The tomato is an excellent model for studies of plants bearing berry-type fruits and for experimental studies of the Solanaceae family of plants due to its conserved genetic organization. In this study, a comprehensive mutant tomato population was generated in the background of Micro-Tom, a dwarf, rapid-growth variety. In this and previous studies, a family including 8,598 and 6,422 M2 mutagenized lines was produced by ethylmethane sulfonate (EMS) mutagenesis and γ-ray irradiation, and this study developed and investigated these M2 plants for alteration of visible phenotypes. A total of 9,183 independent M2 families comprising 91,830 M2 plants were inspected for phenotypic alteration, and 1,048 individual mutants were isolated. Subsequently, the observed mutant phenotypes were classified into 15 major categories and 48 subcategories. Overall, 1,819 phenotypic categories were found in 1,048 mutants. Of these mutants, 549 were pleiotropic, whereas 499 were non-pleiotropic. Multiple different mutant alleles per locus were found in the mutant libraries, suggesting that the mutagenized populations were nearly saturated. Additionally, genetic analysis of backcrosses indicated the successful inheritance of the mutations in BC1F2 populations, confirming the reproducibility in the morphological phenotyping of the M2 plants. To integrate and manage the visible phenotypes of mutants and other associated data, we developed the in silico database TOMATOMA, a relational system interfacing modules between mutant line names and phenotypic categories. TOMATOMA is a freely accessible database, and these mutant recourses are available through the TOMATOMA (http://tomatoma.nbrp.jp/index.jsp).

As metabolomics can provide a biochemical snapshot of an organism's phenotype it is a promising approach for charting the unintended effects of genetic modification. A critical obstacle for this application is the inherently limited metabolomic coverage of any single analytical platform. We propose using multiple analytical platforms for the direct acquisition of an interpretable data set of estimable chemical diversity. As an example, we report an application of our multi-platform approach that assesses the substantial equivalence of tomatoes over-expressing the taste-modifying protein miraculin. In combination, the chosen platforms detected compounds that represent 86% of the estimated chemical diversity of the metabolites listed in the LycoCyc database. Following a proof-of-safety approach, we show that % had an acceptable range of variation while simultaneously indicating a reproducible transformation-related metabolic signature. We conclude that multi-platform metabolomics is an approach that is both sensitive and robust and that it constitutes a good starting point for characterizing genetically modified organisms.

High-level accumulation of the target recombinant protein is a significant issue in heterologous protein expression using transgenic plants. Miraculin, a taste-modifying protein, was accumulated in transgenic tomatoes using an expression cassette in which the miraculin gene was expressed by the cauliflower mosaic virus (CaMV) 35S promoter and the heat shock protein (HSP) terminator (MIR-HSP). The HSP terminator was derived from heat shock protein 18.2 in Arabidopsis thaliana . Using this HSP-containing cassette, the miraculin concentration in T0 transgenic tomato lines was 1.4-13.9% of the total soluble protein (TSP), and that in the T1 transgenic tomato line homozygous for the miraculin gene reached 17.1% of the TSP. The accumulation level of the target protein was comparable to levels observed with chloroplast transformation. The high-level accumulation of miraculin in T0 transgenic tomato lines achieved by the HSP terminator was maintained in the successive T1 generation, demonstrating the genetic stability of this accumulation system.

We previously developed a transgenic tomato that expresses the miraculin gene using a constitutive promoter. In this study, we profiled the developmental and spatial accumulation of the miraculin protein and mRNA in transgenic tomato fruits. Miraculin mRNA expression was almost constant up to orange stage, and then the expression increased at red stage. The miraculin protein accumulated gradually during fruit development and reached its highest level at the overripe stage. At the red stage of fruit, miraculin protein was accumulated at the highest level in the exocarp, and similar in other fruit tissues: mesocarp, dissepiment, upper placenta, lower placenta and jelly. Moreover, the pattern of miraculin accumulation in fruit tissues was the same regardless of genetic background and position at which the miraculin gene was inserted in the genome. We also discuss suitable tomato types expressing miraculin for their commercial use.

In our previous study, a transgenic tomato line that expressed the MIR gene under control of the cauliflower mosaic virus 35S promoter and the nopaline synthase terminator (tNOS) produced the taste-modifying protein miraculin (MIR). However, the concentration of MIR in the tomatoes was lower than that in the MIR gene's native miracle fruit. To increase MIR production, the native MIR terminator (tMIR) was used and a synthetic gene encoding MIR protein (sMIR) was designed to optimize its codon usage for tomato. Four different combinations of these genes and terminators (MIR-tNOS, MIR-tMIR, sMIR-tNOS and sMIR-tMIR) were constructed and used for transformation. The average MIR concentrations in MIR-tNOS, MIR-tMIR, sMIR-tNOS and sMIR-tMIR fruits were 131, 197, 128 and 287 μg/g fresh weight, respectively. The MIR concentrations using tMIR were higher than those using tNOS. The highest MIR accumulation was detected in sMIR-tMIR fruits. On the other hand, the MIR concentration was largely unaffected by sMIR-tNOS. The expression levels of both MIR and sMIR mRNAs terminated by tMIR tended to be higher than those terminated by tNOS. Read-through mRNA transcripts terminated by tNOS were much longer than those terminated by tMIR. These results suggest that tMIR enhances mRNA expression and permits the multiplier effect of optimized codon usage.