MS based method for assessing saffron (Crocus sativus L.) adulteration

Journal Pre-proofs A rapid MALDI MS/MS based method for assessing saffron (Crocus sativus L.) adulteration Donatella Aiello, Carlo Siciliano, Fabio Mazzotti, Leonardo Di Donna, Constantinos M. Athanassopoulos, Anna Napoli PII: DOI: Reference:

S0308-8146(19)31646-2 https://doi.org/10.1016/j.foodchem.2019.125527 FOCH 125527

To appear in:

Food Chemistry

Received Date: Revised Date: Accepted Date:

12 September 2018 2 May 2019 12 September 2019

Please cite this article as: Aiello, D., Siciliano, C., Mazzotti, F., Di Donna, L., Athanassopoulos, C.M., Napoli, A., A rapid MALDI MS/MS based method for assessing saffron (Crocus sativus L.) adulteration, Food Chemistry (2019), doi: https://doi.org/10.1016/j.foodchem.2019.125527

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Elsevier Ltd. All rights reserved.

1

A rapid MALDI MS/MS based method for assessing saffron (Crocus sativus L.) adulteration

2

3

4

Donatella Aiello±, Carlo Siciliano♯*, Fabio Mazzotti±, Leonardo Di Donna±,

5

Constantinos M. Athanassopoulos‡, Anna Napoli±*

6

7

8

9

10

±Department

♯Department

of Chemistry and Chemical Technologies, University of Calabria, Italy.

of Pharmacy, Health and Nutritional Sciences, University of Calabria, Italy

‡Department

of Chemistry, University of Patras, Patras, Greece.

11

12

13

14

15 16 17 18 19 20

* Corresponding authors: Prof. Anna Napoli, Department of Chemistry and Chemical Technologies Via P. Bucci Cubo12/d 87036 Arcavacata di Rende (CS), Italy e-mail: [email protected]

21 22 23 24 25 26

Prof. Carlo Siciliano, Department of Pharmacy, Health and Nutritional Sciences, Edificio Polifunzionale I-87036 Arcavacata di Rende (CS), Italy e-mail: [email protected]

27

1

28

ABSTRACT

29

We report on a sensitive and fast quantitative MALDI-MS/MS method used to assess saffron

30

authenticity by direct analysis through the determination of picrocrocin as the saffron authenticity

31

marker, and using curcumin as the non-isotopic isobaric internal standard. The internal standard

32

curcumin yielded good linearity (R2 = 0.994), and with confidence intervals at 95% for intercept.

33

The detectable maximum adulteration percentage (99.0%) was estimated interpolating the limit of

34

detection (LOD) for the isobaric internal standard in linear regression. The LOD was 47.63 ppm,

35

and LOQ was 56.53 ppm. Good accuracy and precision were obtained for all concentrations. The

36

capability of the MS approach to monitor analytes in a specific, selective fashion was used to obtain

37

a semi-quantitative adulteration percentage and to establish the adulterant by additional

38

experiments. The detection of gardecin and its derivatives in commercial samples indicated that

39

Gardenia jasminoides Ellis was used as the adulterant.

40

41

Keywords: saffron, quantitation, picrocrocin, curcumin, mass spectrometry, adulteration.

42

43

Abbreviations: MALDI, matrix assisted laser desorption; MS, mass spectrometry; MS/MS, tandem

44

mass spectrometry; LOD, limit of detection; LOQ, limit of quantitation.

45

46

1. Introduction

47

Saffron, the red dried stigmas of Crocus sativus L., is the world's most expensive spice and

48

thereby is considered within the major candidates for economically motivated fraud (Moore, et

49

al., 2012). Saffron authentication through established methodologies is a challenging task, as 2

50

saffron of higher quality may intentionally be blended with plant-derived adulterants. The most

51

frequently used adulterants are saffron stamens, safflower, calendula, turmeric rhizomes or dried

52

gardenia fruits (M. Carmona, et al. 2006, Sabatino, et al.2011, Petrakis, et al. 2015, Johnson,

53

2014). Fruits of Gardenia jasminoides Ellis represent a bio-adulterant which is difficult to detect

54

by classical methods, because it contains crocins (C-1÷C-3) and flavonoids as does saffron itself

55

(Pfister, et al., 1996).

56

The quality of saffron and its commercial value are determined by specifications described

57

within the ISO/TS-3632 standard (ISO 3632-1, 2011; ISO 3632-2, 2010) that established

58

spectrophotometric (for picrocrocin and safranal) and chromatographic (for crocins and polar

59

dyes) measurements. According to the ISO/TS-3632 standard, the maximum mass fraction of

60

foreign matter permitted in the third-class products is 1% (w/w). The standard UV–vis

61

spectrophotometric method of ISO 3632-2 for grading saffron may not reveal saffron

62

adulteration with amounts lower to 20% (w/w) of safflower, turmeric, or calendula (Sabatino, et

63

al., 2011). Many analytical methods have been developed for authentication of saffron

64

(Sabatino, et al., 2011, Alonso, et al., 1998, Zalacain, et al., 2005, Maggi, et al., 2011, Ordoudi,

65

et al. 2014, Sereshti, et al. 2018, García-Rodríguez, et al., 2014), including strategies based on

66

the use of NMR (Petrakis, et al., 2015), LC-MS, and molecular techniques (Sabatino, et al.,

67

2011, Rubert, et al., 2016, Guijarro-Díez, et al. 2017, Guijarro-Díez, et al., 2017). At present,

68

there is a growing tendency to find quick, simple and powerful tools to differentiate pure and

69

adulterated saffron. These methods should also measure the adulteration levels regardless the

70

adulterant.

71

Mass spectrometry is a powerful tool for the high-throughput detection and quantitation of

72

metabolites, amino acids (De Marco, et al., 2010) and their synthetic analogues and proteins.

73

Matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) and tandem mass

74

spectrometry (MS/MS) techniques have seldom been considered for the analysis of saffron 3

75

extracts (Koulakiotis, et al., 2012), and for quantifying adulterants in saffron. Several studies

76

have shown that MALDI can be used as an alternative to LC-ESI for the highly sensitive

77

analysis of low molecular weight compounds in complex matrices (Persike, et al., 2010, van

78

Kampen, et al., 2011, Aiello, et al. 2018, Persike, et al., 2009). Direct MS analysis of foods and

79

food extracts has been proposed as a useful and robust approach to the chemical fingerprinting,

80

when a rapid classification of food-sample types or rapid screening of food adulteration is

81

wanted.

82

Authenticity assessments can successfully been performed by powerful analytical approaches

83

based on MALDI-TOF/TOF-MS (Herrero, et al. 2012, Aiello, et al., 2015). This MS technique

84

is extremely advantageous due to short analysis times, high sensitivity, tolerance to

85

contaminants, and the ability to detect different components in highly complex mixtures.

86

Moreover, MALDI-MS analysis can be combined with a rapid and simple preparation of the

87

sample, preventing any possible analyte loss (Napoli, et al., 2014). Several MS and MS/MS

88

based methods have been developed to achieve relative and absolute quantitative measurements

89

of target low molecular weight analytes, using both isotopically labeled and unlabeled synthetic

90

compounds as the standards. MALDI MS/MS provided quantitation of target compounds and

91

small sets of analytes in a complex matrix with great sensitivity, dynamic range, and precision

92

(Persike, et al., 2010, van Kampen, et al., 2011, Persike, et al., 2009). The high-quality MS/MS

93

quantification called for the synthesis of stable isotope (2H,

94

analogues as the internal standards (Persike, et al., 2010, Di Donna, et al. 2015, Mazzotti, et al.,

95

2014). In fact, the common workflow requires the construction of a calibration curve with

96

standard solutions containing the same (fixed) amount of the stable isotope internal standard,

97

and variable amounts of the single specific analyte of interest. The constant of proportionality

98

for a single analyte can be established, and ion abundance ratios can be converted into absolute

13C

and

15N)

labeled analyte

4

99

amounts. This approach is suitable only when stable isotope internal standards are available for

100

each analyte of interest, and the analyte concentration levels to be measured are known.

101

Saffron authentication processes through the targeted quantitative measurements should require

102

the preparation of standard solutions of a specific saffron metabolite mixed with fixed amount

103

of the corresponding stable isotope internal standard. However, synthetic isotope labeled

104

markers of saffron are not available. To overcome this drawback, we evaluated the use of whole

105

extracts obtained from sets of standard sample (w/w), with the addition of a non-isotopic

106

isobaric internal standard (IIS). The use of an IIS might represent the best choice, because

107

quantification by MALDI MS/MS can be performed on whole extracts without prior

108

chromatographic separations of analytes.

109

The aim of this work was to develop a fast and sensitive method based on MALDI MS/MS for

110

the quality control of saffron regardless of the adulterant employed, by quantifying the

111

biomarker picrocrocin in the presence of a non-isotopic isobaric internal standard. The MALDI

112

MS spectrum acquired from a crude extract of powdered saffron might reveal the presence of

113

the adulterant by the simultaneous detection of picrocrocin and the target marker of the

114

adulterant.

115

116

2. Materials and Methods

117

2.1 Chemicals.

118

Solvents (CH3CN, and H2O, HPLC grade), -cyano-4-hydroxy-trans-cynnamic acid (-CHCA,

119

pure 99.0%), sinapinic acid (SA, pure 99.0%) and curcumin (assay ≥ 98.0%, CAS Number 458-37-

120

7) were purchased from Sigma Aldrich Fluka (Milano, Italy).

121

2.2 Spice. 5

122

Samples of saffron spice were directly obtained from producers, with a guarantee of their origin and

123

freedom from fraud. Dried Crocus sativus L. stigmas were obtained from Cooperative of Saffron,

124

(Krokos Kozanis, Greece). Five powdered saffron samples from different brands, suspected of

125

adulteration on the basis of their low costs and the questionable origins, and Calendula officinalis

126

L., were purchased from a local store.

127

128

2.3 Sample preparation.

129

Crocus sativus L. stigmas, Calendula officinalis L. and Citrus leaves were ground into a powder.

130

Aliquots of powdered calendula (chosen as the blank matrix) were used to prepare standard samples

131

for the picrocrocin dilution series. Aliquots of powdered calendula and citrus leaves were also used

132

to prepare spiked samples.

133

Solvent system used for extraction. The solvent system for extraction was as follows:

134

H2O/CH3CN (40:60, v/v) with 0.3% TFA.

135

Extraction procedure. A portion (5 mg) of each standard sample was extracted with the solvent

136

system (1 mL), under magnetic stirring at room temperature for 2 min. After 2 min. centrifugation

137

at 9660 g, the pellet was discarded and the resulting solution was used for MALDI-MS and MS/MS

138

experiments. Sample aliquots (1 µL) were spotted 3-fold and in triplicate on the MALDI plate, and

139

dried at room temperature. Matrix solution (1 µL) was pipetted onto dried samples. After the

140

crystals had dissolved completely, the spots were dried under a continuous air stream.

141

Preparation of spiked saffron samples. Spiked samples were prepared by adding the required

142

amounts of saffron and calendula (w/w), or citrus leaves (w/w), corresponding to the desired

143

adulteration percentage.

6

144

145

2.4 Stock solution of the non-isotopic isobaric internal standard (IIS).

146

A stock solution was prepared by dissolving curcumin in CH3CN, reaching the final concentration

147

of 0.5 g/L.

148

149

2.5 Working calibration solutions.

150

Extracts were obtained from the eight standard samples 1-8 (Table 1S, supplementary material).

151

Two sets of standard samples were prepared by adding the required amount of authentic saffron and

152

blank matrix (calendula).

153

Set I. A mixture of the solvent system (980 µL), and curcumin stock solution (20 µL) was added to

154

standard samples 1-8 (5 mg). The IIS final concentration was 10 mg/L (27 pmol/µL). The

155

concentrations of standard solutions were 5000-3500 mg/L, referred to the dry material (Table 1S).

156

This set was used to calculate the calibration curve.

157

Set II. The solvent system (1 mL) was added to standard samples 1-8 (5 mg). The concentrations of

158

standard solutions were 5000-3500 mg/L, referred to the dry material (Table 1S). This set was used.

159

to determine the possible blank matrix effect on specific ion signals.

160

2.6 MALDI-TOF-MS and CID-MS/MS analysis.

161

Each sample was directly spotted three times on a 384-well insert Opt-TOFTM stainless steel

162

MALDI plate (AB SCIEX, Darmstadt, Germany). Mass spectrometric analyses were performed

163

using a 5800 MALDI-TOF-TOF Analyzer (AB SCIEX, Darmstadt, Germany) equipped with an

164

Nd:YLF Laser with λ = 345 nm wavelength of < 500 ps pulse length and p to 1000 Hz repetition

165

rate, in reflectron positive mode with a mass accuracy of 5 ppm. Mass spectra were acquired 7

166

automatically in the positive reflector mode between 200 and 2000 with fixed laser intensity.

167

Spectra with signal-to-noise below 200 were discarded automatically by the instrument. The

168

operation parameter was optimized for the mass region of interest. Laser intensity was adjusted

169

manually to avoid detector saturation. At least 4000 laser shots were typically accumulated with a

170

laser pulse rate of 400 Hz in the MS mode, whereas in the MS/MS mode spectra up to 5000 laser

171

shots were acquired and averaged with a pulse rate of 1000 Hz. MS/MS experiments were

172

performed at a collision energy of 1 kV, ambient air was used as collision gas at a pressure of 10-6

173

Torr. The potential difference between the source acceleration voltage and the collision cell was set

174

as 1 kV. Each spot was measured three times with a precursor selection of 369 (± 0.5 Da) for

175

picrocrocin and the non-isotopic isobaric internal standard. After acquisition, spectra were handled

176

using Data Explorer version 4.11 (AB Sciex). To reduce the inhomogeneous co-cristallization of the

177

analyte with the matrix, a fast drying protocol was adopted (Persike, et al., 2009, Persike, et al.

178

2010).Fluctuations in signal intensities were overcome by averaging over a high number of laser

179

shot (5000) and working with a large part of the sample spot. The MALDI matrix sinapinic acid

180

(SA) was prepared at a concentration of 20 mg/mL in H2O/CH3CN (40:60, v/v) with 0.3% TFA.

181

2.7 Nomenclature for crocetin esters and gardecin.

182

To abbreviate the names of crocetin esters and gardecin in this paper, they were labeled as follows:

183

the nomenclature which refers to the isomeric cis and trans forms was written with a hyphen

184

separating the total number of the glucose moieties at both extremes of the base molecule (C-n).

185

Then, C-4 and G-2 would indicate Ct (crocetin) and gardecin with four and two hexose (Hex)

186

residues, respectively.

187

2.8 Method validation.

8

188

Each standard solution 1-8 was spotted three times, and each spot was sampled in triplicate for a

189

total of nine data points for each concentration. All data presented in this work are averages of three

190

replicates. The linear range was assessed by plotting the analyte isotopic cluster area, divided by the

191

isotopic cluster area of IIS, and multiplied by the IIS concentration, versus the analyte concentration

192

in dry material. The dilution series 1-8 is plotted in Figure 1S A,B (for related data see Table 2S in

193

supplementary material).

194

Calibration curve for picrocrocin was calculated in the concentration range of 5000-3500 mg/L,

195

referred to the weighed amount of dry material, using curcumin as the IIS at a concentration of 10

196

mg/L (Table 1S). Picrocrocin concentration was reported in mg/L, and the estimated concentration

197

for spiked and S1-S5 samples was reported in mg/Kg, referred to the weighed amount of dry

198

material. The IIS was tested in the concentration range from 5 mg/L to 15 mg/L, in order to

199

determine the appropriate concentration level to be used for the quantitative analysis. The best level

200

for IIS was found to be 10 mg/L. Linear regression (R2), relative standard deviation (RSD), and

201

accuracy were calculated with the Microsoft Excel software. Accuracy was calculated from the

202

experimentally determined concentrations, compared to the respective nominal values.

203

2.9 Analytical Parameters.

204

The limit of detection (LOD) and the limit of quantitation (LOQ) were calculated by applying Eqs 1

205

and 2, following the directives of IUPAC and the American Chemical Society’s Committee on

206

Environmental Analytical Chemistry. SLOD is the signal at the limit of detection, SLOQ is the signal

207

at the limit of quantitation, SRB is the signal of the blank “authentic saffron samples”, and σRB is the

208

standard deviation.

209

Eq. 1

SLOD = SRB + 3σRB

210

Eq. 2

SLOQ = SRB +10σRB 9

211

2.10 Direct detection of major saffron components.

212

Several glycoconjugated carotenoid breakdown products showing a common trimethylcyclohexene

213

scaffold were isolated from Crocus sativus L. stigmas (Winterhalter, et al. 2000). Among these

214

products, picrocrocin (1, Fig. 1A) has been reported as a valuable authenticity marker for saffron

215

(Tarantilis, et al., 1994; Kanakis, et al., 2004). The hydrophilic properties of the major components

216

of saffron (cis- and trans-crocins, picrocrocin and its related compounds) led to the hypothesis that

217

an acid binary aqueous solvent system could disrupt cell membranes, and should favor the release

218

of metabolites in aqueous media (Aiello, et al. 2016). Therefore, powdered saffron was subjected to

219

a brief (2 minutes) extraction with solvents. An aqueous solution of 0.3% TFA, and a mixture of

220

H2O/CH3CN (40:60, v/v) with 0.3% TFA, were tested. The latter solvent system proved to be the

221

most suitable. Then, a small aliquot of the crude extract was directly placed on the MALDI sample

222

plate and analyzed. The identity of crocins (C-1÷C-4) and picrocrocin was confirmed by MALDI-

223

MS and MS/MS measurements. When the extraction time was extended to 5 minutes, no significant

224

changes in the extent of crocins from sample were observed by MALDI-MS. The extraction of

225

saffron with a solution of H2O/CH3CN (40:60, v/v) with 0.3% TFA was adopted to fulfill the direct

226

detection of spice endogenous metabolites. The recorded spectrum is depicted in Fig. 2S

227

(supplementary material).

228

Figure 1 to be inserted here

229

2.11 Experimental design. Fig. 2 displays the developed approach for the quantitative

230

determination of saffron adulteration. The strategy has two stages: the non-isotopic isobaric internal

231

standard (IIS) selection and validation (Fig. 2, panel A-D), and method development and

232

implementation (Fig. 2, panel E-G). To choose an IIS, a specific set of endogenous analytes of

233

saffron was evaluated (crocins and picrocin), the chosen based on the molecular weight, the

234

fragmentation pattern and the availability. Thus, the choice fell on curcumin ((1E,6E)-1,7-bis(410

235

hydroxy-3-methoxyphenyl)hepta-1,6-diene-3,5-dione (2), Fig. 1C), since it is an isobar of

236

picrocrocin (4-(-D-glucopyranosyloxy)-2,6,6-trimethyl-1-cyclohexene-1-carboxaldehyde (1, Fig.

237

1A), which is present in saffron spice from 0.8% to 26.6% on a dry basis (Alonso, et al., 2001,

238

Iborra, et al., 1992, Sánchez, et al., 2008). The potential IIS was preliminary evaluated by MALDI-

239

MS and MS/MS experiments, providing qualitative information about ionization efficiency and

240

fragmentation via collision induced dissociation (CID MS/MS). Informative and abundant fragment

241

ions of picrocrocin and IIS were selected.

242

Figure 2 to be inserted here

243

The authentication process through the targeted quantitative measurements requires the preparation

244

of standard solutions of a specific saffron metabolite, containing a fixed amount of the

245

corresponding stable isotope internal standard. Synthetic stable isotope-labeled crocin and

246

picrocrocin are not available. The only commercially available standard of crocin does not have a

247

specific purity (Cossignani, et al., 2014). To overcome this drawback, we evaluated the use of sets

248

of standard samples as the source of picrocrocin dilution series. The strategy required the selection

249

of a blank matrix, and validation of dilution series by MALDI MS and MS/MS.

250

MALDI-MS of crude extracts of calendula (Fig. 3S, supplementary material), and saffron did not

251

overlap. Standard samples belonging to set II were further analyzed, and MS/MS of the ion of m/z

252

369.13, assigned to picrocrocin, confirmed any overlap to be absent.

253

The dilution series was validated by additional MALDI-MS experiments. A set of standard samples

254

was prepared containing IIS (Set I, Table 1S). Then, the abundance of a specific fragment ion from

255

both picrocrocin and IIS was measured by MS/MS experiments, as a function of the picrocrocin

256

level (w/w). The experiments resulted in the highly specific and sensitive measurement of both

257

internal standard and analyte, directly from complex mixtures. These ions were monitored in a rapid 11

258

succession by MS/MS performed on standard solutions. Since measurements were performed by

259

direct MS/MS in presence of IIS, any possible analyte loss coming from sample handling, as well as

260

the variability during sample loading, did not affect the picrocrocin/IIS abundance ratios. Since an

261

absolute amount of IIS was added (10 mg/L, 27 pmol/L), the ratio of the areas can be used for

262

quantitation.

263

3. Results and discussion.

264

3.1 Mass spectra of picrocrocin and non-isotopic isobaric internal standard (IIS).

265

Precise quantification of small organic molecules using stable isotope labeled (2H,

266

compounds as internal standards (IS) has been widely implemented in mass spectrometry. The use

267

of a non-isotopic isobaric internal standards (IIS) represents the best choice in the absence of

268

chromatography steps, and it can be considered a test case for MALDI MS/MS quantification

269

performed on crude extracts. Compounds 1 and 2 isobars with Δm/z = 0.002, thus not resolvable by

270

TOF. The powdered saffron samples were spiked with an aliquot of 2 and directly analyzed by

271

MALDI-MS and MS/MS. Fragmentation of ions [1]+ (m/z 369.13) and [2]+ (m/z 369.13) was

272

studied by MS/MS, before to carry out the quantitative experiments. Fig. 1C displays the MALDI

273

MS/MS spectrum of [2]+. The fragmentation channels of the major curcuminoids have extensively

274

been studied, and elucidated by Fourier transform ion cyclotron resonance (FT-ICR) mass

275

spectrometry (Jiang, et al. 2006).Thus, the observed products ions of m/z 351.1, 299.1, 285.1,

276

259.1, 245.1, 177.1, and 175.1 were easily assigned to [C21H19O5]+, [C18H19O4]+, [C17H17O4]+,

277

[C15H15O4]+, [C14H13O4]+, [C10H9O3]+ and [C11H11O2]+, respectively (Fig. 1C). The most intense

278

product ion of m/z 177.06, arising from the 3,4-bond cleavage and neutral loss of one 1-aryl moiety

279

in 2, was chosen as the general marker ion for quantitation of 2. Fragmentation across the rings

280

induced by MS/MS (1 kV) was valuable for the characterization of the trimethylcyclohexene

281

13C,

and

15N)

scaffold, and the sugar moiety of 1. Cleavage of the glycosidic O-linkage, with concomitant H12

282

rearrangement, led to the elimination of 162 Da (hexose), yielding the Y0 ion of m/z 207.1 as the -

283

base peak of the spectrum (Fig. 1A). Losses of 15 Da (CH3), 18 Da (H2O), and 30 Da (CH2O) from

284

the parent ion allowed to confirm the presence of the formyl (COH) and methyl groups on the

285

trimethylcyclohexene skeleton. It turned out that transitions 369207 for 1, and 369177 for 2

286

can successfully be used in a MALDI MS/MS based strategy aiming saffron authentication (Fig.

287

1B).

288

3.2 Mass spectrometry and the selection of blank matrix for dilution series.

289

A series of experiments was planned in order to determine the possible blank matrix effects on

290

specific ion signals, and the reliability of the whole extract from standard samples used as standard

291

solutions (Table 1S, Set II). Then, the eight standard solutions were analyzed. All metabolites of

292

interest were detectable as singly charged cation adducts. Only C-2÷C-4 and 1 signals were used for

293

data analysis. The mole fraction of crocin C-4 (MFC-4) was determined for all standard solutions 1-

294

8. MFC-4 was calculated as MFC-4 = ICAC-4/∑ICAS, where ICAC-4 and ICAS are the isotope cluster

295

areas of specific ion signals from saffron. The linear range was assessed by plotting MFC-4 values

296

versus concentrations of standard samples in the range of 3500-5000 mg/L. MFC-4 calculated by

297

linear regression was 0.4369 ± 0.0001 (RSD% 0.0278) suggesting the absence of interferences

298

along the dilution series.

299

The Set I of standard solutions (Table 1S) was prepared and analyzed by MALDI MS/MS. The

300

internal standard 2 yielded good linearity (R2 = 0.994), with confidence intervals at 95% for

301

intercept. The ANOVA regression model was significant, F(1, 72) = 11746.8742, p < 0.001 (slope =

302

0.0163 ± 0.0001, intercept = 48,8631 ± 0.6574). The LOD was calculated with the signal of a blank

303

“authentic saffron sample” plus 3 times the standard deviation of blank sample. The LOD value was

304

47.63 mg/L (0.95 %). The maximum adulteration percentage (99.0 %) was estimated by

305

interpolating LOD for IIS in linear regression. The LOQ value was 56.53 mg/L. The repeatability 13

306

(RSD%) calculated on spiked samples was found to be lower than 1%. Accuracy of the method was

307

determined by using fortified samples (SP1- SP3), prepared by adding known quantities of the

308

foreign matter (citrus leaves, calendula) to authentic saffron samples (Table 1A). Quantitative

309

recovery and high reproducibility (*RSD%) highlighted the reliability of the method, suggesting

310

that the developed approach is suitable for a rapid screening of saffron. The calibration curve and

311

SP1-SP3 sample were prepared using the same authentic saffron sample therefore the calculation of

312

the adulterant concentration is accurate. Comparison between the determined LOD with its value

313

obtained as previously reported by employing HPLC with PDA and/or ESI-MS detection, namely

314

5% (w/w) for calendula or safflower and 2% (w/w) for turmeric, implied that the proposed approach

315

enables detection of plant-derived adulterants at lower levels in saffron (Sabatino, et al., 2011). The

316

recently published and more sensitive methods combining LC and MS to assess the authenticity of

317

saffron, through the analysis of a group of kaempferol derivatives and geniposide, have LOD of 0.2-

318

2% and 10 ng/mL, respectively (Guijarro-Díez, et al., 2017a, Guijarro-Díez, et al., 2017b). However,

319

the accuracy of the most sensitive method (LOD 10 ng/mL), assessed by evaluating the recovery

320

obtained for geniposide in spiked saffron sample with 1 g/mL of geniposide, was 89 ± 14%

321

(Guijarro-Díez, et al., 2017a).

322 323 324

Table 1. (A) Reproducibility (*RSD%), and accuracy for spiked Samples SP1-SP3; (B) calculated concentrations (mg/L and %) of the adulterant in Samples S1-S5; (C) calculated mole ratios of Gardenia in Samples S1-S5. (A) Sample

Adulterant amount

*RSD%

Accuracy (%)

SP1

603.64 ± 5.0

0.83

96.6

SP2

546.76 ± 2,5

0.46

98.8

SP3

462.09 ± 3.9

0.85

98.6

(B)

MS/MS experiments

(C) MS experiments

Sample

Adulterant amount

*RSD%

Adulteration (%)

Mole ratio (%)

S1

1250.92 ± 13.15

1.05

25.01 ± 0.13

27.01 ±0.40

S2

500.45 ± 15.67

3.13

10.00 ± 0.31

10.59 ± 0.30

14

325 326

S3

806.52 ± 8.18

1.20

16.13 ±0.19

17.32 ± 0.42

S4

635.08 ± 12.93

2.04

12.70 ± 0.26

13.51 ± 0.31

S5

1071.0 ± 17.25

1.61

21.42 ± 0.34

23.05 ± 0.44

*The reproducibility of the measurements was determined by extracting the same samples in triplicate over a period of 1 week. Adulterant amounts are expressed as mg/kg of dry material.

327

328

The relatively poor temporal resolution, due to sampling times up to 20 minutes, is the major limit.

329

Throughput of the LC MS systems is restricted to a limited number of samples. The MALDI-MS

330

method here described showed LOD comparable to that of LC/MS methods. Since chromatographic

331

separations or desalting are not required, in our case the recovery is quantitative. Moreover, the

332

MALDI MS/MS method enables a high sample throughput because of the minimal sample

333

preparation and the very short measuring time per sample. The assessment of the specificity of the

334

method requires the m/z values and the monitored transitions to be free of interferences from the

335

internal standard and endogenous molecules. No interferences were observed for both endogenous

336

molecule and the IIS during the analysis of authentic and spiked samples. Thus, the developed

337

approach can be used to determine saffron adulteration by turmeric in unknown samples because

338

IIS is a specific marker. A limit of the developed approach could be represented by the selection of

339

picrocrocin as general marker of saffron because of its high variability (0.8% ± 26.6% on a dry

340

basis) (Alonso, et al., 2001; Iborra, et al., 1992; Sánchez, et al., 2008). However, the MALDI

341

MS/MS quantification method was further tested using five powdered saffron sachets purchased

342

from market. The selected five samples (S1-S5), characterized by different color intensities and

343

grain sizes, were processed as described above. The capability of the MS approach in specifically

344

and selectively monitoring analytes can be used to establish the adulterant identity. The

345

identification of the adulterant present in S1-S5 extracts, represented the next step of the present

346

study and it was performed by the adulterant marker identification based on accurate mass, isotopic

347

pattern, MS/MS analysis and literature.

15

348

Therefore, MS and MS/MS experiments on extracts from saffron samples (S1-S5), in the absence of

349

IIS, were included in the planned workflow (Fig. 2, panel F). MS/MS experiments on the ion signal

350

of m/z 369 from S1-S5 extracts allowed to establish the absence of turmeric as the adulterant.

351

Turmeric is one of the most frequently reported plant to adulterate saffron. It comprises three

352

curcuminoids: curcumin, demethoxycurcumin, and bisdemethoxycurcumin, with curcumin the

353

major component (0.3–5.4 %) of raw turmeric. MS/MS spectrum of the ion signal of m/z 369 from

354

S1-S5 extracts did not show daughter ions attributable to the curcumin structure (i.e. m/z 177).

355

Figure 3 to be inserted here

356

MS spectra of samples S1-S5 showed similar molecular profiles, suggesting that all samples were

357

intentionally blended with the same adulterant. Fig. 3 (panel B) displays partial spectra of

358

suspicious samples S1-S5. The structures of crocins C-1÷C-6 from saffron have been elucidated

359

(Carmona, et al., 2006). The MALDI-MS spectrum (Fig. 3, panel B) displayed ion signals of m/z

360

691.27, 837.32, 853.33, 999.38 and 1015.38 corresponding to cation adducts of crocins (C-2÷C-4),

361

that were unequivocally determined by the accurate mass of each peak and MS/MS analysis.

362

Spectra of samples S1-S5 showed an extra specific m/z spacing patterns of glycoforms (n162)

363

within 0.5-1 kDa. A difference of 3.94 mass units between the peaks at m/z 811.39

364

([C42H60NaO14]+) and 815.33 ([C38H55O19]+) associated to a more complex isotopic pattern of peaks

365

suggested that the extracts contained a mixture of at least two components (Fig.3; Fig. 4, panel A).

366

Figure 4 to be inserted here

367

The ion signal of m/z 815.33 ([C38H55O19]+) was assigned to crocin C-3, while that of m/z 811.39

368

([C42H60NaO14]+) could be a glycosylated compound arising from the adulterant. Among the known

369

saffron adulterants, only gardenia contains glycosyl ester of crocetin and gardecin. The main

370

structural difference between gardecin and crocin is the presence of ,-epoxyketone group 16

371

substituting one ester group by a ketoneic bond (Chen, et al. 2008). This structural feature leads to a

372

mass difference of 3.9457 u. Therefore the ion signals of m/z 811.39 ([C42H60NaO14]+) can be

373

assigned to gardecin-2 (G-2). The theoretical calculated isotopic distribution, and the sum of the

374

isotopic distributions of both crocins confirmed the presence of both compounds in the examined

375

mixture (Fig. 3, panel A). The MS/MS experiments of the ions of m/z 815.33 and m/z 811.39

376

validated the structure of crocins C-3 and G-2, respectively (Fig. 4, B-C) highlighting the presence

377

of gardenia in the extract. Consequently, the ions of m/z 649.33 ([C36H50NaO9]+), 811.39

378

([C42H60NaO14]+) and 973.45 (C48H72NaO19]+), were assigned to sodium adducts of gardecins G-1

379

(Hex-G), G-2 (Hex-Hex-G), G-3 (Hex-Hex-Hex-G), respectively (Fig.3). The ion of m/z 487.29

380

([C30H40NaO4]+) was assigned to the apocarotenoid sodium adduct. Fig. 5 displays MS/MS spectra

381

of G-1 and its aglycone. The product ions of m/z 207.1 ([C11H20NaO2]+) and 326.2 ([C21H26O3] •+)

382

were diagnostic for the ,-epoxyketone group and the apocarotenoid counterpart (Fig. 5A). The

383

MS/MS spectrum of ion of m/z 649.33 (Fig. 5B) displayed the neutral loss of 162u (Y0, m/z 487)

384

and the formation of product ion m/z 326.2 ([C21H26O3] •+) confirming that ion of m/z 649.33 is

385

monoglycosyl ester of gardecin (G-1).

386

Figure 5 to be inserted here

387

The detection of gardecin (G) and its derivatives (G-1÷G-3) in all samples S1-S5 indicated

388

adulteration by Gardenia jasminoides Ellis.

389

The UV–Vis spectra of crocins are characterized by some absorption bands in the range 250–470

390

nm. A band between 400 and 470 nm (max = 440 nm) is typical of all trans-carotenoids

391

(Cossignani, et al., 2014). Gardecin is characterized by a bathochromic shifts (max = 450 nm),

392

with respect to crocin-1 (max = 439 nm) (Chen, et al., 2008). Therefore, UV–vis analysis of a

393

sample of saffron adulterated with Gardenia jasminoides Ellis might underestimate the saffron 17

394

fraud occurrence due to the addition of this adulterant. The standard ISO 3632-2 UV–vis

395

spectrophotometric method, recommended for grading saffron, may fail to reveal saffron

396

adulteration with gardenia, in particular when the adulterant is added in quantities less than 20%

397

(w/w). On the contrary, the method based on MALDI-MS and MS/MS experiments enables the

398

semi-quantitative assay of saffron adulteration, and the identification and structural characterization

399

of specific markers of the adulterant.

400

The adulteration of samples S1-S5 was assessed by calculating the mole ratio of gardenia (MRG),

401

and from the calibration curve. Mole ratio was obtained by the equation MRG=

402

∑ICAG/(∑ICAS+∑ICAG), where ICAG and ICAS are the isotope cluster areas of specific ion signals

403

from gardenia and saffron, respectively. In all samples, only the ion signals of picrocrocin, crocins

404

C-2÷C-4, and gardecins G-1÷G-3 were used for data analysis (Table 1C).

405

The calibration curve was calculated by the specific authentic saffron sample. However, attention

406

must be given to the confidence limits of the regression line. In fact, its use may not be appropriate

407

to calculate the absolute adulterant concentrations in other commercial and unknown saffron

408

samples, due to the variability of picrocrocin contents. Notwithstanding, the developed method can

409

be appropriate for the semi-quantitative measurement of the adulterant in samples (Table 1B). The

410

adulteration assays carried out by the use of mole ratio and linear regression are in accordance,

411

confirming the presence of gardenia in the commercial samples.

412

4. Conclusions

413

The performed studies demonstrate that the quantification of turmeric as adulterant in saffron can be

414

achieved by MALDI MS and MS/MS following the optimized sample preparation protocol. The

415

established quantitation strategy yielded excellent linearity, precision, and accuracy. The LOD and

416

LOQ were 47.63 ppm and 56.53 ppm of dry material, respectively for curcumin where these limits 18

417

are comparable to or even better than those reported when other methods were used (Sabatino, et

418

al., 2011). The method presented here displays the following advantages: first, employing the

419

method as described significantly reduces the analysis time. Second, a simple extraction method

420

was applied to obtain the markers of saffron and adulterant and only about 1 μL of sample is needed

421

per data point. The calibration curve was calculated by the specific authentic saffron sample.

422

Notwithstanding, the developed method can be appropriate for the semi-quantitative measurement

423

of the adulterant in samples.

424

425

426

Acknowledgements

427

Ministero dell'Istruzione, dell'Università e della Ricerca (MIUR) is thanked for financial support

428

through Project PRIN 2015 (Progetti di Rilevante Interesse Nazionale, Prot. 201545245K_002).

429

430

431

432

References

433

Aiello, D., Casadonte, F., Terracciano, R., Damiano, R., Savino, R., Sindona, G., Napoli,

434

A.(2016) Targeted proteomic approach in prostatic tissue: a panel of potential biomarkers for

435

cancer detection. Oncoscience 3, 220–241.

436

Aiello, D., Giambona, A., Leto, F., Passarello, C., Damiani, G., Maggio, A., Siciliano, C.,

437

Napoli, A. (2018) Human coelomic fluid investigation: A MS-based analytical approach to

438

prenatal screening. Scientific Reports 8,10973, DOI:10.1038/s41598-018-29384-9. 19

439

Aiello, D., Materazzi, S., Risoluti, R., Thangavel, H., Di Donna, L., Mazzotti, F., Casadonte, F.,

440

Siciliano, C., Sindona, G., Napoli, A. (2015) A major allergen in rainbow trout (Oncorhynchus

441

mykiss): complete sequences of parvalbumin by MALDI tandem mass spectrometry. Molecular

442

Biosystems, 11, 2373-2382.

443

Alonso, G. L., Salinas, M. R, Garijo J.(1998) Method to Determine the Authenticity of Aroma

444

of Saffron (Crocus sativus L.). Journal of Food Protection 61, 1525–1528.

445

Alonso, G. L., Salinas, M.R., Garijo, J., Sánchez, M.A. (2001) Composition of crocins and

446

picrocrocin from Spanish saffron (Crocus sativus L.). Journal of Food Quality 24, 219–233.

447

Carmona, M., Zalacain, A., Sanchez, A.M.,. Novella, J. L., Alonso, G.L (2006) Crocetin Esters,

448

Picrocrocin and Its Related Compounds Present in Crocus sativus Stigmas and Gardenia

449

jasminoides Fruits. Tentative Identification of Seven New Compounds by LC-ESI-MS. J.

450

Agric. Food 54, 973-979.

451

Chen, Y., Cai, L., Zhao, C., Xu, H., Cao, C., Liu, Y., Ji, L., Yin, H., Chen, C., Zhang, H. (2008)

452

Spectroscopic, stability and radical-scavenging properties of a novel pigment from gardenia

453

Food Chem. 109, 269–72.

454

Cossignani, L., Urbani, E., Simonetti, M. S., Maurizi, A. C., Chiesi, F. (2014) Characterisation

455

of secondary metabolites in saffron from central Italy (Cascia, Umbria). Blasi Food Chem.143,

456

446-451.

457

Di Donna, L., Benabdelkamel, H., Taverna, D., Indelicato, S., Aiello, D., Napoli, A., Sindona,

458

G., Mazzotti, F. (2015) Determination of ketosteroid hormones in meat by liquid

459

chromatography tandem mass spectrometry and derivatization chemistry. Analytical and

460

Bioanalytical Chemistry 407, 5835-5842.

461

De Marco, R., Di Gioia, M.L., Leggio, A., Liguori, A., Perri, F., Siciliano, C., Viscomi, M.C.A

462

new non-natural arginine-like amino acid derivative with a sulfamoyl group in the side-chain.

463

(2010) Amino Acids, 38 , 691-700. 20

464

García-Rodríguez, M. V., Serrano-Díaz, J., Tarantilis, P.A., López-Córcoles, H., Carmona, M.,

465

Alonso, G. L. (2014) Determination of Saffron Quality by High-Performance Liquid

466

Chromatography. J. Agric. Food Chem. 62, 8068–8074.

467

Guijarro-Díez, M., Castro-Puyana, M., Crego, A. L., Marina, M. L. (2017) Journal of Food

468

Composition and Analysis 55, 30-37.

469

Guijarro-Díez, M., Castro-Puyana, M., Crego, A. L., Marina, M. L. (2017) Detection of saffron

470

adulteration with gardenia extracts through the determination of geniposide by liquid

471

chromatography–mass spectrometry. Food Chem. 228, 403–410

472

Herrero, M., Simò, C., García-Cañas, V., Ibáñez, E., Cifuentes, A. (2012) Foodomics: MS-

473

based strategies in modern food science and nutrition. Mass Spectrom. Rev. 31, 49–69.

474

Iborra, J. L., Castellar, M. R., Canovas, M., Manjón, A. (1992) A. TLC preparative purification

475

of picrocrocin, HTCC and crocin from saffron. J. Food Sci. 57, 714–731.

476

ISO 3632-1. (2011). Saffron (Crocus sativus L.): Specifications (2003 revised 2011).

477

Switzerland: Geneva, International Standards Organization.

478

ISO 3632-2. (2010). Saffron (Crocus sativus L.): Test Methods (2003 revised 2010).

479

Switzerland: Geneva, International Standards Organization.

480

Jiang, H., Somogyi, A., Jacobsen, N.E., Timmermann, B.N., Gang, D.R. (2006) Analysis of

481

curcuminoids by positive and negative electrospray ionization and tandem mass spectrometry.

482

Rapid Commun. Mass Spectrom. 20, 1001–1012.

483

Johnson, R. (2014). Food Fraud and economically motivated adulteration of food and food

484

ingredients.

485

https://www.fas.org/sgp/crs/misc/R43358.pdf.

486

Kanakis, C. D., Daferera, D. J., Tarantilis, P. A., Polissiou, M.G.(2004) Qualitative

487

Determination of Volatile Compounds and Quantitative Evaluation of Safranal and 4-Hydroxy-

Washington,

DC:

Congressional

Research

Service

Report.

21

488

2,6,6-trimethyl-1-cyclohexene-1-carboxaldehyde (HTCC) in Greek SaffronJ. Agric. Food

489

Chem. 52, 4515–4521.

490

Koulakiotis, N. S., Pittenauer, E., Halabalaki, M., Tsarbopoulos, A., Allmaier, G. (2012)

491

Comparison of different tandem mass spectrometric techniques (ESI‐IT, ESI‐ and

492

IP‐MALDI‐QRTOF and vMALDI‐TOF/RTOF) for the analysis of crocins and picrocrocin from

493

the stigmas of Crocus sativus L. Rapid Commun. Mass Spectrom. 26, 670–678

494

Maggi, L., Sanchez, A. M., Carmona, M., Kanakis, C.D., Anastasaki, E., Tarantilis P.A. (2011)

495

Rapid determination of safranal in the quality control of saffron spice (Crocus sativus L.)Food

496

Chem. 127, 369-373.

497

Mazzotti, F., Di Donna, L., Napoli, A., Aiello, D., Siciliano, C., Athanassopoulos, C. M. (2014)

498

N-hydroxysuccinimidyl p-methoxybenzoate as suitable derivative reagent for

499

isotopic dilution assay of biogenic amines in food. Journal of Mass Spectrometry, 49,

500

802–810.

501

Moore, J.C.; Spink, J.; Lipp M. (2012) Development and application of a database of food

502

ingredient fraud and economically motivated adulteration from 1980 to 2010. J. Food Sci. 77,

503

118–126.

504

Napoli, A., Aiello, D., Aiello, G., Cappello, M.S., Di Donna, L., Mazzotti, F., Materazzi, S.,

505

Fiorillo, M., Sindona, G. (2014) Mass Spectrometry-Based Proteomic Approach in Oenococcus

506

oeni Enological StarterJ. Prot Res. 13, 2856–2866.

507

Ordoudi, S.A., de los Mozos Pascual, M., Tsimidou, M.Z. (2014) On the quality control of

508

traded saffron by means of transmission Fourier-transform mid-infrared (FT-MIR) spectroscopy

509

and chemometrics. Food Chem. 150, 414–421.

22

510

Persike, M., Karas. M., (2009) Rapid simultaneous quantitative determination of different small

511

pharmaceutical drugs using a conventional matrix‐assisted laser desorption/ionization

512

time‐of‐flight mass spectrometry system. Rapid Commun Mass Spectrom. 23, 3555–3562.

513

Persike, M., Zimmermann, M., Klein, J., Karas, M. (2010) Quantitative Determination of

514

Acetylcholine and Choline in Microdialysis Samples by MALDI-TOF MS. Anal Chem 82, 922–

515

929.

516

Petrakis, E. A., Cagliani, L. R., Polissiou, M. G., Consonni, R. (2015) Evaluation of saffron

517

(Crocus sativus L.) adulteration with plant adulterants by 1H NMR metabolite fingerprinting.

518

Food Chem.173, 890-896.

519

Pfister, S., Meyer, P., Steck, A., Pfander, H. (1996) Isolation and structure elucidation of

520

carotenoid glycosyl esters in gardenia fruits (Gardenia jasminoides Ellis) and saffron (Crocus

521

sativus Linne). J. Agric. Food Chem. 56, 257–259.

522

Rubert, J., Lacina, O., Zachariasova, M., Hajslova, J. (2016) Saffron authentication based on

523

liquid chromatography high resolution tandem mass spectrometry and multivariate data

524

analysis. Food Chem. 204, 201–209.

525

Sabatino, L., Scordino, M., Gargano, M., Belligno, A., Traulo, P., Gagliano G. (2011)

526

PLC/PDA/ESI-MS evaluation of saffron (Crocus sativus L.) adulteration. Nat. Prod. Commun.

527

6, 1873–1876.

528

Sánchez, A. M., Carmona, M., Zalacain, A., Carot, J.M., Jabaloyes, J. M., Alonso, G.L. (2008)

529

Rapid Determination of Crocetin Esters and Picrocrocin from Saffron Spice (Crocus sativus L.)

530

Using UV–Visible Spectrophotometry for Quality Control. J. Agric. Food Chem. 56, 3167–

531

3175.

532

Sereshti, H., Poursorkh, Z., Aliakbarzadeh, G., Zarre, S. (2018) Quality control of saffron and

533

evaluation of potential adulteration by means of thin layer chromatography-image analysis and

534

chemometrics methods. Food Control 90, 48-57 23

535

Tarantilis, P.A., Polissiou, M., Manfait, M. (1994) Separation of picrocrocin, cis-trans-crocins

536

and safranal of saffron using high-performance liquid chromatography with photodiode-array

537

detection. J. Chromatogr. A 664, 55–61.

538

van Kampen, J. J., Burgers, P. C., de Groot, R., Gruters, R. A., Luider, T. M. (2011) Biomedical

539

application of MALDI mass spectrometry for small‐molecule analysis. Mass Spectrom Rev, 30,

540

101–120.

541

Winterhalter, P., Straubinger, M. (2000) Saffron-renewed interest in an ancient spice. Food Rev.

542

Int. 16, 39–59.

543

Zalacain, A., Ordoudi, S. A., Díaz-Plaza, E. M., Carmona, M., Blazquez, I., Tsimidou M. Z.

544

(2005) Near-Infrared Spectroscopy in Saffron Quality Control:  Determination of Chemical

545

Composition and Geographical Origin. J. Agric. Food Chem. 53, 9337-9341.

546

24

547

Figure Captions

548

Fig. 1. MALDI MS/MS of ion of m/z 369.1 from (A) authentic saffron sample, (B) authentic

549

saffron and curcumin mixture, (C) curcumin.

550

Fig. 2. Pictorial description of the developed approach.

551

Fig. 3. Panel A displays the structure and m/z value of sodium adduct of gardecin and its

552

derivatives, respectively. Panel B displays partial MS spectra of authentic and suspicious saffron

553

samples S1-S5.

554

Fig.4. (A) Isotopic distribution of ion of m/z 811.39 (G-2) and 815.33 (C-3); (B) MS/MS spectrum

555

of crocin-3 (m/z 815.33); (C) MS/MS spectrum of Gardecin-2 (m/z 811.39).

556

Fig. 5. MALDI MS/MS spectra of (A) m/z 487.29, R = H, and (B) m/z 649.33, R=Hex.

n=2

and

n=1

are crocetin and gardecin derivatives, respectively.

25

557

26

558

27

559

28

560

29

561

562

Highlights

563

564

A MALDI-MS/MS quantitative method was developed for saffron authenticity.

565

The method was sensitive and fast, not requiring chemical manipulation of samples.

566

Picrocrocin was chosen as the saffron authenticity biomarker.

567

Percentages of adulteration in commercial saffron were evaluated by using curcumin.

568

569

30