-MRM

259

method was validated as described in “Method Validation”. Table 3 and 4 list validation

260

parameters for the two kinds of mozzarella. Since saponification conditions were optimized to

261

maximize the extraction of ergocalciferol and β-carotene, absolute recoveries of lutein were around

262

50% (Table 3). Overall, considering that calibration curves were constructed by fortifying sample

263

aliquots pre-extraction, the determination coefficients were quite satisfactory (Table 4). As an

264

example, Figure S4 illustrates calibration curves for β-carotene extracted from water buffalo (a)

265

and cow (b) mozzarella. From the values of recovery, precision and R2 (Tables 3 and 4), it is



266

possible to infer that β-carotene-d6 is not a good IS for lutein. On the other hand, all validation

267

parameters were optimal for the two most abundant biomarkers, i.e. -carotene and ergocalciferol.

268

When the validated LC-MRM method was applied for the characterization of the lab-made

269

mozzarellas, it unraveled that concentrations of biomarkers were 3-5 times higher than the levels

270

found in milks used for their preparation. Table 5 compares concentration levels of these

271

micronutrients in milk and in the relative finished products. As it can be seen, the resulting

272

enrichment made it possible to detect β-carotene also in mozzarellas prepared with 100 % water

273

buffalo milk, even if its concentration was significantly lower than that observed in 100% cow

274

mozzarellas. On the other hand, ergocalciferol was exclusively detected in water buffalo-based milk

275

products. Figures 1 and 2 illustrate LC-MRM chromatograms of the four biomarkers in cow and

276

water buffalo mozzarellas.

277

Therefore, in accordance with the literature, β-carotene was not detected in water buffalo milk,

278

but it occurred at concentrations between LOD and LOQ in the final products (Table 5). This

279

outcome was not a result of an accidental contamination since extreme care had been taken during

280

the preliminary preparation phase. On the contrary, it is possible that water buffalo milk contains β-

281

carotene at an endogenous level that is below LOD of the LC-MRM method.20 Thus, this molecule

282

becomes detectable in mozzarella because the preparation process causes pre-concentration of lipids

283

and fat-soluble compounds in finished products. This also means that pure water buffalo mozzarella

284

may not be distinguishable from the mixed ones merely on the detection or not of β-carotene, as

285

supposed by Cerquaglia et al.5 As a matter of fact, the authors based their deduction on results

286

obtained analyzing water buffalo ricotta, a low-fat cheese (lower concentration effect), by a HPLC-

287

UV/Vis method (lower sensitivity).

288

Sensitivity Threshold of the LC-MRM Method and its Reliability in Performing

289

Quantitative Discriminations. Notwithstanding the results above described, β-carotene can still be

290

considered a cow milk biomarker from a quantitative point of view, since its concentration levels in

291

cow mozzarellas are statistically significantly higher (P < 0.01 by t-test) than those found in water 12 ACS Paragon Plus Environment

Page 12 of 27

Page 13 of 27


292

buffalo mozzarellas. By looking at data concerning the mixed mozzarellas (panel (a) of Figure 3), it

293

can be observed that the minimum cow milk addition that corresponds to a β-carotene concentration

294

statistically different from its endogenous level in pure water buffalo mozzarella is equal to 5%.

295

However, in general, declared water buffalo mozzarellas with β-carotene concentrations ≥20 ng/g

296

might be considered suspect of adulteration. Detection of ergocalciferol indicates that water buffalo

297

milk was surely used to prepare the product, but it is not possible to define exactly in what

298

percentage since, as it can be seen in panel (a) of Figure 3, variability ranges are partially

299

superimposable. Actually, an effective, easy and quick criterion to decide if a product is suspected

300

of being adulterated comes from the simultaneous evaluation of both biomarkers. Then, proteomic

301

or PCR analysis can be applied for confirmation.

302

Lutein and β-cryptoxanthin represent cow’s biomarkers of less practical utility with respect to β-

303

carotene, due to their much lower endogenous concentrations (Table 5). In effect, lutein and β-

304

cryptoxanthin were only detected in mozzarellas with high cow milk percentages (lutein from 75%;

305

β-cryptoxanthin from 25%).

306

Finally, the HPLC-MRM method was applied to analyze seven commercial cow mozzarellas and

307

eight PDO water buffalo mozzarellas. In almost all water buffalo mozzarellas, declared PDO by

308

vendors, β-carotene was detected at concentration levels too low (< LOQ) to be consistent with

309

fraudulent purposes. As can be seen in panel (b) of Figure 3, each series is distributed near the axes

310

of the semi-log graph, allowing a clear distinction among 100% cow mozzarellas (x = 0 ng/g; y >

311

600 ng/g) and 100% water buffalo mozzarellas (x> 175 ng/g; y = 5-20 ng/g). Only one sample of

312

PDO water buffalo mozzarella appears suspect since biomarkers give contrasting indications: β-

313

carotene concentration is that of a mixture with a 3-4 % of cow milk (44 ng/g), while ergocalciferol

314

concentration is as high as 100 % water buffalo mozzarellas. Finally, a certified PDO water buffalo

315

mozzarella, kindly provided by the Italian Breeder Association, was analyzed in triplicate. As it can

316

be seen, the coordinates fall within the box relative to 100% lab-made water buffalo mozzarellas.



Page 14 of 27

317

Surely, the method discriminating power should be confirmed/enhanced by analyzing a higher

318

number of mozzarellas prepared in lab, preferably using certified milks.

319

In conclusion, a new analytical approach to recognize adulterations of water buffalo mozzarella

320

with cow milk has here been proposed. A sensitive screening approach, based on SMRM

321

acquisition mode, was able to identify -carotene and ergocalciferol as biomarkers of cow milk and

322

water buffalo milk, respectively. This work has also verified that -carotene becomes detectable in

323

water buffalo mozzarella because of its enrichment in the finished product. In spite of this, -

324

carotene can be still used as a biomarker since its concentration in water buffalo mozzarella is

325

nearly two orders of magnitude smaller than that in cow mozzarella (Table 4). The simultaneous

326

evaluation of -carotene and ergocalciferol allows the detection of cow milk additions as low as

327

5%. This discriminating threshold can be considered sufficiently sensitive for a preliminary

328

screening analysis because, as considered by other authors,12 adulterations with less than 5% of cow

329

milk are uneconomical. As advantages in using LMWBs consist in simplicity, analysis speed,

330

cheapness and robustness, this method can be conveniently used to check authenticity of water

331

buffalo mozzarellas in routine analyses with high throughput and significant save of time.

332

Our method has also been able to highlight the different micronutrient nutritional properties of

333

mozzarellas from cow and from water buffalo: the former contains antioxidant carotenoids and

334

provitamin A carotenoids, while the latter is a source of vitamin D. This is an important finding,

335

never published so far, with potential implications for counteracting hypovitaminosis-D, especially

336

in the poorest countries. In fact, water buffalo milk is the second most produced in the world and

337

water buffalo milk products are foods that could be easily integrated in a balanced diet.

338

Finally, this study has verified that mozzarellas contain such precious micronutrients at higher

339

concentrations than milks used for their preparation.

340 341

ABBREVIATIONS USED 14 ACS Paragon Plus Environment

Page 15 of 27


342

PDO, Protected Designation of Origin; LMWB, low molecular weight biomarker; BHT, butylated

343

hydroxytoluene;

344

chromatography-atmospheric pressure chemical ionization-tandem mass spectrometry; MRM,

345

multiple reaction monitoring; SMRM, scheduled multiple reaction monitoring; NARP, non-aqueous

346

reversed phase; R2, coefficient of determination.

IS,

internal

standard;

HPLC–APCI-MS/MS,

high

performance

liquid

347 348

ACKNOWLEDGEMENT

349

We would like to acknowledge our appreciation to Italian Breeder Association (Rome, Italy;

350

http://www.aia.it/aia-website/en/home) for their help in providing us a certified PDO Mozzarella di

351

Bufala Campana.

352 FUNDING SOURCES The present work has been funded by Sapienza University of Rome, Research Project 2017 (grant number: RM11715C7DE61769). 353 354

ASSOCIATED CONTENT

355

Supporting Information

356

Names and structures of cow and water buffalo biomarkers (Figure S1); LC-MRM chromatograms

357

of β-carotene and ergocalciferol in cow and water buffalo milks (Figure S2); LC-MRM

358

chromatograms of lutein and β-cryptoxanthin in cow and water buffalo milks (Figure S3);

359

calibration curves of β-carotene obtained from the analysis of cow or water buffalo mozzarellas

360

(Figure S4); spike levels employed to construct the calibration curves (Table S1). This material is

361

available free of charge via the Internet at http://pubs.acs.org.

362 15 ACS Paragon Plus Environment


Page 16 of 27

363

REFERENCES

364 365

1. Mayer, H.K.; Bürger, J.; Kaar, N. Quantification of cow’s milk percentage in dairy products – A myth? Anal. Bioanal. Chem. 2012, 403, 3031–3040.

366 367 368

2. Molina, E.; Martín-Álvarez, J. P.; Ramos, M. Analysis of cows’, ewes’ and goats’ milk mixtures by capillary electrophoresis: Quantification by multivariate regression analysis. Int. Dairy J. 1999, 9, 99–105.

369 370 371

3. Addeo, F.;Pizzano, R.; Nicolai, M.A.; Caira, S.; Chianese, L. Fast isoelectric focusing and antipeptide antibodies for detecting bovine casein in adulterated water buffalo milk and derived mozzarella cheese. J. Agric. Food Chem. 2009, 57, 10063–10066.

372 373 374

4. Cartoni, G.P.; Coccioli, F.; Jasionowska, R.; Masci, M. Determination of cow milk in buffalo milk and mozzarella cheese by capillary electrophoresis of the whey protein fractions. Ital. J. Food Sci. 1998, 10, 127-135.

375 376

5. Cerquaglia, O.; Sottocorno, M.; Pellegrino, L.; Ingi, M. Detection of cow’s milk, fat or whey in ewe and buffalo ricotta by HPLC determination of β-carotene. Ital. J. Food Sci. 2011, 23, 367-372.

377 378 379

6. Chen, R. K.; Chang, L. W.; Chung, Y. Y.; Lee, M. H.; Ling, Y. C. Quantification of cow milk adulteration in goat milk using high-performance liquid chromatography with electrospray ionization mass spectrometry. Rapid Commun. Mass Spectrom. 2004, 18, 1167-1171.

380 381 382

7. De Noni, I.; Tirelli, A.; Masotti, F. Detection of cows’ milk in non-bovine cheese by HPLC of whey protein: Application to goat milk cheese. Scienza e Tecnica Lattiero-Casearia. 1996, 47, 7– 17.

383 384 385 386

8. Enne, G.; Elez, D.; Fondrini, F.; Bonizzi, I.; Feligini, M.; Aleandri, R. High performance liquid chromatography of governing liquid to detect illegal bovine milk’s addition in water buffalo mozzarella: Comparison with results from raw milk and cheese matrix. J. Chromatogr. A. 2005, 1094, 169–174.

387 388 389

9. Ferreira, I.M.; Caçote, H. Detection and quantification of bovine, ovine and caprine milk percentages in protected denomination of origin cheeses by reversed-phase high-performance liquid chromatography of betalactoglobulins. J. Chromatogr. A. 2003, 1015, 111–118.

390 391 392

10. Russo, R.; Severino, V.; Mendez, A.; Lliberia, J.; Parente, A.; Chambery, A. Detection of buffalo mozzarella adulteration by an ultra-high performance liquid chromatography tandem mass spectrometry methodology. J. Mass Spectrom. 2012, 47, 1407-1414.

393 394 395

11. Angeletti, R.; Gioacchini, A. M.; Seraglia, R.; Piro, R.; Traldi, P. The potential of Matrixassisted Laser Desorption/Ionization Mass Spectrometry in the quality control of water buffalo mozzarella cheese. J. Mass Spectrom. 1998, 33, 525-531.

396 397 398

12. Cozzolino, R.; Passalacqua, S.; Salemi, S.; Garozzo, D. Identification of adulteration in water buffalo mozzarella and in ewe cheese by using whey proteins as biomarkers and matrix-assisted laser desorption/ionization mass spectrometry. J. Mass Spectrom. 2002, 37, 985–991 16 ACS Paragon Plus Environment

Page 17 of 27


399 400 401

13. Czerwenka, C.; Műller, L.; Lindner, W. Detection of the adulteration of water buffalo milk and mozzarella with cow’s milk by liquid chromatography–mass spectrometry analysis of βlactoglobulin variants. Food Chem. 2010, 122, 901–908.

402 403 404

14. Sassi, M.; Arena, S.; Scaloni, A. MALDI-TOF-MS platform for integrated proteomic and peptidomic profiling of milk samples allows rapid detection of food adulterations. J. Agric. Food Chem. 2015, 63, 6157−6171.

405 406 407

15. Hurley, I. P.; Ireland, H. E.; Coleman, R. C.; Williams, J. H. H. Application of immunological methods for the detection of species adulteration in dairy products. Int. J. Food Sci. Technol. 2004, 39, 873–878.

408 409

16. Dalmasso, A.; Civera, T.; La Neve, F.; Bottero, M. T. Simultaneous detection of cow and buffalo milk in mozzarella cheese by Real-Time PCR assay. Food Chem. 2011,124, 362-366.

410 411 412 413

17. Commission Regulation (EC) No 273/2008 of 5 March 2008 laying down detailed rules for the application of Council Regulation (EC) No 1255/1999 as regards methods for the analysis and quality evaluation of milk and milk products. Official Journal of the European Union L88, 51, 1– 115.

414 415 416

18. Fontecha, J.; Mayo, I.; Toledano, G. M.; Juárez, M. Triacylglycerol Composition of Protected Designation of Origin Cheeses During Ripening. Authenticity of Milk Fat. J. Dairy Sci. 2006, 89, 882-887.

417 418

19. Abd El-Salam, M. H.; El-Shibiny, S. A comprehensive review on the composition and properties of buffalo milk. Dairy Sci. & Technol. 2011, 91, 663–699.

419 420 421

20. Gentili, A.; Caretti, F.; Bellante, S.; Ventura, S.; Canepari, S.; Curini, R. Comprehensive profiling of carotenoids and fat-soluble vitamins in milk from different animal species by LC-DADMS/MS hyphenation. J. Agric. Food Chem. 2013, 61, 1628-1639.

422 423 424

21. Nozière, P.; Grolier, P.; Durand, D.; Ferlay, A.; Pradel, P.; Martin, B. Variations in Carotenoids, Fat-Soluble Micronutrients, and Color in Cows’ Plasma and Milk Following Changes in Forage and Feeding Level. J. Dairy Sci. 2006, 89, 2634-2648.

425



Page 18 of 27

426

Figure Captions

427

Figure 1. LC-MRM profiles of ergocalciferol and β-carotene in mozzarella from cow (a;c) and

428

from water buffalo (b;d).

429 430

Figure 2. LC-MRM profiles of lutein and β-cryptoxanthin in mozzarella from cow (e;g) and from

431

water buffalo (f;h).

432 433

Figure 3. Panel (a) illustrates the semi-log graph constructed by analyzing the concentrations of β-

434

carotene and ergocalciferol in mozzarellas prepared in lab from calibrated mixtures of cow milk and

435

water buffalo milk. Panel (b) shows analytical results of retail cow mozzarellas and water buffalo

436

mozzarellas, including a certified PDO product.

437


Page 19 of 27


438 439

Tables Table 1. LC-S-MRM parameters used for the targeted screening of fat-soluble micronutrients in cow’s and buffalo’s milk. Analytes

Retention timea Average± SD (min)

MRM transitions (m/z)

1

retinol

4.7  0.3

269.1/93.1

2

δ-tocotrienol

6.5  0.3

397.4/137.2

3+4

+γ-tocotrienol

7.1  0.4

411.5/151.2

5

α-tocotrienol

7.8  0.3

425.3/165.2

6

δ-tocopherol

7.9 ± 0.3

402.4/177.2

7+8

+γ-tocopherol

8.5 ± 0.3

416.3/151.1

9

ergocalciferol

8.7  0.4

397.3/379.3

10

cholecalciferol

8.7  0.4

385.3/367.3

11

α-tocopherol

9.2 ± 0.4

430.2/165.1

menaquinone-4

9.6 ± 0.3

445.3/187.1

all-trans-lutein

9.8 ± 0.4

551.5/175.0

all-trans-zeaxanthin

10.5 ± 0.6

569.6/477.2

15

phylloquinone

11.9 ± 0.3

451.5/187.1

16

all-trans-β-cryptoxanthin

12.8 ± 0.4

553.5/119.1

17

15-cis-phytoene

15.0 ± 0.4

545.5/69.0

18

all-trans-phytoene

15.2 ± 0.4

545.5/69.0

all-trans-phytofluene

15.5 ± 0.4

543.4/81.0

13-cis-β-carotene

16.2 ± 0.4

537.0/177.0

all-trans-β-carotene

16.8 ± 0.5

537.5/177.2

22

9-cis-β-carotene

17.5 ± 0.5

537.0/177.0

23

all-trans-ζ-carotene

17.5 ± 0.5

541.7/69.0

24

all-trans-γ-carotene

19.3 ± 0.5

537.5/119.1

25

all-trans-lycopene

25.8 ± 0.6

537.5/119.1

5-cis-lycopene

26.4 ± 0.6

537.2/119.0

12 13 14

19 20 21

26 a

The retention times are reported as arithmetic average of ten replicates.

440



Page 20 of 27

Table 2. LC-MRM parameters used for the quantitative analysis of the identified LMWBs in cow’s and buffalo’s milk and mozzarella. Retention time MRM transitions Ion ratio (%) Analytes Average± SD (min) (m/z) Average± SD (min) 397.3/107.1 ergocalciferol 947 8.7  0.4 397.3/379.3 all-trans-lutein

9.8 ± 0.4

551.5/135.2 551.5/175.0

859

all-trans-β-cryptoxanthin

12.8 ± 0.4

553.5/135.1 553.5/119.1

60 10

all-trans-β-carotene

16.8 ± 0.5

537.5/119.1 537.5/177.2

789

8.5 0.4

400.3/382.4

-

16.7  0.5

543.6/180.4

-

Internal standards ergocalciferol-d3 β-carotene-d6 a

The retention times are reported as arithmetic average of ten replicates. b The first line reports the least intense MRM transition (qualifier, q) and the second line the most intense one (quantifier, Q).c The ion ratio (relative abundance) between the two MRM transitions is calculated as percentage intensity ratio of Iq/IQ; the results are reported as arithmetic average of ten replicates.


Page 21 of 27


Table 3. Recovery, precision, LOD and LOQ of the target analytes in cow’s and water buffalo’s mozzarellasa. Precision (%) Recoverya (%)

LOD (ng/g)

Analyte

441 442 443

intraday

LOQ (ng/g)

interday

cow

buffalo

cow

buffalo

cow

buffalo

cow

buffalo

cow

buffalo

β-carotene

77

91

4

3

4

6

6.5

6.4

22

21

β-cryptoxanthin

91

93

12

20

19

20

1.9

4.1

6.3

14

Lutein

52

50

20

20

18

20

8.0

8.7

27

29

Ergocalciferol

83

77

4

6

5

7

9.9

14

33

47

a

b

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

444 445



Page 22 of 27

Table 4. Linear regression parameters for the analyte quantification in cow’s and water buffalo’s mozzarellasa. R2

calibration curves Analyte

446

cow

buffalo

cow

buffalo

β-carotene

y = 0.004x - 0.998

y = 0.003x - 0.007

0.983

0.980

β-cryptoxanthin

y = 0.006x - 0.768

y = 0.004x - 0.008

0.981

0.995

Lutein

y = 0.0009x – 0.095

y = 0.0004 x – 0.009

0.862

0.850

Ergocalciferol

y = 0.038x + 0.869

y = 0.016x + 0.584

0.984

0.986

a

See “Method Validation” in the section “Materials and Methods” for the details.

447 448 449


Page 23 of 27


450 Table 5. Comparison of biomarker concentrations in cow’s and buffalo’s milk and mozzarella (2 replicates per each of the 4 samples). Concentration (ng/g) Analyte

milk

mozzarella

cow

buffalo

cow

buffalo

β-carotene

160  90

n.d.

790  90

105

β-cryptoxanthin

3.5  0.9

n.d.

11  6

Screening and Assessment of Low-Molecular-Weight Biomarkers of

Recommend Documents