Multi-gene metabolic engineering of tomato plants results in increased fruit yield up to 23%

The capacity to assimilate carbon and nitrogen, to transport the resultant sugars and amino acids to sink tissues, and to convert the incoming sugars and amino acids into storage compounds in the sink tissues, are key determinants of crop yield. Given that all of these processes have the potential to co-limit growth, multiple genetic interventions in source and sink tissues, plus transport processes may be necessary to reach the full yield potential of a crop. We used biolistic combinatorial co-transformation (up to 20 transgenes) for increasing C and N flows with the purpose of increasing tomato fruit yield. We observed an increased fruit yield of up to 23%. To better explore the reconfiguration of metabolic networks in these transformants, we generated a dataset encompassing physiological parameters, gene expression and metabolite profiling on plants grown under glasshouse or polytunnel conditions. A Sparse Partial Least Squares regression model was able to explain the combination of genes that contributed to increased fruit yield. This combinatorial study of multiple transgenes targeting primary metabolism thus offers opportunities to probe the genetic basis of metabolic and phenotypic variation, providing insight into the difficulties in choosing the correct combination of targets for engineering increased fruit yield.

: Analysis of the effect of genotypes and growth conditions. P-value of two-way (factorial) ANOVA to test the effect of the genotype (gene expression) and the condition (i.e. environmental condition of the two experiments carried out in glasshouse and polytunnel) on fruit and leaves matrixes. * denotes significant effects of genotype, condition or interaction of them on gene expression (P<0.001).  Figure S1. Photosynthesis, dark respiration, stomatal conductance, chloroplast electron transport rate (ETR) parameters in transgenic lines measured under A) glasshouse and B) polytunnel conditions. Data presented are means ± SD (n between six and 10 plants per line). An asterisk indicates the values that were determined by the t-test to be significantly different (P < 0.05) from control.

B) Polytunnel
Supplementary Figure  The variables X and Y are represented through their projections onto the plane defined either by Xvariates or Y-variates. The variables X and Y being assumed to be of unit variance, their projections are inside a circle of radius 1 centered at the origin. Strongly associated variables are projected in the same direction from the origin. The greater the distance from the origin the stronger the association. Two circumferences of radius 1 and 0.5 are plotted to reveal the correlation structure of the variables under glasshouse (A) and polytunnel (B) conditions.

Supplementary Data
Evaluation of this entire gene expression dataset by two-way ANOVA analysis on leaves and fruits separately, revealed that most of genes are influenced significantly by genotype (line) suggesting differential gene expression responses of individual lines to the two growth conditions. When the effect of growth conditions and the interaction of genotype and growth conditions were tested, ANOVA analysis revealed far fewer genes were influenced (Supplementary  Table S2). In leaves, only three genes (AtSUS1, AtTMT1, and AtSUC9) on a few transgenic lines were significantly influenced by the interaction of growth conditions and genotype, suggesting that only these three genes responded to the tested growth conditions in a specific manner. On the other hand, ANOVA analysis on fruits, revealed that only EcPP, FpGLDH, and AtAAP1 gene expression on few transgenic lines were affected by interaction of growth conditions and genotype (Supplementary Table S2).

Supplementary Discussion
We were not able to find strong and common statistical support for many such features, we made some interesting observations. Sucrose levels increase during tomato fruit ripening. This rise may be due to the incoming sucrose from the photosynthate translocation from the leaf, where it is loaded into phloem in either an apoplastic or a symplastic manner 80,81,111,112 . This sugar is likely used to support respiration, as well as providing substrate to be metabolized into storage polymers and into primary metabolites needed for growth 113 . We observed a slight increase in sucrose levels correlated with similar increase in fructose, fructose 6-P, in mature fruit irrespective of the growth conditions indicating that the different amount of sugars in fruit may be dependent on endogenous metabolic processes. However, we cannot currently formally discard differences in the degree of phloem unloading, since tomato fruit has been demonstrated to have low photosynthetic activity 114 . This is particularly evident when it is considered that previous studies suggest that sucrose import ceases during tomato fruit ripening due to the formation of an abscission layer between the calyx and fruit 115,116 .
In plants grown under high light condition (polytunnel), we observed a general decrease in the levels of raffinose. Unlike the situation observed in the Curcubitaceae, raffinose does not constitute a significant component of phloemtransported sugars in tomato 117 . However, there could conceivably be a role for raffinose, in stress tolerance since it has been implicated to have membrane stabilization and antioxidative functions 118-120 . Raffinose is synthesized by transferring a galactose residue from galactinol to sucrose, and myo-inositol is used to synthetize galactinol 121 . In addition, myo -inositol itself is implicated to function as an osmolyte to enhance tolerance to abiotic stress 122 . Consistently, galactinol was reduced in these fruits. The lower raffinose level may indicate the use of the raffinose family of oligosaccharides as carbon sources 119 .
During normal tomato ripening at the initiation of ethylene biosynthesis, aspartic acid increases in addition to putrescine, one of the three major plant polyamines 123 . Recently, ethylene and polyamines have been reported to possess opposing biological roles: ethylene promotes senescence, whereas polyamines are known to suppress it, by slowing down membrane deterioration and loss of chlorophyll and enhancing protease and RNAse activities 124 . Interestingly, in mature fruits from transgenic plants grown under high light intensity, we observed high increase in aspartate acid level while putrescine decreased in comparison to control plants; however, we observed the opposite behavior in plants grown under low light intensity and limited soil conditions (greenhouse). Further investigations are, however, needed to examine the biological significance of these changes.
It has been long documented that proline accumulates under stress conditions 125,126 , as well as being one of many well-known compatible solutes in plants 127 . We observed accumulation of proline in mature fruit in the high yielding transgenic plants grown in high light conditions, indicating it may confer tolerance. Another general trend was the commonly increased level of phenylalanine observed in mature fruit from the high-yielding transgenic plants grown in the greenhouse while the opposite behavior was observed in fruits from the same plants grown in high light conditions. Interestingly, aromatic amino acids can act as alternative respiratory substrates in instances in which carbohydrates are not abundant 128 . Intriguingly, the mitochondrial electron transfer system which renders this possible is very highly expressed in tomato fruit tissues 129,130 , suggesting that this metabolic shift may support the enhanced energy requirements associated with elevate fruit growth in the greenhouse but that another mechanism is invoked under high light conditions when the phenylalanine is likely utilized in the production of phenylpropanoid sunscreen 131 .