© 2006 Nature Publishing Group Transgeneration memory of stress in plants Jean Molinier 1 †, Gerhard Ries 1 †, Cyril Zipfel 1 † & Barbara Hohn 1 Owing to their sessile nature, plants are constantly exposed to a multitude of environmental stresses to which they react with a battery of responses. The result is plant tolerance to conditions such as excessive or inadequate light, water, salt and temperature, and resistance to pathogens. Not only is plant physiology known to change under abiotic or biotic stress, but changes in the genome have also been identified 1–5 . However, it was not determined whether plants from successive generations of the original, stressed plants inherited the capacity for genomic change. Here we show that in Arabidopsis thaliana plants treated with short- wavelength radiation (ultraviolet-C) or flagellin (an elicitor of plant defences 6 ), somatic homologous recombination of a trans- genic reporter is increased in the treated population and these increased levels of homologous recombination persist in the subsequent, untreated generations. The epigenetic trait of enhanced homologous recombination could be transmitted through both the maternal and the paternal crossing partner, and proved to be dominant. The increase of the hyper-recombina- tion state in generations subsequent to the treated generation was independent of the presence of the transgenic allele (the recombi- nation substrate under consideration) in the treated plant. We conclude that environmental factors lead to increased genomic flexibility even in successive, untreated generations, and may increase the potential for adaptation. Plants are influenced by abiotic and biotic environmental factors on several levels; apart from changes in plant physiology and the mounting of resistance responses, the dynamics of the genome can also be altered. Examples include the activation of transposable elements by abiotic and biotic stress conditions 7–9 , induction of mutations by chemical and physical agents 10 , and enhancement of homologous recombination by elevated temperatures 11 or ultraviolet-B (UV-B) (ref. 2). Especially interesting is the genomic flexibility shown by plant genomes in response to pathogen attack 3,4,7 . Whenever possible, such changes were monitored at the level of the sequence of affected genes. The influence these changes have in evolutionary terms, however, remained poorly understood, because most changes were detected in somatic tissue and not considered in further generations. In plants, the reproductive cell-lineage emerges from somatic tissue late in development 12 , thus some genomic changes acquired during the life of a plant can be transmitted to the next generation. Indeed, with progeny of UV-B- or pathogen-treated plants, the frequency of occurrence of genetically fixed mutation (in this case, homologous recombination) was reproducibly elevated 2,4 . The degree of genomic change in the offspring of the stressed population was expected to return to the basal level. We show here that increased levels of homologous recombination persist for several generations in the lineage from the original parent plants that were exposed to stresses, including ultraviolet radiation or flagellin. We measured the rate of homologous recombination in the untreated offspring of plants exposed to conditions of environmental stress. We used A. thaliana plants harbouring b-glucuronidase (GUS)-based constructs in which truncated but overlapping parts of the gene allow quantification of somatic homologous recombina- tion. The results of this event are visualized as blue spots on a white background following histochemical staining of plants (Fig. 1a, b). Previous molecular analyses of the plant DNA confirmed that the blue spots, which represent GUS activity, indeed symbolize bona fide recombination events 13,14 . Using this assay, the influence of ultra- violet-C (UV-C) was tested in six independent transgenic lines that carried the recombination reporter in different relative orientations of the GUS sequence fragments: ‘GU’ and ‘US’ (ref. 15). The basal levels of homologous recombination, indicated as numbers of recombination sectors per plant, varied among the six lines; the degrees of stimulation were also different, but in all cases the treatment with UV-C stimulated the level of homologous recombi- nation (Fig. 1c). UV-C induction of homologous recombination together with variation between independent transgenic lines is consistent with previous reports 2 . LETTERS Figure 1 | Somatic homologous recombination in UV-C- and flg22-treated plants. a, Schematic representation of a recombination substrate used for monitoring somatic homologous recombination (lines IC1 and IC9). GUS, b-glucuronidase gene; Hpt, hygromycin-resistance gene. Homologous region is shown in dark blue. b, Recombination events (blue spots highlighted by black arrows) giving a measure of homologous recombination frequency (HRF; see Methods) in line IC1 after flg22 treatment. Scale bar, 1 mm; inset, £3 original magnification. c, Somatic HRF in untreated and UV-C-treated S 0 plants. Results are means ^ s.e.m. (n . 50 plants; t-test *P , 0.05). d, Somatic HRF in either untreated plants, plants treated with flg22 A. tum., or treated with flg22. Results are means ^ s.e.m. (n . 40 plants; t-test *P , 0.05). 1 Friedrich Miescher Institute for Biomedical Research, Maulbeerstrasse 66, CH-4058 Basel, Switzerland. †Present addresses: Institut de Biologie Mole ´culaire des Plantes, 12 Rue du Ge ´ne ´ral Zimmer, F-67084 Strasbourg Cedex, France (J.M.); BioMedinvestor AG, Elisabethenstrasse 23, CH-4051 Basel, Switzerland (G.R.); The Sainsbury Laboratory, John Innes Centre, Colney Lane, Norwich NR4 7UH, UK (C.Z.). Vol 442|31 August 2006|doi:10.1038/nature05022 1046