**4. Early cultivation in Europe and in the world**

Probably only a few tomato seeds were brought back from Mexico to Europe, leading, after domestication, to a new genetic bottleneck. George McCue has extensively reviewed the history of tomato diverse uses, tracking back the first references by country upon the time (McCue 1952). Most remote reference available comes from Petrus Matthiolus, an Italian Physician (1544). Due to its botanical closeness with toxic *Solanum* species common in Europe (Mandrake, Belladonna), tomato was for long mostly used as an ornamental. Two centuries later, it was referred as a cultivated plant in Italy by Saccardo (1769). Southern Europe was precursor in use of tomato for human consumption. In France, Bois at first mentions it as ornamental (1760). The same author reported it as vegetable seeds sold in the catalogue of the seeds of the "Maison grainière Andrieux Vilmorin" (1778). Lamarck mentioned it in 1798. Extensive consumption in Spain is described by Quer (1784).

<sup>1</sup> Approximate Bayesian Computation

center theory, Harlan suggested that biloculed domesticated forms found in south Mexico and Guatemala are the oldest cultivated types (Harlan 1971). Quoting Sahagun, Diez argued that tomato was totally "integrated" in the Aztec civilization food consumption in XVI century, contrary to South American Incas (Diez and Nuez 2008). Nevertheless, two authors identified Quechua names possibly referring to tomato: "pirca" (Horkheimer 1973) and "pesco-tomate" (Yakovleff 1935). However, botanists consider the origins of tomato domestication as unsolved (Peralta and Spooner 2007b). These authors mention recent evidences showing that the Mexican hypothesis is not supported by comparative data, as South American and Mexican tomato accessions share similar isozymes (Rick and Fobes 1975) as well as molecular markers (Villand, Skroch et al. 1998). So far, no evidence appears to be enough conclusive and tomatoes may have been domesticated independently in both areas. To go further a more extensive analysis of molecular polymorphism in the wild and cultivated tomatoes is needed. This would allow investigating demographic scenarios and estimating the parameters of these scenarios (bottleneck intensity, ancestral population size, migration rates) using Markovian model implemented in tools such as IM\* program (Hey and Nielsen 2004) or ABC1 methodology (Beaumont, Zhang et al. 2002; Lopes and Beaumont 2010). Very recently, this approach has been implemented to infer past demography and ecological parameters of two tomato wild relatives, *S. chilense* and *S. peruvianum* (Tellier,

Many authors consider *S. lycopersicum* var. *cerasiforme* as ancestral form of the cultivated tomato. It is present in both Mexico and Peru, on the contrary to *S. pimpinellifolium* which is absent from Mexico. If we assume that *S.l. cerasiforme* results from direct domestication from *S. pimpinellifolium*, a consequence of this domestication is that *S.l. cerasiforme* suffered a decrease of its population effective size during domestication (Bai and Lindhout 2007). Subsequent changes occurred for domestication traits such as growth habit, mating system, gigantism and fruit morphological diversity. Notably a change from exerted to inserted stigmas is responsible for the change from partial allogamy to strict autogamy. Selection for self-pollinating as well as shortening of the stigma compared to close wild relatives such as *S. pimpinellifolium* has allowed a yield increase (Rick 1977b). This "selfing syndrome" (Sicard and Lenhard 2011) is striking in tomato where a mutation in gene controlling stigma length

Probably only a few tomato seeds were brought back from Mexico to Europe, leading, after domestication, to a new genetic bottleneck. George McCue has extensively reviewed the history of tomato diverse uses, tracking back the first references by country upon the time (McCue 1952). Most remote reference available comes from Petrus Matthiolus, an Italian Physician (1544). Due to its botanical closeness with toxic *Solanum* species common in Europe (Mandrake, Belladonna), tomato was for long mostly used as an ornamental. Two centuries later, it was referred as a cultivated plant in Italy by Saccardo (1769). Southern Europe was precursor in use of tomato for human consumption. In France, Bois at first mentions it as ornamental (1760). The same author reported it as vegetable seeds sold in the catalogue of the seeds of the "Maison grainière Andrieux Vilmorin" (1778). Lamarck mentioned it in 1798. Extensive consumption in Spain is described by Quer (1784).

has been identified in cultivated germplasm (Chen, Cong et al. 2007).

**4. Early cultivation in Europe and in the world** 

Laurent et al. 2011).

 1

Approximate Bayesian Computation

Progressively, following South-North gradient, tomato consumption reached Northern Europe (Sabine 1819). Similarly in USA, Bartram (1766) reported tomatoes being used as food plants. Boyd (1784) mentioned that The David Landreth Seed Co. started to sell tomato seeds for vegetable consumption. Selection for diverse fruit shapes and local adaptation probably rapidly occurred through bulk selection. The crop gained in economic importance by the end of XIXth century with the establishment of tomato breeding programs. Most of the plant material at that time can be considered as landraces: selected for subsistence agriculture environments, producing low but relatively stable yield. At the end of the XIXth century, tomato cultivars (nowadays called landraces or heirlooms) were open pollinated from which seeds were saved by the farmers from a year to the other. Selection of new genotypes within heterogeneous cultivars (or selection of chance variance) resulted from spontaneous mutations, natural outcrossing or recombination of pre-existing genetic variation. Thus, *S. lycopersicum* found in Europe a secondary centre for diversification (García-Martínez, Andreani et al. 2006). In the XIXth century, establishment of commercial routes and colonies contributed to spread the species worldwide (Diez and Nuez 2008). In United States, prior to 1850 and "Trophy" the first commercially successful variety, no breeding programs were effective (Smith 2000). On an evolutionary perspective, domestication and implementation of breeding programs induced physiological changes. Artificial selection has reduced the genetic diversity of the crop which suffered a new bottleneck.

### **5. Tomato breeding in the XXth century: Seeking for diversity and intensive production**

After domestication and adaptation to North hemisphere growing conditions for two centuries, the crop started the XXth century with benefits of two major scientific discoveries: The rediscovery of Mendel pioneering work to set up the basis of experimental methods on the use of plant hybridization (Mendel 1866). Second are established domestication concept (Darwin and Wallace 1858) and selection theory (Darwin 1859).

This context has seen the emergence in public institutes of plant germplasm banks, starting point for collecting existing genetic diversity, preserving and valorizing it, following the pioneer work of Nikolai Vavilov (1887-1943) (Kurlovich, Rep'ev et al. 2000). Later on, he was followed by Charles Rick (1915-2002) who dedicated his life to discover, collect and characterize exotic tomato germplasm (Tanksley and Khush 2002). Today, more than 83,000 tomato accessions are stored in seed banks worldwide, ranking 1st among vegetable species collected (FAO 2010). The main collections in the world are: In USA, the Tomato Genetic Resources Center in California (TGRC), (www.tgrc.ucdavis.edu ) and the USDA2 collection (www.ars.usda.gov), the World Vegetable Center in Taiwan (www.avrdc.org) and several Europeans collections. The establishment of tomato resource collections made major contributions to understand the distribution of its diversity around the world. Nevertheless the lack of coordination and conflicting passport data is a pitfall for an efficient tomato germplasm management. Efforts are now made to coordinate national initiatives in global or regional approaches. Since 2007, The European Cooperative Programme for Plant Genetic Resources (www.ecpgr.cgiar.org) is a collaborative project between most European countries for long-term conservation and utilization of plant genetic resources in Europe.

<sup>2</sup> United States Department of Agriculture, Geneva

This project is based on large network of national centers for tomato genetic resources including COMAV3 (Spain), CGN4 (Netherlands), INRA (France), IPK5 (Germany), Vavilov Institute (Russia) and others. These institutions share their germplasm informations through a database (http://documents.plant.wur.nl/cgn/pgr/tomato/). More recently, in the context of a European Solanaceae project (EU-SOL, www.eu-sol.wur.nl), a collection of more than 6,000 domesticated tomato accessions was established and phenotyped, accompanied by an *ad hoc* database (Finkers, de Weerd et al. 2011). Finally, since 2008, a world initiative, is conducted under the Plant Biodiversity Inventories (www.nhm.ac.uk/researchcuration/projects/solanaceaesource/). Project aim is to produce a worldwide taxonomic monograph of the species occurring within the plant genus *Solanum.* As well, tomato is part of long term collection of plant species project, launched by the Svalbard Global Seed Vault initiative (Food 2008).

Due to its broad use for food consumption and adaptation to many environmental conditions (from Alaska summers to tropical conditions) and different crop systems, tomato experienced an important phenotypic diversification. Hundreds of past and present cultivated varieties are now available. Cultivars are dedicated to two main markets, processing and fresh market. Processing tomatoes are cultivated as a field crop, whereas fresh market tomatoes are grown outdoor or indoor (heated and non-heated greenhouses). Breeding objectives have evolved over time, with the evolution of production systems. Nevertheless, three main objectives remain: adaptation to growth constraints, disease and pest resistances and fruit productivity and quality. Wild species were first used as source of adaptation to biotic stress. Disease resistance selection started in United States early XXth century. The first *fusarium* wilt –resistant cultivar "Tennessee red" was released in 1912. Early 1920's breeders used hybridization with selection in segregating generations. By the mid-30's, plant breeders developed technical procedures to improve selection, such as pedigree selection. Later, existing or emerging private companies enhanced their development with the release of F1 hybrid varieties. Selection for disease resistance was successful as dominant resistance genes were found in the wild relatives for most of the diseases and pests. Modern cultivars can cumulate up to 12 different disease resistance genes which all derive from wild species. Wild germplasm has been primarily used as a source of major resistances.

Processing tomato industry was developed to provide North American and European households canned tomato, tomato paste and ketchup. Processing tomato varieties differ from fresh market ones in their pulp volume. Their growing conditions are dramatically different from fresh tomatoes (open field, mechanical harvesting). Thus the main criteria for processing tomatoes are fruit firmness, plant type with short fruit set period to produce a high percentage of ripe fruits simultaneously. Compact fruit set was obtained from a natural mutation discovered at the beginning of the XXth century, named *sp* (for self-pruning), conferring to the plant a determinate growth. This mutation was introduced in the well known "Roma" variety, whose long fruit type became a specific trait in processing tomatoes. In the 60's, VF145 was the first variety mechanically harvested. This cultivar has been the major cultivar for tomato ketchup industry for more than a decade in California. Apart from *sp*, several other mutations (detected in cultivated tomato or through interspecific

<sup>3</sup> Centro de Conservación y Mejora de la Agrodiversidad Valenciana, Valencia

<sup>4</sup> Center for Genetic resources, Wageningen

<sup>5</sup> Institut für Pflanzengenetik und Kulturpflanzenforschung, Gatersleben

This project is based on large network of national centers for tomato genetic resources including COMAV3 (Spain), CGN4 (Netherlands), INRA (France), IPK5 (Germany), Vavilov Institute (Russia) and others. These institutions share their germplasm informations through a database (http://documents.plant.wur.nl/cgn/pgr/tomato/). More recently, in the context of a European Solanaceae project (EU-SOL, www.eu-sol.wur.nl), a collection of more than 6,000 domesticated tomato accessions was established and phenotyped, accompanied by an *ad hoc* database (Finkers, de Weerd et al. 2011). Finally, since 2008, a world initiative, is conducted under the Plant Biodiversity Inventories (www.nhm.ac.uk/researchcuration/projects/solanaceaesource/). Project aim is to produce a worldwide taxonomic monograph of the species occurring within the plant genus *Solanum.* As well, tomato is part of long term collection of plant species project, launched by the Svalbard Global Seed Vault

Due to its broad use for food consumption and adaptation to many environmental conditions (from Alaska summers to tropical conditions) and different crop systems, tomato experienced an important phenotypic diversification. Hundreds of past and present cultivated varieties are now available. Cultivars are dedicated to two main markets, processing and fresh market. Processing tomatoes are cultivated as a field crop, whereas fresh market tomatoes are grown outdoor or indoor (heated and non-heated greenhouses). Breeding objectives have evolved over time, with the evolution of production systems. Nevertheless, three main objectives remain: adaptation to growth constraints, disease and pest resistances and fruit productivity and quality. Wild species were first used as source of adaptation to biotic stress. Disease resistance selection started in United States early XXth century. The first *fusarium* wilt –resistant cultivar "Tennessee red" was released in 1912. Early 1920's breeders used hybridization with selection in segregating generations. By the mid-30's, plant breeders developed technical procedures to improve selection, such as pedigree selection. Later, existing or emerging private companies enhanced their development with the release of F1 hybrid varieties. Selection for disease resistance was successful as dominant resistance genes were found in the wild relatives for most of the diseases and pests. Modern cultivars can cumulate up to 12 different disease resistance genes which all derive from wild species. Wild germplasm has been primarily used as a

Processing tomato industry was developed to provide North American and European households canned tomato, tomato paste and ketchup. Processing tomato varieties differ from fresh market ones in their pulp volume. Their growing conditions are dramatically different from fresh tomatoes (open field, mechanical harvesting). Thus the main criteria for processing tomatoes are fruit firmness, plant type with short fruit set period to produce a high percentage of ripe fruits simultaneously. Compact fruit set was obtained from a natural mutation discovered at the beginning of the XXth century, named *sp* (for self-pruning), conferring to the plant a determinate growth. This mutation was introduced in the well known "Roma" variety, whose long fruit type became a specific trait in processing tomatoes. In the 60's, VF145 was the first variety mechanically harvested. This cultivar has been the major cultivar for tomato ketchup industry for more than a decade in California. Apart from *sp*, several other mutations (detected in cultivated tomato or through interspecific

3 Centro de Conservación y Mejora de la Agrodiversidad Valenciana, Valencia

5 Institut für Pflanzengenetik und Kulturpflanzenforschung, Gatersleben

initiative (Food 2008).

source of major resistances.

4 Center for Genetic resources, Wageningen

hybridization) were used in tomato breeding. For example, the jointless (Szymkowiak and Irish 2005; Quinet, Kinet et al. 2011) *j2* allele was introgressed from *S. Cheesmani*i. In combination with *sp*, *j2* has been bred into many processing varieties, allowing a large scale mechanical harvest of tomato fruits. Major genes used in tomato breeding are listed in Table2.

Today, after a rapid evolution towards very firm fruits and long shelf life varieties (with the major success of the variety *Daniela*, which carried the spontaneous *rin* mutation), consumers request more diverse texture and tastes (Causse, Friguet et al. 2010). The fresh tomato market faces rapid developments and diversification (Navez 2011). New products and varieties are emerging always faster and their life cycle gets shorter, 5 years in average (Bai and Lindhout 2007). Consumption trend is for broader and diverse choice of fruit types. After the development of truss and cherry type tomatoes, new cultivars resembling to old heirloom varieties are developed. If access to allelic diversity is a must to improve fruit quality, choices in breeding objectives are critical to maintain organoleptic fruit values. Strong associations are often made by consumers between morphology and sensory values. Association of "Oxheart" or "Marmande" fruit shape with a pleasant texture is a good example. This link can be lost through modern breeding (Casals, Pascual et al. 2011). Improved content in potential health beneficial components such as anti-oxidants (lycopene, vitamin C) is also promoted. This can be obtained thanks to specific mutations like *hp* (Lieberman, Segev et al. 2004) or, again, by the introgression of genes from wild relatives.

Farmers and breeders have shaped diversity over years in an ever-evolving process that is hard to track and to record. Intensive breeding of crop varieties by modern science has increased the genetic erosion which started with domestication. Nevertheless the introgression from wild relatives allowed major progress and introduced a new source of diversity. Charles Rick observed that crosses between wild and cultivated species generated a large diversity of novel phenotypic diversity. Rick's work represents milestones for the modern use of genetic diversity in tomato. It led to uncover positive transgressive variation within interspecific progenies. This encouraged a greater use of exotic germplasm and thus larger gene pools to unlock causal polymorphism.
