Introduction
The maternal environment in which seeds develop and mature have a profound influence on seed properties such as germination vigour. The sink-source connection between the mother plant and the seeds allows the seeds to accumulate reserves required for seed germination and seedling growth (Baud et al. 2008). Metabolites such as amino acids, sugars and organic acids play a vital role in the different stages of seed development such as maturation, desiccation and germination (Borisjuk et al. 2004; Fait et al. 2006). During seed maturation, the content of these metabolites in seeds decreases and storage reserves, including starch, oil and seed storage proteins increase (Fait et al. 2006; Galili et al. 2014). It has also been shown that the subsequent metabolite content and composition of dry seeds may reflect the maturation environments in which they developed (He et al. 2016). For example, in different species it has been reported that nitrogen related metabolites such as asparagine, allantoin and GABA show a lower content in seeds developed under low nitrate maternal environments (Geshnizjani et al. 2019; Heet al. 2016). Although many studies have been performed related to the effect of maternal environments on dry seed metabolic content, more information is required to understand the genetic and molecular mechanisms governing the metabolic changes in response to the maternal environment.
In general, each observed phenotype in plants is the consequence of different cellular processes such as gene transcription, protein translation and, finally, metabolite production (Kooke & Keurentjes 2011). Therefore, genetic variation is not only confined to phenotypic traits such as seed and seedling quality traits. Many studies have revealed that metabolite composition and content, which play a very critical role in plant growth and development, is also controlled by genetic variation within a plant species (Windsor et al. 2005). The existing natural variation for both phenotypic traits and metabolite content is displayed by a continuous distribution, considered as quantitative variation. Such variations are often regulated by multiple loci and can be detected in mapping populations like recombinant inbred line (RIL) populations where the different loci are known as phenotypic or metabolite quantitative trait loci (QTLs and mQTLs, respectively) (Keurentjes & Sulpice 2009; Lisec et al. 2008). Many QTL analyses have been performed in seeds and many QTLs that regulate complex quantitative traits such as seed germination characteristics, seed size, seedling traits as well as seed metabolites have been described (Kazmi et al. 2012; Kazmi et al. 2017; Khanet al. 2012; Schauer et al. 2006).
Plants are a rich source of biochemical compounds that are mainly contributing to plant development, adaptation and final appearance and yield (Binder 2010). Therefore, the quantitative variation of these metabolites may have an influence on different physiological traits like seed germination and seedling establishment. The integrative analysis of metabolites and genetics has provided valuable information and knowledge on how natural variation regulates metabolite levels and their subsequent effect on growth of plants and their adaptation and how this knowledge can be used in plant breeding (Kliebenstein 2009).
Genetical genomics in which QTL analysis is integrated with proteomics, transcriptomics and metabolomics has provided in-depth understanding of molecular mechanisms regulating complex traits (Jansen & Nap 2001; Keurentjes et al. 2006; Kliebenstein et al. 2006; Schaueret al. 2006). Nonetheless, more advanced approaches are required for further determination of the complexity of quantitative traits. In addition to genotype (G), molecular networks are also influenced by the environment (E) and the interaction between genotype and the environment (G×E). Thus, the incorporation of different environments in genetic studies is a prerequisite for comprehensive perception of the regulation of molecular mechanisms. Li, Breitling and Jansen (2008) proposed a new strategy which is called generalized genetical genomics (GGG) by which both genetic and environmental perturbations can be studied. This approach allows QTL analysis governing the interesting molecular traits under consideration of multiple environments. It is a cost-effective method to not only determine the genotype but also the environmental effects and their interaction for detected QTLs (Li, Breitling & Jansen 2008). In principle, by creating similar subpopulations of RILs and subjecting each of these to a different environment, G, E and G×E effects can be investigated in a cost-effective experimental design (Joosen et al. 2013).
Although the QTLs governing dry seed metabolite content have been previously detected in many plants, including tomato (Kazmi et al. 2017; Toubiana et al. 2012), the effect of G×E interactions has been studied to a much lesser extent (Albert et al. 2016; Kazmi et al. 2017; Rosental et al. 2016). In this study we used a RIL population derived from a cross between two tomato species:Solanum lycopersicum (cv. Moneymaker) and Solanum pimpinellifolium (Voorrips et al. 2000). We have exploited the existing natural variation in this population to investigate how QTLs are influenced by the environment to which the mother plants are exposed. Moreover, metabolic profiling of the seeds which have matured in different environments will be useful to illustrate important metabolic differences that regulate the development and adaptation of plants (Joosen et al. 2013). By using a GGG approach we performed metabolite analysis for the RIL population and their parental lines, grown in high phosphate and low nitrate environments. By generating metabolite correlation networks and performing mQTL analysis, genetic and molecular aspects of seed metabolic changes in response to the maternal environments have been discovered.