Heterosis occurs when the F₁ individuals exhibit increased performance for a number of traits compared to their parental lines. On the other hand, reciprocal hybrids are obtained by changing the cross direction (or the sexual role) of parental genotypes in a cross. Both biological phenomena could affect the external and internal attributes of fleshy fruits. This work aimed to detect reciprocal effects and heterosis in different tomato (Solanum lycopersicum) for fruit quality traits and metabolite content. Twelve agronomic fruit traits and 28 metabolites identified and estimated by ¹H NMR were evaluated in five cultivars grown in two environments. Due to the genotype component was the most important over the phenotype, the traits were evaluated following a full diallel mating design among those cultivars, in greenhouse. Data demonstrated that hybrids performed a higher phenotypic diversity than parental lines. Although both type of traits displayed reciprocal effect and heterosis, the metabolites had a bigger impact on generating this variation, mainly the amino acids. This coincided with the fact that agronomic traits were more influenced by GCA and metabolites by SCA. Furthermore, a relationship between genetic distance between parental lines and reciprocal effect or heterosis was not found. Hybrids with heterosis and high content of metabolites related to tomato flavour and nutritional components were obtained. Results of this work highlight the impact of selecting the role of a cultivar as male or female in a cross to enhance the variability of fruit attributes through hybrids as well as the possibility to exploit heterosis. 

1. Description of methods used for collection/generation of data:

1.1 Characterization of parental genotypes by DNA molecular markers
Genomic DNA of frozen leaves (40 mg) from five parental genotypes was extracted with a commercial kit (Wizard® Genomic DNA Purification Kit, Promega®, Madison, WI, USA).  Two plants per genotype were used for the DNA extraction according to the manufacturer protocol, with some minor modifications. Samples were incubated 15 min at 65°C with 600 µl of Nuclei Lysis Solution. Then, 200 µl of protein precipitation solution was added and the samples were centrifuged for 3 min at 13000 rpm. The supernatant obtained was transferred to a clean 1.5 ml tube with 600 µl of room temperature isopropanol. The solution was gently mixed by inversion and centrifuged for 1 min at 13000 rpm. The supernatant was discarded and the pellet was washed with ethanol 70% v/v. The solution was centrifuged for 1 min at 13000 rpm and the pellet was dried at room temperature. Finally, the pellet was re-suspended in TE buffer (10 mM Tris pH 7.4 and 1 mM EDTA pH 8.0). Quantity and quality of DNA extracted was evaluated with agarose gel 1% dyed with SYBR® Safe (Thermo Fisher Scientific®, Waltham, MA, USA) by comparison with a Lambda DNA.
Molecular markers were used to estimate polymorphism and genetic distance among parental genotypes. A total of 16 SSR (Single Sequence Repeats) and 64 InDel (Insertion/ Deletion) markers covering the entire tomato genome were used (Cambiaso et al., 2019a,b). Also, four functional markers: FW2.2 (Blanca et al., 2015), FW3.2 (Chakrabarti et al., 2013), LC (Muños et al., 2011) and FAS (Rodríguez et al., 2011) were used.

The PCR amplification protocol consisted of 60 s at 94°C, 7 cycles (30 s at 94°C, 30 s at 58 °C – 52 °C and 2 min at 72°C), followed by 30 cycles including 30 s at 94°C, 30 s at 52°C and 2 min at 72°C. The last step was 5 min at 72°C. Amplified fragments were visualized in agarose gel 3% dyed with SYBR® Safe (Thermo Fisher Scientific®, Waltham, MA, USA).

1.2 Agronomic fruit traits

Six agronomic fruit traits were evaluated at breaker stage (fruits with 10 to 30% of final ripe colour, according to Giovannoni (2004)) in at least eight fruits per plant of each genotype: weight (FW, in g), height (H, in cm), diameter (D, in cm), shape index (FS), firmness (F, measured on the equatorial plane with a tester type Shore A) and shelf life (SL, measured as the number of days elapsed from harvesting to visualization of the first symptoms of deterioration in the fruits stored at 25±3°C, according to Green et al. (2016); Cambiaso et al. (2019a).

Six additional agronomic fruit traits were evaluated at red ripe stage (fruits with 90% of mature colour): locule number (LN), soluble solids content (SS, in ºBrix, determined with a hand refractometer in the homogenized juice from the pericarp tissue), pH, titratable acidity (TA, g of citric acid/100 g of homogenized juice) and the fruit colour. The fruit colour was estimated by the absorbance index (a*/b*, where a* is the absorbance at 540 nm and b* at 675 nm) and reflectance percentage (L*) with a Minolta Chroma Meter reflectance colorimeter (model CR-400, Minolta Co., Osaka, Japan), as the average of three measures per fruit.

1.3 Metabolic traits
First or second fruit from the 2nd and 3rd inflorescence was collected from three independent plants at ripe stage per genotype. The harvest was done from 10 AM to noon. Fruit tissue was conditioned by removing the exocarp and locular tissue; the seeds were discarded while the mesocarp was cut into small pieces, quick-frozen in liquid nitrogen and then stored at -80 °C. A 1H NMR was done according to the protocol described by López et al. (2015). One gram of frozen mesocarp was powdered and 300 µl of 1M phosphate buffer (pH 7.4) prepared in D2O was added. The samples were centrifuged at 12000 rpm for 30 min at 4°C and the supernatant was transferred to a new tube. The pH of solutions was adjusted to 7.4 using NaOH 1N. As internal standard 1mM of TSP (3-(trimethylsilyl) propionic-2,2,3,3-d4 acid sodium salt) prepared in D2O was added. An NMR Bruker Avance II spectrometer (Bruker, Germany) was used for the spectral analysis at 600.13 MHz. Proton spectra were acquired at 298 K by adding 512 transients of 32 K data points with a relaxation delay of 5 s. Intense signals corresponding to H2O protons were suppressed by pulse gradient techniques. The 90° flip angle pulse was always *10 ls. TSP was used for both, chemical shift calibration and quantitation, that is, proton spectra were referenced to the TSP signal (d = 0 ppm) and their intensities were scaled to that of TSP. Spectral calibration was made using the software TopSpin v.3.5 pl 7, Bruker BioSpin GmbH. The identification and quantification of specific metabolites were done using the software “Mixtures v2.0”, developed ad hoc as an alternative to commercial programs (Abriata, 2012). Metabolites concentration was expressed in µmol/gFW.

2. Environmental/experimental conditions: Eight plants of each parental genotype were transplanted to open field (2) and greenhouse (1). Tomato plants were pruned and conducted to a single stem. Eight plants of each hybrid were transplanted into a greenhouse together with the parental genotypes. Tomato plants were pruned to a single stem and tied vertically to a top wire. In both environments, the plants were distributed according to a completely randomized design with 1.4 m between rows and 40 cm between plants.
Plants were drip irrigated and supplemented with nitrogen, phosphorus and potassium + magnesium oxide + magnesium and trace elements fertilizer once a week. The crop was grown according to the standard cultural recommendations for the area at the Estación Experimental ‘‘José F. Villarino’’ (33° SL and 61° WL), Argentina in 2015-2016.