Tomatine and dehydrotomatine are two steroidal secondary metabolites found in tomato plants that have pharmacological and nutritional roles in the human diet due to anticancer and chemopreventive, anticholesterol, anti-inflammatory, antipyretic, antifungal and antibacterial properties. Tomato glycoalkaloids content depends on cultivar, fruit ripening stage, and agricultural practices. Present work was conducted to analyze the α-tomatine and dehydrotomatine content in leave, green and red fruits of fifteen tomato genotypes by HPLC. The extraction and chromatographic methods used, based on procedures reported in literature, were improved with minor modifications enhancing the percentages of recovery and the performances. The recovery efficiencies for extraction method ranged from 96 to 105 %, while the chromatographic analysis gave a linear response in the ranges 2.3-150.0 µg and 0.8-32.0 µg, with detection limits of 0.10 µg and 0.03 µg, and percentages of 90.8±4.8% and 87.8±4.4% respectively for α-tomatine and dehydrotomatine. The α-tomatine content varied from 0.54±0.07 to 1993.89±13.32 in leaves, from 0.24±0.04 to 527.97±3.62 in green fruits and from no detected to 119.00±2.82 in red fruits. The dehydrotomatine content varied from 0.08±0.004 to 635.11±4.39 in leaves, from 0.08±0.01 to 91.87±7.68 in green fruits and from no detected to 26.19±0.47 in red fruits.

Keywords: α-tomatine, dehydrotomatine, glycoalkaloid, tomato, HPLC

Introduction

Tomatoes (Solanum lycopersicum) are a major source of lycopene compounds (Agarwal and others 2001), and are also considered an important source of carotenoids, antioxidants, carbohydrates, fibers, flavor compounds, minerals, proteins, vitamins, calistegines, and glycoalkaloids in the human diet (Friedman 2002; Tommonaro and others 2011). Plants belonging to Solanaceae family, including eggplants (Solanum melongena), potatoes (Solanum tuberosum), and tomatoes (Solanum lycopersicum) produce typically two glycoalkaloids, such as α-solanine and α-chaconine for S. tuberosum, α-tomatine and dehydrotomatine for S. lycopersicum, α-solasonine and α-solamargine for S. melongena. Generally one glycoalkaloid is more abundant than the second one, and the mixture exhibits synergistic activity in hostplant resistance (Milner and others 2011). The presence of paired glycoalkaloids has been principally attributed to plant evolution (Friedman 2002).

The glycoalkaloids α-tomatine and dehydrotomatine, whose mixture is called tomatine (Friedman and others 1994) are the steroidal secondary metabolites found in all parts of the tomato plant (Friedman and Levin 1998) and their chemical structures show an identical hydrophilic tetrasaccharide part called lycotetraose, consisted of xylose, galactose and two glucose units, and an hydrophobic steroidal part, called tomatidine and tomatitidenol for α-tomatine and dehydrotomatine respectively, that differs by a double bond between C-5 and -6 of the ring B of the aglycon chain present only in the dehydrotomatine structure. The carbohydrate side chain seems to be essential to confer bioactivity properties to the glycoalkaloids (Blankemeyer and others 1995). The content of tomatine up to 2000 mg/kg (FW) in leaves (Kozukue and others 2004), up to 500 mg/kg (FW) in immature green tomatoes and up to 5 mg/kg (FW) in red ripe tomatoes (Friedman 2004) because it is known that these compounds are degraded during ripening in the fruits. However, in some Peruvian genotypes the content of tomatine can range from 500-5000 mg/kg (DW) in red tomatoes (Rick and others 1994). The Peruvians usually eat both green and red fruit of these high-glycoalkaloids genotypes without poisoned effects (Rick and others 1994; Friedman 2004). These glycosylated alkaloids have been shown to be involved in the natural mechanism of plant resistance against insects, fungi, bacteria and viruses (Milner and others 2011).

In humans, poisoning or toxic effects due to tomato glycoalkaloids consumptions have not been reported (Milner and others 2011); however, because it is notorious that potato glycoalkaloids are associated with food poisoning (Van Gelder 1991), the tomato glycoalkaloids α-tomatine and dehydrotomatine are considered as potentially toxic, although several researches indicated the important pharmacological and nutritional roles in the human diet that these compounds might have anticancer and chemopreventive (Friedman and others 2009; Lee and others 2011; Lee and others 2004), anticholesterol (Cayen 1970; Friedman and others 2000), anti-inflammatory (Kovacs and others 1964; Chiu and others 2008), antipyretic, antifungal and antibacterial properties (Milner and others 2011). Some of possible mechanisms involved in the tomatine biological activity might include the cholesterol complex producing and the potentiation of immune system (Milner and others 2011). However, furthermore studies need to assure the glycoalkaloid safety concentration and to elucidate the mechanism of their bioactive properties.

Thus, in the light of the plant high variability, human health insights and possible nutritional importance, present work was conducted to evaluate the α-tomatine and dehydrotomatine content in leaves, green and red tomato fruits of fifteen tomatoes genotypes by HPLC. We selected two high-tomatine accession of S. lycopersicum var. cerasiforme, LA2213 and LA2262, the known low-tomatine S. lycopersicum accession LA2295 (Rick and others 1994), several wild accessions, cultivated tomato Super Marmande, and the Slmlo line showing highly pathogens resistance. We employed some minor modification of chromatographic techniques and extraction procedure described in literature. The results, beyond to amply the knowledge on glycoalkaloids and their content in tomato plants, could be utilized to design breeding assays in order to investigate on GLA inheritance of segregant populations.

Materials and Methods

Plant samples and laboratory materials. Fifteen tomato plant accessions were grown from seeds in a glasshouse of University of Aberdeen, Scotland, UK. The green fruits (collected when the maximum size fruits was reached and when no color changing from green to red were observed), red ripe fruits (collected when whole fruit color was red) and leaves of tomato plants were harvested and analyzed separately for dehydrotomatine and α-tomatine contents. Both green and red fruits were collected at the ripening stage considered the most suitable for marketing (Leonardi and others 2000). Commercially available standard were purchased from Sigma (St. Louis, MO) (a 10:1 mixture of α-tomatine and dehydrotomatine) and from Santa Cruz Biotechnology (Santa Cruz, USA). HPLC grade acetonitrile, methanol, and analytical grade KH2PO4 were obtained from Sigma (St. Louis, MO). Before use, solvents were passed through a 0.45 μM membrane filter (Millipore, Bedford, MA) and degassed in an ultrasonic bath.

Extraction procedure. The extraction method was adapted from established extraction procedures (Friedman and others 1994; Friedman and Levin 1998; Kozukue and others 2004; Pegg and Woodward 1986), modifying slightly with some measures in the following manner. Samples (leaves, 5 g; fruits, 20 g) were homogenized in 60 ml of methanol-glacial acetic acid (98:2 v/v), placed in 50 ml tubes and extracted in an ultrasound bath (Fisherbrand, UK) for 20 minutes. The extract was filtered under vacuum through Whatman No3 filter paper. The tissue residue was re-extracted in 40 ml of solvent, as above, and filtered. Filtrates were combined and rotary evaporated to dryness at 37°C. The residue was dissolved in 40 ml 0.2 N HCl and centrifuged at 2350g for 10 minutes at 4°C. The supernatant was decanted into a clean vessel and the pellet rinsed in 0.2 N HCl and re-centrifuged, as above. The supernatants were combined and sufficient ammonia solution (0.88 sp. gr.) added to raise the pH 11. The solution was placed in a water bath at 65°C for 50 min, with shaking at approximately 10-15 min intervals. Solutions were maintained at 4°C overnight before centrifuging at 2350g for 35 min at 4°C. Supernatants were discarded and the pellet washed in 20 ml 2% aqueous ammonia solution and re-centrifuged. The final pellet was dried at 25°C for 10 minutes, suspended in 2 ml methanol/KH2PO4 (80:20; buffer pH 3.0) and centrifuged at 13000 x g for 15 min at 4°C. The supernatant was collected and used for HPLC analysis.

Commercial tomatine was added at the concentration of 0.5, 2, 4 mg/100 of fresh weight to green and red homogeneous tomatoes sample, both from stock solution and adding solid tomatine. These samples were extracted as usual to determine percent recovery from the extraction.

HPLC method. HPLC analyses were performed using an Agilent 1100 Series Liquid Chromatograph with diode array detector (DAD)

sets at 208 nm, on an Inertsil ODS-2 column (250 mm x 4.0 mm i.d.; 5 uM particle diameter) (Hichrom, UK). The conditions used for analysis were: temperature 25 °C, mobile phase acetonitrile (A)/ 20 mM KH2PO4 (B) pH 3.0, elution program (20, 20, 30, 40, 100, 100, 20, 20% acetonitrile at times 0,4, 15, 30, 31, 36, 37, 40 min), flow rate, 1 ml/min. α-Tomatine and dehydrotomatine both from standards and from tomato fruits and leaves were analyzed. The identities of the two peaks, corresponding to the two glycoalkaloids, were confirmed by Orbitrap Discovery LCMS (Thermo Scientific, UK). Data analysis was performed using ChemSation v. B.02.01 SR1.The injection volume was 25 µL for both standards and samples.

An 8-point standard curve was prepared by making dilutions of the tomatine stock standard (6, 5, 4, 2, 1, 0.500, 0.250, 0.125 µg/µL).

To determine the percent recovery, tomato extracted were analyzed before and after addition of known amounts of commercial tomatine solution (0.25 and 0.5 µg/µL).

Results and Discussion

The extraction and chromatographic methods used, based on procedures reported in literature, were improved with minor modifications enhancing the percentages of recovery and the chromatographic performances. Extraction methods reported in literature (Friedman and others 1994; Friedman and Levin 1998; Kozukue and others 2004; Pegg and Woodward 1986) were modified obtained recoveries from spiked tomato extracts, using precipitation method without solid phase extraction (SPE) step, from 96 to 105 % that were comparable to the recoveries had in previous studies using SPE reported to yield increased recovery (Friedman and others 1994). Probably using the ultrasound bath helps to disperse solvent and vegetal matrices well through the sample and, in addition, avoids the long stirring times used in the traditional methods of extraction (Friedman and others 1994; Friedman and Levin 1998; Pegg and Woodward 1986). Besides, this extraction approach allows us to obviate the recourse of the organic solvent chloroform adopted by Kozukue (2004), to operating safety, and to reduce the organic solvent waste.

Separation by HPLC of the α-tomatine and dehydrotomatine standards compared with the two glycoalkaloids isolated from tomato is shown in Figure 1 (a). Calibration plots for the glycoalkaloids (Figure 2 a,b) gave determination coefficients of 0.99. The intra-day and inter-day precision analysis was carried out by estimating the corresponding responses 3 times on the same day and on 3 different days of 8 different concentrations of commercial tomatine.

Retention times for the alkaloids were highly repeatable between individual analyses, at ~16 min and ~17 min, for dehydrotomatine and α-tomatine, respectively. The limit of detection, the concentration that yielded signal to noise ratio (S/N) 3:1, for α-tomatine was 0.10 µg and for dehydrotomatine 0.03 µg. These detection limits are approximately one order of magnitude lower than those published previously, based on HPLC-UV analyses (Kozukue and others 2004). The HPLC-UV method gave a linear response in the concentration range tested, 0.8 -32.0 µg and 2.3 -150.0 µg respectively, for dehydrotomatine and α-tomatine. The limit of quantification was the concentration that yielded signal to noise ratio (S/N) 10:1. This response also represented an improvement on previously reported methods. The accuracy of the method was determined by calculating recoveries of α-tomatine by method of standard additions. The average recovery of α-tomatine added to a sample was 90.8±4.8% (n=5). The corresponding value for dehydrotomatine was 87.8±4.4% (n=5). Mass spectrometry analysis confirmed the presence of the two glycoalkaloids (not shown). The ratio of α-tomatine/dehydrotomatine found for Santa Cruz tomatine standard was 2:1.

The method described in this paper was applied successfully to analyze the contents of the two glycoalkaloids in green and red fruits, and leaves in a range of tomato genotypes. Figure 1 b and c show the HPLC chromatograms of the α-tomatine and dehydrotomatine glycoalkaloids isolated from green and red tomatoes of S. lycopersicum accession LA2213. Overall, the α-tomatine and dehydrotomatine content in green and red fruits, and leaves found in all cultivars considered in this study and the corresponding diameter size are presented in Tab 1, 2 and 3. In Figures 3, 4, and 5 the histograms show the amount of tomato glycoalkaloids in green fruits, red fruits and leaves respectively.

The α-tomatine and dehydrotomatine content (Fig 3) were detected for all cultivars and accessions in green tomato fruits, ranging from 0.24±0.04 to 527.97±3.62 µg/g (FW) and from 0.08 ±0.01 to 91.87±7.68 µg/g (FW) respectively. The percentage of α-tomatine was from 70% to 87%. S. lycopersicum var. cerasiforme accessions LA2213 and LA2262 (number 1 and 2) have been shown by Rick and others (1994) to have a high content in glycoalkaloids in the red fruits. In the present study, the accession LA2262 have been revealed to have the highest total glycoalkaloids amount of 602.25±4.68 µg/g (FW) and the main percentage of α-tomatine (87.50%±0.28), while the accession LA2213 presented a glycoalkaloid content of 289.73±16.20 µg/g (FW) that was lower than the total amount detected for other samples including S. lycopersicum accession LA2295 that was also analyzed by Rick (1994) and considered a tomato line presenting low content in glycoalkaloids. Instead, our analysis showed that cultivar Super Marmande and the line Slmlo exhibited the lowest dehydrotomatine and α-tomatine content in the green tomato fruits and the lowest percentage of α-tomatine.

In the red tomato fruits, α-tomatine and dehydrotomatine are also detected in all tomato samples (Fig 4), varying from no detected to 119.00±2.82 µg/g (FW) and from no detected to 26.19±0.47 µg/g (FW). The percentage of α-tomatine ranged from 77% to 87%.The accessions LA2213 and LA2262 showed highest total glycoalkaloids amount of 145,19±3,25 and 120.06±4.28 µg/g (FW), and also the main percentage of α-tomatine (81.96%±0.16 and 87.27%±0.08), confirming the characteristic of these plants to have a high content in glycoalkaloids as studied by Rick (1994). Analogously with what has been obtained above, the cultivar Super Marmande and the line Slmlo exhibited the lowest tomato glycoalkaloids content and the lowest percentage of α-tomatine.

The percentage of total tomatine degraded during the maturation of the fruits was around 50 and 80% for the accessions LA2213 and LA2262 respectively, and 99% for other samples.

Fig 5 showed that the α-tomatine and dehydrotomatine content in the leaves varied considerably in the cultivars and accessions, ranging from 0.54±0.007 to 1993.89±13.32 µg/g (FW) and 0.08 ±0.04 to 635.11±4.39 µg/g (FW) respectively.

The cultivars S. hirsutum, S. peruvianum e S. lycopersicum show the highest glycoalkaloids content in the leaves, 635.11±4.39, 342.43±19.60 and 345.45±5.79 µg/g (FW) for dehydrotomatine and 1993.89±13.32, 1699.15±85.36 and 1613.06±167.34 µg/g (FW) for α-tomatine. Conversely, the two accessions of S. pennellii show very low content both in dehydrotomatine and α-tomatine. We suggested that the different result is probably due to other resistance mechanism against phytopathogens in the low content cultivars. In fact, contrarily to other cultivated tomato species, S. pennellii secrets sugar polyesters weeping by trichomes, involved in insect resistance (Li and others 1999; Cesio and others 2006).

The percentage of α-tomatine amount varied from 44% to 89%, and there was no evident relationship between these values and the total glycoalkaloid contents; in fact, S. hirsutum accession LA0094, exhibiting the highest α-tomatine and dehydrotomatine content, has been shown to present a percentage of α-tomatine (75.7%) that was inferior to S. pennellii accession LA1275 (89.2%), exhibiting the lowest α-tomatine and dehydrotomatine content.

Overall, quantitative and qualitative analyses by HPLC confirmed the presence of two tomato glycoalkaloids in the samples and endorsed the expected differences between varieties and accessions. Moreover, the findings corroborate the hypothesis that the biosynthesis/degradation of the dehydrotomatine and α-tomatine may be under distinct genetic control in the fruits and in the leaves of tomato plants (Kozukue and others 2004), as we detected different total amounts of two glycoalkaloids and also variations in the percentage of α-tomatine in the different parts of the tomato plant. We found a relationship between percentage of α-tomatine and total glycoalkaloid content, increasing with the rise of the glycoalkaloids content. We did not observe the same tendency for the percentage of α-tomatine in the leaves.

The results, beyond to amply the knowledge on glycoalkaloids and their content in tomato plants, leaded us to identify the accessions S. lycopersicum var. cerasiforme LA2213 and LA2262 as high glycoalkaloids content and the line Slmlo and Super Marmande as low glycoalkaloids content. The accessions S. lycopersicum var. cerasiforme LA2213 and the line Slmlo were utilized to design breeding assays in order to investigate on GLA inheritance of segregant populations. The segregant population F2 has been obtained and molecular marker analyses have been performed to investigate the genes involved in the glycoalkaloid synthesis (unpublished data).

Conclusions

It is well known that tomato is one of the major fruit largely consumed by humans and that tomato plat can produce glycoalkaloid tomatine. Poisoning cases caused by the ingestion of tomato correlating to glycoalkaloids content have not been reported in humans; however tomatine is considered preventively as potentially toxic based on well-known toxicity of potato glycoalkaloids which are chemically similar, even though many health-beneficial effects of tomatine have also been described. On one hand, whether further research needs to assure the full safety profile and the pharmacokinetics for this compound, on the other hand, also genetic studies are requested to investigate the glycoalkaloid biosynthesis, degradation and inheritance in the tomato plants. This work enhances the knowledge on glycoalkaloids content correlated to vegetal genotype, and could be suitable for development of inheritance study and selection of new tomato varieties with beneficial total glycoalkaloid content.