Differentiation of tomatoes based on isotopic, elemental and organic markers

Abstract

In this study, 41 tomato samples were investigated by means of stable isotope ratios (δ¹³C, δ¹⁸O and δ²H), elemental content, phenolic compounds and pesticides in order to classify them, according to growing conditions and geographical origin. Using investigated parameters, stepwise linear discriminant analysis was applied and the differences that occurred between tomato samples grown in greenhouses compared to those grown on field, and also between Romanian and abroad purchased samples were pointed out. It was shown that Ti, Ga, Te, δ²H and δ¹³C content were able to differentiate Romanian tomato samples from foreign samples, whereas Al, Sc, Se, Dy, Pb, δ¹⁸O, 4,4′-DDT could be used as markers for growing regime (open field vs. greenhouse). For the discrimination of different tomato varieties (six cherry samples and fourteen common sorts) grown in greenhouse, phenolic compounds of 20 samples were determined. In this regard, dihydroquercetin, caffeic acid, chlorogenic acid, rutin, rosmarinic acid, quercetin and naringin were the major phenolic compounds detected in our samples. The phenolic profile showed significant differences between cherry tomato and common tomato. The contents of the chlorogenic acid and rutin were significantly higher in the cherry samples (90.27–243.00 µg/g DW and 160.60–433.99 µg/g DW respectively) as compared to common tomatoes (21.30–88.72 µg/g DW and 24.84–110.99 µg/g DW respectively). The identification of dihydroquercetin is of particular interest, as it had not been reported previously in tomato fruit.

Electronic supplementary material

The online version of this article (10.1007/s13197-020-04258-z) contains supplementary material, which is available to authorized users.

Introduction

Tomato is the second most consumed vegetable in the world, after potato, with a worldwide harvest of over 162 million tons annually (Martínez Bueno et al. 2018; Slimestad and Verheul 2009). It is a good source of nutrients and bioactive compounds that are important for human health, including mineral matter, vitamins, lycopene, flavonoids, organic acids, phenolics and chlorophyll (Choi et al. 2014; Bressy et al. 2013; Sánchez-Rodríguez et al. 2012). Recent studies have demonstrated that the consumption of tomato is associated with a reduced risk of inflammation, diverse cancers, cardiovascular diseases, diabetes and obesity and can increase cell protection from DNA damage by oxidant species (Erba et al. 2013; Raiola et al. 2016). In particular, phenolic compounds, including phenolic acids (chlorogenic, caffeic, ferulic, gallic, p-hydroxybenzoic, protocatechuic and p-coumaric acids) and flavonoids (rutin, quercetin, naringenin, kampferol and derived), are of great interest due to their antioxidant and radical scavenging properties (Espin et al. 2016; Raiola et al. 2016). In this context, the characterization of the phenolic acids and flavonoids present in tomatoes is extremely important as potentially useful compounds with respect to health benefits (Vallverdú-Queralt et al. 2012).

During the last years, an increased interest, concerning the quality and traceability issues of food commodities, was observed on the consumer side. This because, it was demonstrated that the growing conditions, geographical origin and plant variety are directly related to products quality. Thus the nature and concentration of synthesized compounds are influenced by agricultural practices, environmental factors, variety, and ripeness (Choi et al. 2014; Slimestad and Verheul 2009). In addition, the accumulation of metals varies greatly both between species and cultivars (Bressy et al. 2013).

Among the analytical techniques which have an acknowledged potential for authenticity and traceability studies, stable isotope ratios of so-called bio-elements (²H/¹H; ¹⁸O/¹⁶O; ¹³C/¹²C; ¹⁵N/¹⁴N) are widely used. These were successfully used for differentiation of geographical origin of food and beverages and for identification of agricultural practices used for their production (Magdas et al. 2018). Thus, the determination of stable isotope ratios of ¹³C/¹²C, ¹⁸O/¹⁶O and ²H/¹H demonstrated to be a good instrument for characterization of geographical origin, because isotope ratios change in compliance with latitude and environmental condition of the place where the plant was grown. The isotopic fingerprint of ²H/¹H and ¹⁸O/¹⁶O can differentiate location and climate, and thus they represent important markers in geographical origin verification (Magdas et al. 2018).

In literature, different association markers were proposed for geographical and growing regime assessment. In this regard, stable isotopes in combination with mineral composition and chemical markers were used for investigation of the geographical origin of tomatoes (Mahne Opatić et al. 2018). Stable isotopes or elemental composition, either independently or in combination were also used for discrimination of other agroproducts with respect to their geographical origins such as: hot pepper (Marincas et al. 2018), potato (Magdas et al. 2017), wine (Hosu et al. 2016), or to differentiate the agriculture type (organic vs. conventional vegetables as well as greenhouses vs. field) (Cristea et al. 2017; Magdas et al. 2018). Most commercial tomatoes, especially in northern European countries, are produced in greenhouses that allow better control of agronomic and environmental factors. Nevertheless, greenhouse-grown crops are often affected by powdery mildew whose infection may result in considerable economical damage and low fruit quality, and require the use of fungicides (Erba et al. 2013). Therefore, monitoring of pesticide residues in tomatoes grown in greenhouses has become essential in order to ensure food safety.

The objectives of this paper were (1) to investigate and compare the phenolic profile of different tomato varieties, (2) to investigate the suitability of the use of stable isotope ratios in combination with multi-elemental analyses to provide a screening method for the control of geographical origin, (3) to evaluate the influence of agricultural practices (greenhouses vs. field) on the content of isotopic, elemental, and trace pesticides.

Materials and methods

Chemicals and reagents

The standards of phenolic compounds: gallic acid, dihydroquercetin, chlorogenic acid, caffeic acid, sinapic acid, vitexin, rutin, isoquercetin, quercetrin hydrate, rosmarinic acid, quercetin, cinamic acid, kaempferol, naringenin and apigenin were purchased from Sigma-Aldrich (San Luis, MO, USA). Analytical standards of organochlorinated pesticides mix: Appendix IX, 2 mg/mL each in 1:1 toluene:n-hexane were obtained from Supelco Analytical (Beleffonte, PA, USA) and AOAC QuEChERS QC Spike Mix, 40 μg/mL each in 99.9:0.1 acetonitrile:acetic acid from Restek (Bellefonte, PA, USA). DisQuE extraction tubes (50 mL containing 6 g magnesium sulfate and 1.5 g sodium acetate and 2 mL containing 150 mg of MgSO4, 25 mg primary secondary amine, and 25 mg of C18) were supplied by Waters (Dublin, Ireland). Acetonitrile and methanol (LiChrosolv) were of HPLC gradient grade and were purchased from Promochem (LGC standards, Germany) and Merck (Darmstadt, Germany). Acetic acid and n-hexane (≥ 98.0%) was also purchased from Merck.

Water was purified on a Milli-Q water system (MilliPore Simplicity). Stock solutions of polyphenols were prepared in a methanol at 1 mg/mL. All stock solutions were stored at − 20 °C until their use.

Tomatoes samples

Forty-one samples of tomato were purchases either directly from small farms of local producers or from commercial available sources. From the samples collected from local farmers, 14 of them were grown in greenhouse while 7 tomatoes were produced in open field.

Analysis of phenolic compounds

In this study 20 samples of tomatoes grown in greenhouse of different varieties (six cherry samples and fourteen common sorts) were taken for analysis.

Extraction methods

Fresh tomato samples were crushed in a blender until a completely homogenized mixture was obtained. After this step the samples were stored at − 60 °C until analysis. Phenolic acids and flavonoids were analyzed according to the method described by Del Giudice et al. (2015). Briefly, 3 g of tomato frozen powder were extracted with 30 mL of 70% methanol into an ultrasonic bath for 60 min at 30 °C. Mixture was centrifuged at 3500 × g for 10 min at 4 °C, and the supernatant was stored at − 20 °C until evaluation of phenolic compounds.

To check the efficiency of extraction procedure, the internal quality control was used. For quality control purpose the recovery of one spiked sample with 5 mg/kg target compounds, and extracted using the method described above (C1), was performed. The same tomato sample was extracted without spiking (C0) and recovery was calculated as follows: R = 100 × (C1 − C0)/Cspiked.

Standard solutions and calibrations

Stock solutions of 15 standards were prepared in methanol. Standards of gallic acid, chlorogenic acid, caffeic acid, sinapic acid, rutin, isoquercetin, quercetrin hydrate, rosmarinic acid, quercetin, cinamic acid, kaempferol and naringenin were accurately weighed and dissolved in volumetric flasks to obtain stock solutions of 1 mg/mL, while stock solutions of dihydroquercetin, vitexin, apigenin were prepared in the same way with concentrations of 500 μg/mL. The appropriate volumes of each stock solution were mixed together, and then diluted serially to prepare the working standard solutions for the construction of calibration curves. All solutions were stored under refrigeration.

HPLC analysis

Chromatographic analysis was performed using an ultra high liquid chromatography system (Accela UHPLC system, Thermo Scientific, USA) equipped with quaternary pump, autosampler and photodiode array detector (PDA) monitored at 190–400 nm. A BDS Hypersyl C18 (150 mm × 4.6 mm) column was used for all separations. The binary mobile phase, consisted of 2% (v/v) acetic acid in water (solvent A) and 2% (v/v) acetic acid in acetonitrile (solvent B), was pumped at a flow rate of 700 μL/min for a total run time of 60 min. The system was run with a gradient program: 0–5 min, 5% B; 5–15 min, 15% B; 15–45 min, 30% B; 45–50 min, 50% B; 50–55, 5% B and kept for 5 min for re-equilibration of the column to initial solvent conditions. The injection volume was 10 μL.

Quantifying of phenolic compounds

The identification of the peaks was performed on the basis of their retention times and UV spectra. Quantification of the identified compounds was performed by HPLC–PDA detection using the external standard method with calibration graphs, as a function of concentration based on peak area, detected at the wavelength corresponding to their maximum absorbance. Calibration curves in the 2.5–50 μg/mL range with good linearity (r² > 0.999) for a nine point plot were used to determine the concentration of phenolic compounds in tomato samples.

Limits for detection and quantification were between 0.39–0.96 and 1.30–3.21 μg/mL, respectively. Recoveries obtained in tomato sample ranged from 71.73 to 103%. Values are also given in Table 1.

Table 1.

HPLC parameters of the optimized HPLC–PDA method in tomato

Nr. crt.	Phenolic compounds	tr	r2	Recovery %	LOD (μg/mL)	LOQ (μg/mL)
1	Gallic acid	3.76 ± 0.05	0.9992	89.92	0.42	1.39
2	Dihydroquercetin	11.83 ± 0.25	0.9993	99.11	0.96	3.21
3	Chlorogenic acid	13.01 ± 0.19	0.9994	95.39	0.64	2.10
4	Caffeic acid	14.45 ± 0.14	0.9990	86.06	0.55	1.84
5	Sinapic acid	22.25 ± 0.08	0.9987	94.85	0.53	1.76
6	Vitexin	22.95 ± 0.09	0.9991	103.32	0.49	1.65
7	Rutin	23.54 ± 0.10	0.9994	71.94	0.39	1.30
8	Isoquercetin	24.34 ± 0.10	0.9990	93.26	0.49	1.62
9	Quercetin hydrate	28.07 ± 0.12	0.9993	106.41	0.42	1.40
10	Rosmarinic acid	30.31 ± 0.13	0.9991	71.73	0.49	1.64
11	Quercetin	38.02 ± 0.12	0.9991	88.63	0.49	1.62
12	Cinnamic acid	39.44 ± 0.10	0.9972	99.25	0.68	2.26
13	Naringenin	43.53 ± 0.12	0.9988	85.98	0.50	1.66
14	Apigenin	45.36 ± 0.12	0.9987	40.60	0.52	1.74
15	Kaempferol	46.41 ± 0.12	0.9990	84.93	0.50	1.66

As shown in Fig. 1, the separation of a standard mixture of 15 phenolic compounds can be achieved in 60 min. Table 1 lists the retention times of individual phenolic compounds.

Fig. 1.

HPLC chromatogram of a mixture of 15 phenolic compounds: (1) gallic acid, (2) dihydroquercetin, (3) chlorogenic acid, (4) caffeic acid, (5) sinapic acid, (6) vitexin, (7) rutin, (8) isoquercetin, (9) quercetin hydrate, (10) rosmarinic acid, (11) quercetin, (12) cinnamic acid, (13) naringenin, (14) apigenin, (15) kaempferol

Pesticide analysis

For this part of our study, only the tomato samples directly collected from local farmers were used. Thus a number of 21 samples of tomatoes (14 samples of tomatoes grown in the greenhouses and 7 tomato samples grown in open field) were compared.

Samples preparation

An optimized QuEChERS method by adding n-hexane in the first stage of extraction of pesticides was used for the determination of 30 organochlorine pesticides in tomatoes (Covaciu et al. 2017). Fresh tomatoes were crushed and homogenized using a blender and stored at − 20 °C until analysis. A total of 15 g of homogenized sample was treated with 10 mL of 1% acetic acid in acetonitrile and 2 mL of n-hexane. Samples were allowed to stand for an hour. The QuEChERS extraction salt containing 6 g MgSO₄ and 1.5 g CH₃COONa was added directly to the tube that were sealed and shaken for 2 min to prevent salt agglomeration before centrifugation at 4000 rpm for 5 min. The upper hexane layer was discarded by a transfer pipette. 1.6 mL of the acetonitrile extract was transferred into a 2-mL DisQuE tube containing 150 mg of MgSO₄, 25 mg of primary secondary amine (PSA), and 25 mg of C18. The tubes were capped tightly and vortexed for 1 min. The 2-mL tubes were centrifuged at 4500 rpm for 5 min. The upper extract was concentrated to dryness under nitrogen. The tomato residue was reconstituted in 1 mL of n-hexane, vortexed, and 1 μL was used for analysis by gas chromatography.

GC–ECD analysis

A gas chromatograph (GC, Thermo Fisher Scientific Trace GC Ultra) equipped with electron capture detector (ECD) was employed for detection and quantification of pesticides. The separation was performed on a DB-5MS (stationary phase consisting of phenyl 5% and dimethylsiloxane 95%) capillary column (30 m × 0.25 mm i.d., × 0.25 μm thickness). The column oven temperature was controlled with gradient programs and was initiated at 75 °C, held for 1 min, increased to 150 °C at a rate of 25 °C/min, then it was raised to 225 °C at a rate of 6 °C/min and finally ramped up to 290 °C at a rate of 15 °C/min and held for 10 min. The injection port was maintained at 250 °C in splitless mode, and the detection system was maintained at 300 °C. Helium was the carrier gas with a flow rate of 1.1 mL/min. Nitrogen was used as the makeup gas at 30 mL/min.

Quantification of pesticides

The gas chromatography-mass spectrometer (GC–MS, Thermo Fisher Scientific Finnigan Trace GC with a Polaris Q ion trap) was used for the confirmation of pesticides in the tomatoes. The mass spectra were obtained in full scan mode with electronic impact ionization at 70 eV and an ion source temperature of 250 °C.

Quantification of the identified compounds was performed by GC–ECD. An external calibration plot, as a function of concentration based on peak area was constructed for the analysis of tomato samples at levels from 10 to 150 μg/kg. The response for all pesticides was linear in the concentration range with a correlation coefficient higher than 0.998. The pesticides were detected within 30 min. Adequate separation of all pesticides was achieved with sharp peaks.

Stable isotopes analysis

Tomato samples were firstly freeze-drying and transformed afterwards in a fine powder. δ¹³C measurements were carried out on an Elemental Analyser (Flash EA1112 HT, Thermo Scientific), coupled with an isotope ratio mass-spectrometer IRMS (Delta V Advantage, Thermo Scientific). The used standard was NBS-22 oil with a certified value of − 30.03‰ versus PDB (Pee Dee Belemnite).

Isotopic values are denoted in delta versus Vienna Pee Dee Belemnite (V-PDB) international standards, according to equation (Brand et al. 2014):

$${\delta }^{i}X=\frac{R_{\mathit{sample}}}{R_{\mathit{standard}}}-1$$

where i is the heavier isotope mass number of the element X (i.e. ¹³C) while, R_sample is the isotope number ratio of sample (for example, ¹³C/¹²C), meanwhile R_standard is that of internationally standard. Delta values are multiplied by 1000 and being expressed in units “per mil” (‰).

For water extraction from tomato samples, without isotopic fractionation, a cryogenic distillation system under static vacuum was used. For δ²H and δ¹⁸O determinations of extracted water, a Liquid–Water Isotope Analyzer (DLT-100, Los Gatos Research) was used. The results are reported using conventional δ notation relative to the Vienna-Standard mean Ocean Water (V-SMOW).

The analytical reproductibility was ± 0.2‰ for δ¹³C, ± 0.2‰ for δ¹⁸O values and ± 0.6‰ for δ²H values.

Elemental analysis

All 41 samples were stored in clean polythene bags. They were washed with ultrapure water and dried at 60 °C. The dried tomatoes were ground into a fine powder and passed through a 0.45-mm diameter sieves before digestion. The method for microwave-assisted vegetable digestion was optimized to be: 0.1 g of sample and 2.5 mL of 60% ultrapure HNO₃ were placed in a clean Teflon digestion vessel. The vessel was tightly closed and placed in the microwave oven. The digestion was performed using a controlled pressure and temperature program at 200 °C. After digestion, each sample was transferred quantitatively with ultrapure water and diluted to 50 mL. These digests were filtered through 0.45-μm Teflon membranes and analyzed by inductively coupled plasma–mass spectrometry (ICP–MS). Three replicates were analyzed for each sample.

An ICP–MS (Perkin-Elmer Elan DRC-e) with a Meinhard nebulizer and glass cyclonic spray chamber for pneumatic nebulization was used for elemental analysis. The relative standard deviations were less than 10% for all metals.

Statistical analysis

Chemometric interpretation was made using SPSS Statistics version 24 (IBM, USA). The data matrix was formed by elemental along with isotopic content, measured in 40 tomato samples. Linear discriminant analysis (LDA) is widely applied when two or more groups need to be compared, with emphasizing the most representative descriptors, which are comprised in a linear function (model). Firstly, an initial classification is made based on assigned codes of each sample. Sample from one group have similar characteristics, and are different from those from another group, which means that LDA tries to minimize the distance between samples from the same group, and maximize the distance between samples belonging to different groups. The validation of the model is made using the procedure called “leave-one out” which means that every samples is tested as a new one, with the model obtained from the remained samples (Dias et al. 2009). For both initial and validation procedure, the result is expressed in percentages, the higher the percent is, the more accurate the model is. Classification model allows the prediction of membership of new samples to one group or another (Osorio et al. 2015). In this work, LDA aimed the differentiation of tomatoes samples according to geographical origin and growing regime.

Results and discussion

Phenolic profile of different varieties of tomatoes

Six samples of cherry tomatoes and fourteen common tomatoes were evaluated. For the determination of phenolic profile we optimized a HPLC method for the separation and identification of phenolic compounds in tomato extract. Fifteen standard phenolics were analyzed including phenolic acids and flavonoids (gallic acid, dihydroquercetin, chlorogenic acid, caffeic acid, sinapic acid, vitexin, rutin, isoquercetin, quercetrin hydrate, rosmarinic acid, quercetin, cinamic acid, kaempferol, naringenin and apigenin). Only 7 of them were identified in the extract of tomato samples. In Fig. 2 is presented a representative HPLC profile for a tomato sample. Dihydroquercetin, caffeic acid, chlorogenic acid, rutin, rosmarinic acid, quercetin and naringin were the major phenolic compounds detected in the set of samples studied, which agreed well with the studies of other authors (Coyago-Cruz et al. 2019; Espin et al. 2016; Raiola et al. 2016). The results are shown in Table 2 and they were expressed as μg/g dry weight (DW), because the samples were lyophilized. Different concentrations of these were found in each of the analysed varieties, with significant differences among them. In general, the concentration range of these compounds are wide because the content in matrices depends on several factors such as variety, crop type, environmental conditions, location, germination, maturity (Alarcón-Flores et al. 2013).

Fig. 2.

HPLC profile of a tomato extract recorded at 320 nm

Table 2.

Concentrations of phenolic compounds (μg/g DW) determined in tomato samples

Nr. crt.	Tomato sample	Dihydroquercetin	Chlorogenic acid	Caffeic acid	Rutin	Rosmarinic acid	Quercetin	Naringenin
1	Cherry tomatoes	26.73 ± 1.13	190.59 ± 12.01	3.09 ± 0.17	160.60 ± 0.90	4.84 ± 0.19	2.20 ± 0.08	2.20 ± 0.52
2		28.91 ± 0.78	193.86 ± 13.54	3.48 ± 0.22	240.39 ± 7.93	18.69 ± 0.72	5.48 ± 0.26	9.46 ± 1.17
3		22.09 ± 0.77	90.27 ± 1.38	18.56 ± 0.50	433.99 ± 10.97	10.63 ± 0.46	4.48 ± 0.15	14.78 ± 0.81
4		8.56 ± 0.62	243.00 ± 7.59	3.31 ± 0.08	183.30 ± 1.48	5.30 ± 0.25	4.86 ± 0.11	9.88 ± 0.32
5		12.79 ± 0.67	240.16 ± 13.64	10.03 ± 0.34	268.07 ± 4.26	10.29 ± 0.57	4.04 ± 0.18	17.88 ± 0.75
6		38.57 ± 1.56	214.29 ± 16.88	2.20 ± 0.12	186.12 ± 4.27	9.05 ± 0.52	3.60 ± 0.21	7.54 ± 0.42
7	Common tomatoes	20.81 ± 0.57	35.50 ± 1.94	10.32 ± 0.34	96.38 ± 1.27	12.44 ± 0.33	3.21 ± 0.11	8.95 ± 0.31
8		2.95 ± 0.24	29.66 ± 2.23	11.91 ± 0.73	59.09 ± 1.46	14.40 ± 0.50	1.12 ± 0.05	9.32 ± 0.61
9		27.99 ± 1.25	42.39 ± 1.78	10.23 ± 0.27	79.59 ± 1.67	13.36 ± 0.55	2.65 ± 0.08	11.49 ± 0.54
10		10.33 ± 0.44	24.93 ± 0.81	2.85 ± 0.21	60.19 ± 0.47	1.21 ± 0.07	1.21 ± 0.03	2.17 ± 0.27
11		16.79 ± 0.71	21.39 ± 1.12	7.79 ± 0.16	44.29 ± 1.13	11.60 ± 0.48	2.33 ± 0.08	1.83 ± 0.11
12		80.65 ± 2.65	56.54 ± 2.79	11.01 ± 0.50	67.51 ± 0.96	11.61 ± 0.56	1.83 ± 0.06	2.92 ± 0.18
13		12.25 ± 1.17	51.16 ± 2.54	65.11 ± 1.52	110.99 ± 2.34	15.65 ± 0.64	4.50 ± 0.17	9.00 ± 0.35
14		29.44 ± 1.03	21.30 ± 1.82	16.07 ± 0.72	36.26 ± 0.75	22.67 ± 0.79	4.22 ± 0.14	10.38 ± 0.49
15		35.53 ± 2.02	34.03 ± 1.78	10.71 ± 0.32	79.58 ± 4.23	19.24 ± 0.45	2.85 ± 0.12	11.19 ± 0.35
16		19.60 ± 1.04	52.78 ± 5.16	9.20 ± 0.28	44.13 ± 1.33	12.72 ± 0.35	1.13 ± 0.04	7.17 ± 0.30
17		32.01 ± 1.34	51.67 ± 2.52	3.70 ± 0.20	41.86 ± 1.84	19.94 ± 0.47	1.00 ± 0.08	0.98 ± 0.08
18		27.69 ± 1.28	83.09 ± 3.41	9.90 ± 0.26	36.00 ± 1.11	11.96 ± 0.54	0.56 ± 0.03	2.23 ± 0.15
19		14.12 ± 0.40	88.72 ± 2.79	5.99 ± 0.20	36.72 ± 1.23	10.24 ± 0.21	1.12 ± 0.04	2.28 ± 0.05
20		8.25 ± 0.46	33.94 ± 2.44	2.07 ± 0.07	24.84 ± 0.46	9.16 ± 0.32	0.93 ± 0.11	3.65 ± 0.41

Values expressed as mean ± SD (n = 3)

The most abundant phenolic compounds in our samples were rutin and chlorogenic acid. These results are in a good agreement with those reported by other authors who observed that chlorogenic acid was the major phenolic acid and rutin the major flavonoid together with naringenin (Alarcón-Flores et al. 2013; Raiola et al. 2016). Chlorogenic acid levels ranged from 18 to 230 µg/g, while rutin ranged from 11 to 290 µg/g, which are very similar to values obtained in this work, where chlorogenic acid levels ranged from 21.30 to 240.16 µg/g DW and rutin from 24.84 to 433.99 µg/g DW, respectively.

Phenolic profile showed significant differences between varieties of tomatoes. The contents of the chlorogenic acid and rutin were significantly higher in the cherry samples (90.27–243.00 µg/g DW and 160.60–433.99 µg/g DW respectively) as compared to common tomatoes (21.30–88.72 µg/g DW and 24.84–110.99 µg/g DW respectively). Similar results were obtained in previous studies (Fleuriet and Macheix 1981; Slimestad and Verheul 2005) where it was observed that chlorogenic acid has been detected as the main phenol in some cherry tomato varieties and rutin as the dominating flavonol compound. The explanation of these higher quantities of chlorogenic acid and rutin in cherry tomatoes is given by the fact that this variety is a good source of phenolic compounds, containing about 3–4 times the flavonoid content of standard sized tomatoes (U.S. Department of Agriculture, Agricultural Research Service 2011). This is because cherry tomatoes have more surface area per unit volume than standard sized tomatoes and to the fact that the majority of phenolics are believed to be in the epicarp (Choi et al. 2014).

Rosmarinic acid concentrations ranged between 1.21 and 22.67 µg/g DW. In another study by Espin et al. (2016), rosmarinic acid was found as majority phenolic compound in four tree tomato cultivars and reported concentrations between 12.22 and 121.89 mg/100 g DW. Rosmarinic acid is recognized as a powerful antioxidant to which relevant biological activities have been attributed and the identification of rosmarinic acid is to be emphasized (Espin et al. 2016).

The naringenin concentrations ranged in our investigated tomato samples between 0.98 and 17.88 µg/g DW. These results are not in line with those reported by other authors (Del Giudice et al. 2015; Sánchez-Rodríguez et al. 2012) who observed that naringenin, is the main detected flavonoid. Naringerin has been reported as it protects peripheral lymphocytes of diabetic mice from DNA damage and reduces oxidative stress (Del Giudice et al. 2015; Oršolić et al. 2011).

The quercetin concentrations fluctuated between 0.56 and 5.48 µg/g DW. These values were comparable to those found in other studies (Alarcón-Flores et al. 2013; Raiola et al. 2016). However, relatively high amounts of quercetin have been found in tomato by other authors (Coyago-Cruz et al. 2019).

The discrepancies between these studies and our work may have been due to genetic, agricultural practices or environmental differences. It has been found that the flavonoids content of tomato varies due to both abiotic and agronomic factors (Sánchez-Rodríguez et al. 2012).

Dihydroquercetin concentrations ranged from 2.95 to 80.65 µg/g DW. As far as we know, the presence in tomato of dihydroquercetin is reported for the first time herein, none of the consulted authors cited the presence of dihydroquercetin. Dihydroquercetin is the product of the enzymatic hydroxylation at the 3′ position of the dihydrokaempferol and a substrate for the enzyme flavonol synthase (FLS) leading to the production of quercetin (Bovy et al. 2007), therefore it is expected to be found in tomato extracts.

Comparison between Romanian and abroad tomato samples

Our principal aim when we used chemometric analysis was to highlight the differences that occur between Romanian and abroad produced samples. Stable isotope ratios of light elements (Carbon, Oxygen, Hydrogen) along with elemental content measurements, for 41 tomatoes samples (Table S1, Supplementary Data), were compiled into a data matrix used further for chemometric processing. The geographical classification was made by applying stepwise LDA on the above-mentioned matrix. All the samples were coded according to their origin, into a new dependent variable. By applying stepwise method, for both initial and cross validation steps, a percent of 100% was reached. Since two groups were compared (Romania vs. abroad) only one discriminant function was obtained, which was able to explain the entire variance of the dataset. The obtained function was statistically significant, with p = 0.001 and Wilks Lambda value of 0.132. The samples distribution according to their discriminant scores is presented in Fig. 3.

Fig. 3.

Tomatoes distribution according to their origin, after performing stepwise LDA

The best descriptors used for this classification were Ti, Ga, Te, δ²H and δ¹³C content. The highest content of Ti, Ga and Te were measured in the abroad samples. The level of Ti in Romanian samples ranged from 0.12 to 4.07 mg/kg as compared with Ti level in abroad (2.51–10.68 mg/kg), Ga concentration was between 0.024 and 0.10 mg/kg in Romanian tomato samples as compared with 0.005–0.047 mg/kg in abroad samples, and Te level fluctuated between nd–0.14 mg/kg for samples collected from Romania, whereas for abroad samples the value were between nd–0.81 mg/kg.

δ²H is an acknowledged marker for geographical origin assessment, being directly related to the precipitations which fall in a specific area and thus, with the specific fingerprint of tomatoes growing region (Magdas et al. 2017; Marincas et al. 2018). Apart of this, the isotope ratio of carbon (δ¹³C) proved to be a good tool for the characterization of geographical origin. This is because, δ¹³C values of plants are influenced by the availability of water, relative humidity and temperature, which control stomata aperture and the internal CO₂ concentration in the leaf (O’Leary 1995), which are also, specific parameters for each geographical location.

Differentiation between field and greenhouse tomato samples

Another purpose for applying LDA was to evidentiate the elements that could be used as markers for growing regime (open field vs. greenhouse), as was already mentioned in subsection “Pesticide analysis”. The samples were collected directly from local farms that had available information related to their growing conditions (field vs. greenhouse). Thus, a matrix consisting in 21 tomato samples was constructed, with elemental, isotopic and pesticides content as independent variables. With regard to the analyzed pesticides, those that presented values for only one sample were eliminated from the interpretation, only values for δ-BHC, Chlorpyrifos methyl, endosulfan I, endrin, endosulfan II, 2,4-DDD, 4,4′-DDT and lambda cyhalothrin being retained. LDA was applied to the obtained matrix, the two predefined groups being represented by the open field and greenhouse grown tomatoes (the dependent variable). After applying this method, a 100% initial classification (Fig. 4) was obtained and the same percentage for cross-validation. Due to the fact that two categories (groups) were compared, a statistically significant discriminating function was obtained and which is capable of explaining in 100% the variability of the experimental data (Fig. 4). This function has as main parameters the following elements: Al, Sc, Se, Dy, Pb, δ¹⁸O, 4,4′-DDT. These elements proved to be the best predictors to fit an unknown sample belonging to the group of tomatoes grown in the greenhouse or those grown in the open field. A probable source of some of the elements could be represented by the application of different fertilizers to improve the nutritional quality of tomato fruit and which might contain traces of these elements. In plants, low concentrations of Se can enhance plant growth, improve plant yield, and counteract oxidative stress induced by diverse abiotic factors or pathogen infection. In some crops, Se also improves plant quality (Zhu et al. 2018). According to a previous study on tomato fruit, foliar spraying of sodium selenate postponed fruit ripening and maintained fruit quality by reducing the production of ethylene and reactive oxygen species (ROS) (Zhu et al. 2017). Along with Se content, dysprosium (Dy) was found to be a discrimination factor. The content of rare earth elements (REE) in tomato plants has mainly been employed in agricultural research to improve production methods or to study varietal intercropping outputs (Spalla et al. 2009). In tomato berries most of REEs are present only at trace level (Fragni et al. 2018). Our data showed a very low content of REEs in tomato samples, according to those described by Spalla et al. (2009). The mean value of Dy in tomato samples grown in field was 0.95 mg/kg, whereas in those grown in the greenhouse, mean value was 2.30 mg/kg. Lead is one of the most toxic elements for humans, animals, and plants (Tokalıoğlu et al. 2018). The maximum tolerable concentration limit of Pb in plants (WHO 2007) is 10 mg/kg. In this study, the concentrations of Pb did not exceed the permitted level in any of the samples. The Pb content varies from nd to 0.1 mg/kg in samples grown in field and from 0.03 to 0.29 mg/kg in those grown in greenhouse. Differences observed in δ¹⁸O values of tomatoes grown in greenhouses from those on field appear due to several environmental factors like: water supplied to the plants (different in terms of isotopic composition) and distinct evapo-transpiration rate (which exists between the two growing regimes) (Rossmann et al. 1999).

Fig. 4.

Tomatoes distribution according to their growing regime, after performing stepwise LDA

Detected levels of individual pesticide residues ranged from < LOQ to 5.74 μg/kg and were below maximum residue level (MRL) established for tomatoes (EFSA 2016), as can be seen from Table S2 (Supplementary Data). Exception made lambda-cyhalothrin and chlorpyrifos where the concentrations in two tomato samples grown in the greenhouse were 60.07 μg/kg and 63.15 μg/kg, respectively, being over the MRLs. The level of 4,4′-DDT were between < LOQ-5.74 μg/kg in tomato samples grown in the greenhouse and between < LOQ-0.43 μg/kg in tomato samples grown in field. The insecticide 4,4′-DDT has been widely used for pest control because of its low cost, broad spectrum activity and high residual biological activity. Although its use has been prohibited for agricultural purposes, because of the negative impact on wildlife and human health, it remains adsorbed to soil particles, leading to half-lives of up to 30 years in this matrix (Mitton et al. 2014, 2018). Is known that before 1989 (during communism period) extensively use of the land for agricultural purpose implied the use of high quantities of pesticides and chemical fertilizers. These residues still remained in some agricultural fields.

The percentage of pesticide presence for each sample was detected in a larger number in greenhouse samples. These results are a direct consequence of different agricultural practices applied to pest management on each system. Under open field growth, the accumulation of pesticides on tomato can be reduced due to the direct effect of precipitation and solar radiation. On the other hand, despite the higher yields that can be achieved, growing conditions in greenhouses favor the development of outbreaks of arthropod pests and plant disease, which leads to excessive use of pesticides. Practices applied for combating pests, continuous ripening of fruits, intervals between spraying and harvesting in both types of growth (greenhouse/field) are not respected by producers.

Conclusion

The classification of 41 tomato samples was made based on stable isotope ratios of light elements (δ¹³C, δ¹⁸O and δ²H), elemental content, phenolic compound and pesticides. The results obtained from SLDA showed that to differentiate between Romanian and abroad tomato samples, the obtained discriminant variables were Ti, Ga, Te, δ²H and δ¹³C, while the best markers for growing regime (open field vs. greenhouse) were: Al, Sc, Se, Dy, Pb, δ¹⁸O, 4,4′-DDT.

For the differentiation among distinct tomatoes varieties, a HPLC–PDA method was optimized for the identification and quantification of phenolic compounds from tomato extract. Dihydroquercetin, caffeic acid, chlorogenic acid, rutin, rosmarinic acid, quercetin and naringin were the major phenolic compounds detected in the studied samples set. The detection of dihydroquercetin should be emphasized because, to our knowledge, it had not been previously reported in tomato fruit, its presence in tomato being reported for the first time herein. Based on determined parameters it was observed that the phenolic compounds profile obtained for the studied varieties is different. Thus, cherry tomato samples showed a high content of chlorogenic acid and rutin as compared to common tomato.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.