Site-specific adaptation by natural selection. A case study with lentil.

In the context of conservation and use of plant genetic resources, both storage in gene banks (ex situ) and safeguard on-farm (in situ) are important concepts. Increasingly attention focuses on the on-farm management of plant genetic resources as a dynamic maintenance strategy. To date, little is known about the potential of ‘evolutionary’ adaptation to site-specific conditions and the mechanisms involved. Lentil is an autogamous food legume crop with limited recombination. In long-term experiment three old lentil cultivars were exposed to ten generations of natural selection at three farms in Central and Northern Germany under rain-fed conditions. Two of the selection sites were on soils with low fertility and expected frequent shortages in water supply during the growing season. The following four main questions were investigated: i) Did natural selection during ten generations lead to site-specific adaptation in terms of yield? ii) Did natural selection alter phenotypic and genotypic variability? iii) What is the importance of seed weight for natural selection and adaptation? iv) Was drought a significant factor for site-specific development? The three lentil cultivars are distinctly different in their morphological and phenological traits. Populations in the fifth and tenth generation under site-specific natural selection within each cultivar and the corresponding initial material were compared in a two-year field trial at the three selection sites. Singleplant progeny testing during two years at one location and SNP genotyping using KASPar™ assays was done to assess phenotypic and genotypic within-population variability. Selections for extreme seed size for six generations from each cultivar were phenotyped in the field and genotyped by SNP markers. Additionally, populations in the tenth generation of natural selection were tested in induced drought in rainout shelters for their variation in drought stress tolerance during three years at one location. Significant continuous site-specific changes in seed weight and flowering time in both directions were observed for one cultivar. Natural selection at the more stress-prone location led to a lower seed weight and earlier flowering, whereas in the most fertile location selection favored larger seeded genotypes and later flowering. However, these changes did not lead to a significant site-specific adaptation in terms of yield. But a general yield advantage across locations for the population with a lower seed weight and earlier flowering was observed. A general tendency towards better yield for the latest generations across cultivars and locations suggests that natural selection was positive in terms of the general performance. Phenotypic and genotypic variation within the initial material of one cultivar goes back to three main genotypes, which differ significantly in seed weight and flowering time. The observed phenotypic changes and site-specific differences in the genetic variability are supposedly caused by different selection pressure at the three locations. However, at the most stress-prone location, within-population variability is relatively high due to recombination, whereas the population selected at the most fertile location is highly dominated by one genotype. Across all locations, the genetic variation within this cultivar was maintained. The other two cultivars exhibited a very low genetic variability and were maintained at all three farms without changing their characteristic seed weight significantly. A general yield improvement by selection for larger seeds was observed in all cultivars. The full genetic variation within one cultivar was maintained by the selection for both extremes in seed size. Selection within two cultivars with very low genetic variability resulted in an unintended genetic modification from the initial material by selecting ‘off-types’, probably resulting from spontaneous outcrossing, with smaller seeds in one case and with larger seeds in the other. Significant site-specific variation in drought tolerance was not found. Thus, the effect of year may have counteracted a continuous site-specific adaptation or selection sites are not as distinctly different as expected in water regimes. Consequently, the main factors involved in the continuous site-specific changes in one cultivar remain unidentified. It was observed that cultivars reacted differently to drought conditions. On-farm management with lentil can lead to a site-specific development and site-specific adaptation may occur in the long-term. Given a substantial genetic variability, seed yield, straw yield, harvest index, seed weight, and flowering time were altered significantly by natural selection. Our results prove the maintenance and management on-farm of plant genetic resources in an autogamous species a dynamic management strategy. For the maintenance of material with substantial genetic diversity, at least two farms with significantly different environmental conditions are necessary. Homogeneous accessions, when combined with ex situ maintenance, could be managed at one location. I recommend the characterization of the material in advance for its genetic diversity and outcrossing potential to establish a well-grounded management strategy accordingly. Different material of the same crop should not be maintained at one farm. Selection for traits considered negative for agronomic performance and monitoring of characteristic traits may be considered to preserve the integrity of the material and/or to improve its agronomic performance. To study the main factors involved in site-specific adaptation by natural selection for autogamous crops, I suggest using genetically broad populations, which should be well characterized in advance, phenotypically as well as by molecular markers, to monitor their spatial and temporal development in diverse environments