1

SUPPLEMENTARY MATERIAL

Chemometric analysis of minerals and trace elements in Sicilian wines from two

different grape cultivars

Angela Giorgia Potortί, Vincenzo Lo Turco*, Marcello Saitta, Giuseppe Daniel Bua, Alessia

Tropea, Giacomo Dugo, Giuseppa Di Bella.

Dipartimento di Scienze Biomediche, Odontoiatriche e delle Immagini Morfologiche e Funzionali

(BIOMORF), Università di Messina, Viale Annunziata – Polo Universitario 98168 Messina, Italy.

*Address correspondence to Vincenzo Lo Turco; Dipartimento di Scienze Biomediche,

Odontoiatriche e delle Immagini Morfologiche e Funzionali (BIOMORF), Università di Messina,

Viale Annunziata – Polo Universitario 98168 Messina, Italy;

phone: +39 090 3503 997;

e-mail: vloturco@unime.it.

Abstract

Chemometric analysis are used for food authenticity evaluation, correlating botanical and

geographical origins with food chemical composition.

This research was carried out in order to proved that it is possible linked red wines to Nero d'Avola

and Syrah cultivars of Vitis vinifera according to their mineral content, while the values of the

physical and chemical parameters do not affect relevantly this discrimination.

The levels of mineral elements were determined by ICP-OES and ICP-MS.

Samples from cv Nero d’Avola had the highest content of Zn, Cr, Ni, As and Cd, whereas the

highest mineral concentration in cv Syrah samples was represented by K, Mg, Cu, and Sb. The

research highlights that it is possible linked red wines to Nero d'Avola and Syrah cultivars of Vitis

vinifera according to their mineral contents, adding knowledge to the determination studies of the

wine botanical origin.

Keywords: Red wines; Minerals; Trace elements; Chemometric analysis; Botanical discrimination.

2

Experimental

Reagents and materials

High purity water with resistivity of 10 MΩ cm (J.T. Baker, Milan, Italy), was used throughout.

Concentrated (65%, v/v) nitric acid trace metal analysis grade (J.T. Baker, Milan, Italy) was used

together with concentrated (30%, v/v) hydrogen peroxide (J.T. Baker, Milan, Italy) for samples

digestion. The first one was also employed for cleaning glassware.

Single element standards (1000 mgL−1

in 2% nitric acid) were purchased from Fluka (Milan, Italy)

and from Merck (Darmstadt, Germany) and were mixed to prepare a multi-element standard

solution that has been subsequently diluted for calibration analysis.

To correct instrumental drift and variations due to the matrix, standard solutions of 45

Sc, 103

Rh and

209Bi (1000 mgL

−1 in 2% nitric acid) were purchased from Fluka (Milan, Italy) and were used as

on-line internal standards (at level of 1.5 mgL-1

).

To verify the digestion of sample and to correct the volumetric changes, standard solution of Re at

1000 mgL-1

in 2% nitric acid was acquired by Fluka (Milan, Italy) and was used as preparation

standard (at level of 0.5 mgL-1

).

A solution containing 1 gL-1

of 7Li,

59Co,

80Y and

205Tl in 2% HNO3 was obtained from Agilent

(Santa Clara, CA) and was used to tune the ICP-MS instrument.

A diagnostic standard solution containing 1000 mgL-1

of Ba, Mg and Zn in 5% HNO3 (JYICP-

DIAG) was obtained from Horiba Jobin Yvon (Longjumeau, France) and used for the periodic

check of the ICP-OES instrument.

Argon (N 5.0) of 99.9990% purity and helium (N 5.5) of 99.9995% purity were supplied by Rivoira

gases (Milan, Italy).

All reagents used for enological parameters determination, provided by Sigma Aldrich, were

analytical grade.

Wine samples

The analyses were carried out on 39 sicilian red wines, obtained from grapes of 100 % cv Nero

d’Avola and 34 from 100% cv Syrah, and cultivated in the same geographic areas in province of

Syracuse. All samples were obtained directly from producers and were from 2015 vintage. The wine

glass bottles (750 mL) were stored in the dark at 2°C and opened before analysis.

3

Analytical procedure for minerals and trace elements determination

From the freshly opened bottles, about 1 mL of each wine was transferred and accurately weighed

into acid-prewashed PTFE vessels; it was added with internal Re standard and then digested with 6

mL of HNO3 (69%, v/v) and 2 mL of H2O2 (30%, v/v) in a microwave digestion system Ethos 1

(Milestone, Bergamo, Italy) equipped with sensors for temperature and pressure control.

Instrumental parameters and settings were: 10 min for 1000W up to 200°C, 10 min for 1000W at

200°C. Allowed to cool, each sample was made up to volume of 10 mL with HNO3 (2%, v/v). Each

sample was digested in triplicate.

The K, Ca, Mg, Na, Zn, Fe, Mn and Cu determination was carried out by Horiba Jobin Yvon

ULTIMA 2 (HORIBA Scientific, Longjumeau, France) ICP-OES spectrometer, equipped with a

glass concentric pneumatic nebulizer (i.d. 0.3mm) coupling with a quartz cyclonic type spray

chamber (50mL).

For Cr, Pb, Ni, Co, Se, As, Cd and Sb measurements an Agilent 7500cx (Agilent Technologies,

Santa Clara, CA) ICP-MS spectrometer, equipped with a MicroMist glass concentric pneumatic

nebulizer coupling with a cooled Scott double pass type spray chamber made of quartz, was used.

To minimize polyatomic interferences resulting from plasma and matrix, an octopole collision

system with 4 mLmin-1

helium as collision gas and kinetic energy discrimination mode was used

(collision mode) for almost all the elements.

The instrument operating parameters for ICP-OES and ICP-MS analyses are presented in Table S1.

According to literature, ethanol may affects the quantification of the different elements in wine

samples since its content could influence the transport properties towards atomization devices of the

instrument as well as the viscosity and the density of the samples (Aceto et al. 2002). Thus, the

possible interference of alcohol on elemental measurements was corrected by addiction of 1.3%

ethanol to all standard solutions used for calibration since wine samples were diluted 10 times

before analysis (Rodriguez et al. 2011).

The evaluation of the linearity was based on the 6 standard solutions injections. Each solution was

injected three times (n=3). The instrumental detection limits (LODs) and quantification (LOQs)

were experimentally calculated as 3.3σ/S and 10σ/S, respectively, where σ is the standard deviation

of the response of six blanks and S is the slope of the calibration curve (EURACHEM 2000).

A lab-made wine containing 5 gL-1

tartaric acid and 13% ethanol in water was used as blank

solution. It was digested as describe above and it was run with each batch of wines. Moreover, for

recovery studies, 18 spiked lab-made wine solutions were prepared: 9 were used for ICP-OES

analysis (3 at level of 50 mgL-1

, 3 at level of 100 mgL-1

and 3 at level of 300 mgL-1

) and 9 were

4

used for ICP-MS analysis (3 at level of 10 µgL-1

, 3 at level of 20 µgL-1

and 3 at level of 50 µgL-1

).

Each solution was analyzed in triplicate. For repeatability estimation and intermediate precision,

each spiked level was prepared and analyzed in 12 replicates in the same batch and in 24 replicates

in different days.

Enological parameters determination

The main enological parameters, among which alcohol content, pH, total acidity, volatile acidity,

malic acid, SO2 and total SO2 contents, were determined following the procedures specified in detail

in EC Regulation 2676/90 (1990).

Finally, according to Ribéreau Gayon and Stonestreet (1965), anthocyanins quantification was

carried out in wine samples.

Validation of ICP-MS and ICP-OES analysis

Method linearity, sensitivity, accuracy, precision and repeatability are reported in Table S2. Results

showed that the adopted procedures were suitable for the research. Indeed, good linearity was

observed in each investigate concentration range with R2 ≥ 0.99943. Instrumental LOD values

ranged from 0.003 to 0.750 mgL-1

for ICP-OES analysis and from 0.009 to 0.030 µgL-1

for ICP-

MS analysis, while instrumental LOQ values ranged from 0.010 to 2.50 mgL-1

and from 0.029 to

0.100 µgL-1

, respectively. Thus the analytical limit of quantification for elements analyzed by ICP-

OES were between 0.1 to 25 mgL-1

(Cu and K, respectively), while for elements analyzed by ICP-

MS varied from 0.290 (value determined for Sb) to 1.000 µgL-1

(value determined for Se and As).

The recovery for all elements was always within the interval of 75.2-126.6%. The repeatability

RSD% was lower or equal to 5.1%, while for intermediate precision it was lower or equal to 9.5%.

Statistical analysis

The SPSS 13.0 statistical software package for Windows (SPSS Inc., Chicago, IL, USA) was used

for all statistical calculations.

Statistical methods were conducted on starting multivariate matrix where variables were the

concentrations of 24 detected parameters (8 were the enological parameters and 16 were the

concentrations of minerals and trace elements) and the cases were the 73 analyzed wine samples.

Data below LOQ were replaced with the LOD/2 values and all concentrations were loge-

transformed to reduce the effect of outliers on skewing the data distribution and to bring the

concentrations of element within the same range (Škrbić et al. 2010).

The data were subdivided in two groups, according to the cultivar of Vitis vinifera: the first one (39

samples) consisting of wines from Nero d’Avola grapes, and the second one (34 samples)

represented by wines from Syrah grapes.

5

Initially, the non-parametric Mann–Whitney U test was applied to study the significances of

differences. Successively, the data set was normalized and Principal Component Analysis (PCA)

was performed to differentiate samples belonging to the two red wine types based on the enological

parameters values and the concentrations of minerals and trace elements. In addition to PCA, Linear

Discriminant Analysis (LDA) in the stepwise mode was carried out to discriminate among wine

cultivars in according to F-value.

References

Aceto M, Abollino O, Bruzzoniti MC, Mentasti E, Sarzanini C and Malandrino M. 2002.

Determination of metals in wine with atomic spectroscopy (flame-AAS, GF-AAS and ICP-AES): a

review. Food Addit Contam 19:126-133.

D.M. 29 December 1986. Decree of the Minister of Agriculture and Forests, published on

Official Gazette, No. 13, January 17th 1987.

EC Regulation 2676/90 (17 Sept. 1990). Determining Community methods for the analysis of

wines. Off J Eur Communities 1990, No. 272 (Oct 3), 1-192.

EURACHEM 2000, Guide. (2nd

Ed.). Editors: S L R Ellison (LGC, UK), M Rosslein (EMPA,

Switzerland), A Williams (UK).

European Directive EC/1881/2006/. Commission Regulation, No. 1881, December 19th

2006.

OIV (Organisation Internationale de la Vigne et du Vin). 2011. Compendium of international

methods of wine and must analysis.

Ribéreau GP and Stonestreet E. 1965. Le dosage des anthocyanes dans le vin rouge. Bull Soc

Chim Fr 9:2649-2652.

Rodriguez SM, Otero M, Alves AA, Coimbra J, Coimbra MA, Pereira E and Duarte AC. 2011.

Elemental analysis for categorization of wines and authentication of their certified brand of origin. J

Food Comp Anal 24:548-562.

Škrbić B, Szyrwińska K, Đurišić-Mladenović N, Nowicki P and Lulek J. 2010. Principal

component analysis of indicator PCB profiles in breast milk from Poland. Environ Int 36:862-872.

6

Table S1 – Instrument operating parameters for ICP-OES and ICP-MS analyses.

ICP-OES analysis

Rf power 1000W

Auxiliary/nebulizer/plasma Argon flow rate 0.2/1/12 L·min-1

Nebulization pressure 2.98 bar

Nebulizer pump 20 rpm

Sample introduction flow rate 1 mL·min-1

Acquisition mode maxima

Integration time 2 sec for K, Ca, Mg and Na; 4 sec for Zn, Fe, Mn and Ca

Monitored isotopes and wavelengths (nm) K, 766.490; Ca, 393.366; Mg, 279.553; Na, 588.995; Zn, 213.856; Fe, 259.940; Mn, 257.110; Cu, 324.754

ICP-MS analysis

RF power 1500W

Plasma/auxiliary/carrier gas flow rate 15/0.9/1.1 Lmin-1

Helium collision gas flow rate 4 mLmin-1

Spray chamber temperature 2 °C

Sample depth 9 mm

Sample introduction flow rate 1 mLmin-1

Nebulizer pump 0.1rps

Extract lens 1 1.5 V

Octopole collision system setting He mode for Cr, Ni, Co, Se, As and Cd; No-gas mode for Pb and Sb

Monitored isotopes 52Cr, 59Co, 60Ni, 75As, 80Se, 114Cd, 121Sb and 208Pb

On-line internal standards 45Sc for Cr, Co, Ni, As and Se; 103Rh for Cd and Sb; 209Bi for Pb

Integration times 0.8 s/point and for Se; 0.5 s/point for As; 0.2 s/point for Cr, Co and Ni; 0.1 s/point for Cd, Sb and Pb

Point for mass 3 (3 replicates acquisitions)

7

Table S2 – Validation parameters for ICP-OES and ICP-MS analyses.

ICP-OES analysis

Element R2 LODi (mg/L) LOQi (mg/L) LOQa (mg/L) Accuracy (% ± RSD%, n=9) Repeatability (RSD%, n=12) Intermediate precision (RSD%, n=24)

Level I

(50 mgL-1)

Level II

(100 mgL-1)

Level III

(300 mgL-1)

Level I

(50 mgL-1)

Level II

(100 mgL-1)

Level III

(300 mgL-1)

Level I

(50 mgL-1)

Level II

(100 mgL-1)

Level III

(300 mgL-1)

K 0.99993 0.750 2.500 25.000 76.6 ± 2.6 79.4 ± 2.2 99.8 ± 2.8 3.3 3.5 4.0 6.6 7.0 7.6

Ca 0.99943 0.625 2.083 20.833 106.7 ± 3.1 110.8 ± 2.2 126.6 ± 2.5 3.4 2.7 2.5 5.6 4.2 4.0

Mg 0.99973 0.540 1.800 18.000 98.3 ± 4.2 102.4 ± 2.3 119.4 ± 2.4 3.6 2.7 2.6 4.5 3.0 2.7

Na 0.99969 0.300 1.000 10.000 100.8 ± 2.8 104.4 ± 2.2 120.0 ± 3.2 3.6 3.0 2.8 3.9 4.3 6.0

Zn 0.99991 0.035 0.117 1.167 80.5 ± 2.7 85.0 ± 2.1 105.2 ± 2.3 4.5 4.7 5.1 8.8 7.6 7.4

Fe 0.99947 0.029 0.097 0.967 75.2 ± 3.2 78.4 ± 2.5 96.4 ± 1.2 3.8 3.3 3.1 8.2 8.6 6.5

Mn 0.99993 0.018 0.060 0.600 98.2 ± 3.4 101.4 ± 2.8 123.5 ± 1.2 2.8 2.3 2.1 5.8 6.1 6.9

Cu 0.99961 0.003 0.010 0.100 86.4 ± 2.7 90.0 ± 3.0 105.6 ± 1.2 2.6 2.8 3.4 8.1 9.5 6.2

ICP-MS analysis

Element R2 LODi (mg/L) LOQi (mg/L) LOQa (mg/L) Accuracy (% ± RSD%, n=9) Repeatability (RSD%, n=12) Intermediate precision (RSD%, n=24)

Level I

(10 µgL-1)

Level II

(20 µgL-1)

Level III

(50 µgL-1)

Level I

(10 µgL-1)

Level II

(20 µgL-1)

Level III

(50 µgL-1)

Level I

(10 µgL-1)

Level II

(20 µgL-1)

Level III

(50 µgL-1)

Cr 0.99986 0.014 0.047 0.467 86.7 ± 5.6 90.5 ± 4.1 107.8 ± 4.1 4.6 3.9 3.8 8.0 8.3 9.8

Pb 0.99995 0.010 0.033 0.333 80.5 ± 3.2 84.0 ± 4.8 102.1 ± 2.1 2.9 3.1 3.8 10.3 8.7 8.3

Ni 0.99967 0.020 0.067 0.667 83.1 ± 2.2 86.7 ± 3.8 105.4 ± 2.1 3.5 3.7 3.9 7.5 7.8 9.2

Co 0.99995 0.020 0.067 0.667 88.7 ± 4.6 92.2 ± 3.3 110.1 ± 3.1 3.0 2.5 2.4 8.4 7.6 7.3

Se 0.99943 0.030 0.100 1.000 90.2 ± 4.1 93.6 ± 2.2 111.6 ± 3.1 0.9 1.1 1.8 8.2 7.0 6.6

As 0.99996 0.030 0.100 1.000 86.2 ± 2.8 89.8 ± 3.5 106.9 ± 5.1 3.1 2.9 2.7 4.0 4.2 5.6

Cd 0.99993 0.009 0.030 0.300 86.3 ± 1.8 90.0 ± 3.4 106.8 ± 2.1 4.2 3.6 3.4 5.6 6.0 7.7

Sb 0.99973 0.009 0.020 0.290 82.1 ± 7.6 85.6 ± 3.9 103.5 ± 2.1 1.6 1.1 1.0 7.9 7.1 6.8

R2, determination coefficient; LODi, instrumental detection limit; LOQi, instrumental quantification limit; LOQa, analytical quantification limit.

8

Table S3 – Contents and significant differences of the mineral and trace elements composition of wines from cv Nero d’Avola and cv Syrah grapes.

K (mg·L-1) Ca (mg·L-1) Mg (mg·L-1) Na (mg·L-1) Zn (mg·L-1) Fe (mg·L-1) Mn (mg·L-1) Cu (mg·L-1)

Nero d’Avola (n=39)

Min 506.613 68.412 100.434 10.298 3.801 1.006 0.461 0.121

Max 1519.418 216.643 151.667 204.659 11.793 9.446 1.866 1.525

Mean 887.522 124.126 122.234 41.833 6.607 4.661 1.307 0.289

S. D. 258.326 30.412 14.812 44.663 1.982 1.978 0.402 0.264

Syrah (n=34)

Min 1025.955 75.174 125.789 10.787 1.234 2.424 0.659 0.139

Max 2220.167 162.891 223.748 100.806 6.571 9.576 1.776 1.820

Mean 1450.543 110.113 158.865 37.187 3.571 4.706 1.204 0.744

S. D. 266.901 22.795 23.791 24.255 1.771 2.024 0.309 0.500

Mann Whitney U 1228.500 487.000 1207.000 702.000 196.000 625.000 534.500 1157.000

Wilcoxon W 1823.500 1082.000 1802.000 1297.000 791.000 1220.000 1129.500 1752.000

Asymp. Sign. 0.000 0.052 0.000 0.666 0.000 0.674 0.155 0.000

Cr (µg·L-1) Pb (µg·L-1) Ni (µg·L-1) Co (µg·L-1) Se (µg·L-1) As (µg·L-1) Cd (µg·L-1) Sb (µg·L-1)

Nero d’Avola (n=39)

Min 10.127 11.565 17.212 0.743 1.295 1.513 n.d. n.d.

Max 106.434 97.891 202.937 7.617 10.548 13.913 0.871 n.d.

Mean 35.099 30.855 66.801 4.193 3.841 4.018 0.515 _

S. D. 25.816 19.137 51.357 1.824 2.072 2.636 0.144 _

Syrah (n=34)

Min 10.450 10.283 10.004 1.070 1.333 0.961 n.d. n.d.

Max 29.675 59.952 52.853 6.505 6.116 3.268 0.780 0.936

Mean 18.042 30.053 28.271 3.447 3.437 1.764 0.348 0.572

S. D. 5.190 15.784 9.898 1.183 0.975 0.588 0.224 0.214

Mann Whitney U 334.000 678.000 256.000 502.000 655.000 151.000 459.500 916.500

Wilcoxon W 929.000 1273.000 851.000 1097.000 1250.000 746.000 1054.500 1511.500

Asymp. Sign. 0.000 0.868 0.000 0.075 0.930 0.000 0.019 0.000

Asymp. Sign. bold values indicate element concentrations significantly different at 95%.

9

Table S4 – Legal limits of Zn, Cu, Pb, As and Cd, and comparison with results obtained in this study.

Element Legal limits (mg/L) Concentrations in wine (mg/L) Samples exceeded legal limits (%)

Nero d'Avola Syrah Nero d'Avola Syrah

Mean value Max value Mean value Max value

Pb 0.2a 0.03 0.10 0.03 0.06 0 0

0.15b 0 0

Cu 1c,d 0.3 1.5 0.7 1.8 5 24

Zn 5c,d 6.6 11.8 3.6 6.6 74 24

As 0.2d 0.004 0.014 0.002 0.003 0 0

Cd 0.1d 0.0005 0.0009 0.0003 0.0008 0 0

a Regulation EC/1181/2006; b OIV, 2006; c Ministerial Decree of 29 December 1986; d OIV, 2011.

10

Table S5 – Values and significant differences of the enological parameters of wines from cv Nero d’Avola and cv Syrah grapes.

Alcoholic grade

(% vol)

pH

Total acidity

(g·L-1)

Volatile acidity

(g·L-1)

Malic acid

(g·L-1)

SO2

(mg·L-1)

Total SO2

(mg·L-1)

Anthocyanins

(mg·L-1)

Nero d’Avola (n=39)

Min 12.0 3.1 5.0 0.34 0.01 9.90 41.58 125.00

Max 14.4 3.7 6.8 0.91 1.49 38.38 131.30 1.076.00

Mean 13.0 3.5 5.6 0.52 0.29 26.13 85.36 222.20

S. D. 0.5 0.2 0.4 0.16 0.29 8.53 21.37 144.35

Syrah (n=34)

Min 11.6 3.1 4.2 0.35 0.02 9.80 41.16 120.00

Max 14.2 3.8 6.7 0.90 0.80 51.00 150.00 672.00

Mean 12.9 3.6 5.3 0.55 0.24 23.84 85.19 232.24

S. D. 0.67 0.1 0.5 0.14 0.22 10.48 24.46 104.09

Mann Whitney U 518.000 725.500 388.500 799.000 582.500 544.000 615.000 703.500

Wilcoxon W 1113.000 1320.500 983.500 1394.000 1177.500 1139.000 1210.000 1298.500

Asymp. Sign. 0.108 0.489 0.002 0.132 0.373 0.188 0.595 0.654

Asymp. Sign. bold values indicate element concentrations significantly different at 95%.

11

Table S6– The Recommended Dietary Allowance (RDA) of non toxic elements and RDA shares from consumption of Nero d’Avola and Syrah wines.

Non Toxic Elements RDA (mg×day-1)a Mean concentrations in wine (mg/L) % of RDA estimated by mean value

Nero d'Avola Syrah Nero d'Avola Syrah

Zn 10 6.607 3.571 14.5 7.9

Fe 14 4.661 4.706 7.3 7.4

Se 0.055 0.004 0.003 1.5 1.4

Cu 1 0.289 0.744 6.4 16.4

Cr 0.040 0.035 0.018 19.3 9.9

Mn 2 1.307 1.204 14.4 13.2

Ca 800 124.126 110.113 3.4 3.0

K 2000 887.522 124.126 9.8 1.4

Mg 375 122.234 158.865 7.2 9.3

Na 1500b 41.833 37.187 0.6 0.5

a Commission Directive 2008/100/EC; bAI (Adeguate intake) EFSA 2005.

12

Table S7 – Protection Limits (TDI, TWI, PTWI and BMDL01) of potentially toxic elements and Protection Limits shares from consumption of Nero d’Avola and Syrah wines.

Potentially Toxic Elements Mean concentrations in wine (mg/L) % of Protection limit estimated by mean

value Reference

Nero d'Avola Syrah Nero d'Avola Syrah

Pb PTWI (mg×kgb.w.-1×week-1) 0.025 0.031 0.030 3.2 3.1 EFSA, 2010

BMDL01(µg×kgb.w.-1×day-1) 1.5 7.5 7.3

As PTWI (mg×kgb.w.-1×week-1) 0.015 0.004 0.002 0.7 0.3 EFSA, 2009

BMDL01(µg×kgb.w.-1×day-1) 0.3 4.9 2.2

BMDL01(µg×kgb.w.-1×day-1) 8 0.2 0.1

Cd TWI (µg×kgb.w.-1×week-1) 2.5 0.515 0.348 0.5 0.4 EFSA, 2012

Ni TDI (µg×kgb.w.-1×day-1) 22 66.801 28.271 1.1 0.5 WHO, 2005

Sb TDI (µg×kgb.w.-1×day-1) 6 n.d. 0.572 n.d. 0.035 WHO, 2003

n.d., not determinable.

13

Figure S1 - 2D Scatterplot for the 73 wine samples categorized by cultivars of Vitis vinifera. Insert: loading plot for PC1 and PC2.