LETTER doi:10.1038/nature 13291 Horizontal genome transfer as an asexual path to the formation of new species Ignacia Fuentes1*, Sandra Stegemann1*, Hieronim Golczyk2, Daniel Karcher1 & Ralph Bock1 Allopolyploidization, the combination of the genomes from two different species, has been a major source of evolutionary innovation and a driver of speciation and environmental adaptation1-4. In plants, it has also contributed greatly to crop domestication, as the superior properties of many modern crop plants were conferred by ancient allopolyploidization events5'6. It is generally thought that allopolyploidization occurred through hybridization events between species, accompanied or followed by genome duplication6,7. Although many allopolyploids arose from closely related species (congeners), there are also allopolyploid species that were formed from more distantly related progenitor species belonging to different genera or even different tribes8. Here we have examined the possibility that allopolyploidization can also occur by asexual mechanisms. We show that upon grafting—a mechanism of plant-plant interaction that is widespread in nature—entire nuclear genomes can be transferred between plant cells. We provide direct evidence for this process resulting in speciation by creating a new allopolyploid plant species from a herbaceous species and a woody species in the nightshade family. The new species is fertile and produces fertile progeny. Our data highlight natural grafting as a potential asexual mechanism of speciation and also provide a method for the generation of novel allopolyploid crop species. Grafting commonly occurs in nature913 and the observation of natural grafting may have inspired the application of grafting in agriculture and horticulture hundreds of years bce14. Natural stem grafts result from the mechanical pressure of interlocking stems or branches, leading to tissue fusion and establishment of new vascular connections. Natural root grafts1112 may influence population ecology by allowing a community of plants to share water, minerals and metabolites14. In nature, interspecific grafts have been observed between very distantly related species9. Recent work has demonstrated that entire chloroplast genomes can be horizontally transferred across the graft junction, potentially explaining so-called organelle capture events that have occurred in evolution1517. If nuclear genomes are also exchanged between neighbouring cells in natural grafts, this would provide a straightforward asexual mechanism by which new allopolyploid species could form. Indeed, microscopic evidence supports the idea of nuclear material occasionally moving from cell to cell in intact plant tissues, a phenomenon known as cytomixis18-20. To investigate the possibility that grafting enables the transfer of nuclear DNA, we generated two transgenic tobacco (Nicotiana tabacum) plants carrying different selectable marker genes. Line Nt-kan:yfp carries the kanamycin resistance gene nptll (and the gene for the yellow fluorescent reporter YFP) in its nuclear genome, whereas line Nt-hyg contains Figure 1 Grafting and experimental selection for horizontal transfer of nuclear DNA between tobacco plants, a, Map of the transgenic loci in the two tobacco lines used for grafting and selection. Genes above the line are transcribed from left to right, genes below the line in the opposite direction. Expression elements (promoters, terminators) are shown in grey. Pnos, nopaline synthase promoter; P35S, CaMV 35S promoter; Tnos, nopaline synthase terminator; T35S, CaMV 35S terminator; TAg7, Agl terminator from Agrobacterium tumefaciens; nptll, kanamycin resistance gene; hpt, hygromycin resistance gene; yfp, yellow fluorescent protein gene, b, Schematic drawing of the grafted stem regions, the sectioning of the graft site (horizontal lines) and the selection for gene transfer. The two transgenic lines are coloured according to the antibiotic resistance gene they carry (see a). The Petri dish is split in three parts, with the left two quarters containing leaf and stem sections from the two graft partners and the right half containing sectioned graft sites. c, Selection of callus resistant to kanamycin and hygromycin on medium containing both antibiotics. White arrows indicate green resistant tissue. d, Presence of all three transgenes in doubly resistant lines (nuclear gene transfer, NGT lines). PCR assays were performed with wild-type tobacco (Nt-wt), the two graft partners (Nt-kan:yfp, Nt-hyg) and three independently generated NGT lines (see Methods), e, Analysis of YFP fluorescence by ultraviolet microscopy. Note the differences in cell size between Nt-kan:yfp plants and NGT plants. Scale bar, 100 urn. f, Phenotype of NGT plants. Growth of NGT seedlings is delayed compared to wild-type seedlings (left panel, picture taken 23 days after sowing). Comparison between a wild-type leaf and a leaf from an NGT plant reveals enhanced pigmentation of NGT leaves (right panel). ^ax-Planck-lnstitut für Molekulare Pflanzenphysiologie, Am Mühlenberg 1, D-14476 Potsdam-Golm, Germany, department of Molecular Biology, Institute of Biotechnology, John Paul II Catholic University of Lublin, Konstantynow II, 20-708 Lublin, Poland. *These authors contributed equally to this work. 23 2 I NATURE I VOL 511 I 10 IULY 2014 ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH the hygromycin resistance gene hpt (Fig. la). The two lines were grafted onto each other and, after fusion of scion and stock had occurred, the graft site was excised and subjected to double selection for kanamycin and hygromycin resistance (Fig. lb). Although tissue explants from the two graft partners cannot grow on this medium and bleach out (Fig. lc), surprisingly, the formation of resistant calli from the graft sites was observed frequently (29 independent events selected from 13 grafts). Doubly resistant tissue readily regenerated into shoots, which were tested for the presence of the two resistance genes and the yfp reporter gene (Fig. Id). PCR assays detected all three transgenes in the selected lines (referred to as NGT lines, for nuclear gene transfer), tentatively suggesting that nuclear gene transfer across the graft junction had occurred (Fig. Id). As a preliminary investigation to distinguish between transfer of individual genes and transfer of entire nuclear genomes, the macroscopic and microscopic phenotypes of the regenerated plants were examined. Cell size is correlated to nuclear DNA content21 and, therefore, polyploid plants often have larger cells than their diploid ancestors. Indeed, NGT plants had larger cells than N. tabacum (Fig. le), indicating that they contain significantly more DNA. When NGT plants were grown in the greenhouse, they were smaller than wild-type plants and exhibited more intense leaf pigmentation (Fig. If). NGT plants were fertile and produced seeds from which the next generation was raised to determine chromosome number and DNA content in individual seedlings. The chromosome set of N. tabacum comprises 48 chromosomes. Although 48 chromosomes were present in Nt-kan:yfp and Nt-hyg plants, all Ti progeny of the NGT lines displayed increased chromosome numbers (Extended Data Fig. 1). The numbers were variable, but a large fraction of the plants contained 96 chromosomes. Hence, the NGT lines are auto-polyploid (tetraploid or, considering the allopolyploid origin of Nico-tiana tabacum, autoallooctoploid). Newly formed autopolyploids may be meiotically unstable and experience the loss of individual chromosomes due to missegregation4, potentially explaining why some NGT progeny had fewer than 96 chromosomes. Indeed, when meiotic chromosome segregation was analysed in pollen mother cells, aberrant segregation patterns were observed (Extended Data Fig. 2). Consistent with occasional chromosome loss, the Fi progeny displayed considerable phe-notypic variation (Extended Data Fig. 3). Measurements of nuclear DNA content by flow cytometry confirmed the tetraploid status of NGT plants (and the triploid status of progeny from a cross with the wild type; Extended Data Fig. lc). Tetraploidy was further confirmed by inheritance tests analysing the segregation of the two antibiotic resistances (Extended Data Table 1). Grafting and tissue culture procedures can occasionally induce spontaneous polyploidization or formation of chimaeras. Although our inheritance analyses excluded the possibility of NGT plants being chimaeras composed of Nt-kan:yfp cells and Nt-hyg cells, the data do not entirely rule out the remote possibility that the resistance gene was horizontally transferred and the genome duplication occurred independently through spontaneous autopolyploidization in all NGT lines. To definitively prove that entire nuclear genomes are horizontally transferred and to examine whether this transfer provides an asexual mechanism by which new allopolyploid species can form, grafting experiments with two different plant species were performed using the tree tobacco, Nicotiana glauca (a woody species), and the cigarette tobacco, Nicotiana tabacum (a herbaceous species). N. glauca was equipped with the nptll gene, whereas N. tabacum harboured the hpt gene (Fig. la). Grafting experiments were performed both in vitro15 and in the greenhouse, mimicking natural stem grafting as occurring in 'kissing trees' (Fig. 2a). From a total of twelve grafted plants, 45 doubly resistant lines were obtained (Fig. 2b) and regenerated into plantlets (Fig. 2c). DNA content measurements revealed that the plants have genome sizes equalling the sum of the N. glauca and N. tabacum genomes (Fig. 2d). Analysis of molecular markers confirmed the presence of genetic material from both species (Fig. 3), indicating that allopolyploid plants had indeed been obtained through grafting. As plants combining the genomes of N. tabacum and N. glauca potentially represent a new species, we tentatively named them Nicotiana tabauca. Figure 2 Grafting and horizontal transfer of nuclear DNA between different species, a, A natural stem graft between a beech (front) and a maple (back) in a forest near Monroe, New Jersey (left picture), and an similar stem graft between the tree tobacco, Nicotiana glauca (left), and the cigarette tobacco Nicotiana tabacum (right) in the greenhouse, b, Selection for horizontal transfer of nuclear DNA between plants by exposing graft sites to selection for double resistance to hygromycin and kanamycin. Control explants from the two graft partners (left) and explants from four graft sites (right; framed in white) are shown (compare with Figure la-c). c, Growth of plantlets regenerated from grafts between N. tabacum and N. glauca (here termed Nicotiana tabauca) on synthetic medium, d, Genome size determination in N. tabauca by flow cytometry (compare with Extended Data Fig. lc). SI: Solanum lycopersicum (tomato) standard; Ntca: Nicotiana tabauca. Note that the calculated genome size for N. tabauca equals the sum of the genome sizes of N. tabacum (Nt-hyg) andN. glauca (Ng-kan). N. tabauca plants combine many traits from their two progenitor species, for example, perennial growth and axillary anthocyanin accumulation from N. glauca and the lighter green leaf pigmentation from N. tabacum. Other traits are intermediate, such as leaf shape and flower morphology (Fig. 4a, b). Under greenhouse conditions, N. tabauca overgrows its two progenitor species (Fig. 4a), possibly indicating that it could outcompete them. The fitness advantage of many natural allopolyploids and the superior yield of allopolyploid crops have been linked to heterosis-like effects1-5. N. tabauca plants are fertile and produce fertile progeny (Fig. 3b), thus fulfilling all criteria of a new species. Inheritance assays indicate that N. tabauca behaves like a true amphipolyploid in that the genomes of the two parental species are sufficiently different to largely prevent pairing of N. glauca chromosomes with (homeologous) N. tabacum chromosomes. Consequently, nuclear marker genes show disomic inheritance and segregate like in a diploid (Fig. 3b). Interestingly, the phenotypic variation in the progeny of N. tabauca plants was less pronounced than in the autotetraploid NGT lines (Extended Data Fig. 4). None of 24 analysed N. tabauca progeny showed a reduction in nuclear DNA content that would be detectable with the (limited) 10 JULY 2014 I VOL 511 I NATURE I 233 ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH nptll lpt2 Figure 3 Presence of nuclear genes from both Nicotians tabacum and Nicotiana glauca in Nicotiana tabauca and their Mendelian inheritance. a, Wild-type N. tabacum (Nt-wt), wild-type N. glauca (Ng-wt), the two graft partners (Nt-hyg, Ng-kan) and three independently generated N. tabauca lines (Ntca-1, 2 and 3) were analysed by PCR using species-specific primers for endogenous genes and transgenes (see Extended Data Table 2 and Methods). PT30188 is a polymorphic microsatellite marker. Amplification products of the qpt2 gene were digested with a restriction enzyme (for details and fragments sizes, see Methods), b, Segregation analysis of N. tabauca in comparison to its two parental species, N. glauca (Ng-kan) andN. tabacum (Nt-hyg). Segregation of the two antibiotic resistances is shown for the two graft partners (a heterozygous Ng-kan and a homozygous Nt-hyg) and a selfed N. tabauca line. Segregation ratios are given as numbers of antibiotic-resistant to antibiotic-sensitive seedlings. Note that the transgenes show disomic inheritance and segregate like in a diploid. Consequently, as the N. tabacum parent was homozygous for hpt, no hygromycin-sensitive progeny were obtained from selfed N. tabauca. sensitivity of flow cytometry. However, karyotype analyses in root mer-istems revealed, in addition to plants with the full allopolyploid set of chromosomes (Fig. 4c), also cases of chromosome loss (Extended Data Fig. 5), suggesting that there is some somatic genome instability, at least in rapidly dividing root tip cells. Thus, full evolutionary stabilization of the new allopolyploid species will require further genomic adaptations, which occur rapidly after polyploidization events22-23. Cases of chromosome loss and genome instability have been documented in both recently evolved natural allopolyploids and synthetic allopolyploids24,25, indicating that genomic imbalances are a common consequence of allopolyploidization. The movement of entire nuclear genomes from cell to cell and across graft junctions raises questions about the underlying transfer mechanism. Two mechanisms appear plausible: (1) fusion of neighbouring cells at the graft site, or (2) migration of nuclei from cell to cell through plasmodes-mata in a cytomixis-like process, as observed in microscopic studies1820. To obtain insights into the transfer mechanism, we investigated the genomes of plastids and mitochondria. These organelles occur at high numbers per cell and, if cell fusion was involved, one might expect N. tabauca plants to harbour a mixed population of organelles from N. glauca and N. tabacum, a genetic heterogeneity known as heteroplasmy. Analysis of several independently generated N. tabauca plants revealed that some contained the N. glauca plastid genome, whereas others contained the N. tabacum plastid genome, but none was heteroplasmic (Extended Data Fig. 6), arguing against cell fusion as the mechanism underlying genome transfer. By contrast, some heteroplasmy was detected for the mitochondrial genome (Extended Data Figs 7 and 8). However, as mitochondria are often tightly associated with the nuclear membrane and sometimes even found within the nucleus26, this finding does not necessarily provide evidence against cytomixis. Alternatively, segregation of plastids could occur faster than segregation of mitochondria. b N. tabacum N. glauca N. tabauca N. tabacum N. glauca N. tabauca N, glauca & I 48 chromosomes 24 chromosomes 72 chromosomes Figure 4 Phenotype and karyotype of Nicotiana tabauca plants. a, Vegetative phenotypes of N. tabacum, N. glauca andN. tabauca. Plants in the upper panel were photographed 44 days after transfer to soil. Note the difference in leaf colour (blue-green in N. glauca and lighter green in N. tabacum and N. tabauca). The middle panel shows the difference in anthocyanin accumulation. N. glauca has accumulation visible in the whole leaf; N. tabauca has accumulation restricted to the petiole base; N. tabacum has no anthocyanin. The lower panel depicts the difference in leaf morphology. b, Comparison of flower morphology, c, Karyotype analysis by DAPI (4',6-diamidino-2-phenylindole) staining reveals the allopolyploid status of N. tabauca. The initial interspecies genome transfer in a natural graft represents a somatic event. However, it can become heritable by lateral shoot formation from the graft site, which is facilitated by the presence of a meristem-atic ring surrounding the stem, the cambium12. Lateral shoot formation 23 4 I NATURE I VOL 511 I 10 IULY 2014 ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH is commonly induced by wounding, for example, upon herbivory or grafting1014. This would allow the allopolyploid cell to enter the germline by becoming part of a newly formed apical meristem. As most modern crop plants arose through allopolyploidization, the possibility to generate new allopolyploid species by grafting-mediated genome transfer also provides a new tool for crop improvement. Although genome combination can also be achieved synthetically using protoplast fusion techniques27-28, the procedures involved are technically demanding and available only for a limited number of species. In cases where hybridization is possible, the genomes of two species can also be combined sexually, but the resulting hybrids are often sterile unless their chromosome sets are doubled. This is also the case for most interspecific hybrids between Nicotiana species, including N. tabacum and N. glauca29. Finally, genome transfer by grafting creates new combinations of plastid and mitochondrial genomes (Extended Data Figs 6-8). This is important because the organellar genotypes influence important agronomic traits, including growth and stress tolerance30. We have demonstrated that grafting results in the transfer of entire nuclear genomes between species. The significance of this is twofold. First, natural grafting provides a potential asexual mechanism of speciation that is much less restricted by incompatibilities than sexual hybridization914 (Fig. 2a). Thus, grafting should be considered as an alternative mechanism of polyploidization, especially in allopolyploids that formed from distantly related species. Second, interspecies grafting provides a method to produce new allopolyploid crop species. As polyploidy confers the superior properties of modern crops over their diploid progenitor species6-7, this has significant potential in breeding and agricultural biotechnology. METHODS SUMMARY Nicotiana plants were grown under greenhouse conditions or under aseptic conditions on synthetic medium. Transgenic plants were produced by Agrobacterium-mediated transformation. Interspecies grafting experiments were performed reciprocally by transplanting N. glauca scions onto N. tabacum stocks and vice versa. Selection for gene transfer was done on regeneration medium containing kanamycin and hygro-mycin. Regenerated doubly resistant shoots were rooted on synthetic medium, transferred to soil and grown to maturity in the greenhouse. For molecular analyses, microscopy, karyotyping and flow cytometry, see the Methods section. Online Content Any additional Methods, Extended Data display items and Source Data are available in the online version of the paper; references unique to these sections appear only in the online paper. Received 25 November 2013; accepted 31 March 2014. Published online 8 June 2014. 1. Comai, L. The advantages and disadvantages of being polyploid. Nature Rev. Genet. 6,836-846 (2005). 2. Sobel, J. M., Chen, G. F., Watt, L. R. & Schemske, D. W. The biology of speciation. Evolution 64, 295-315 (2010). 3. Madlung, A. Polyploidy and its effect on evolutionary success: old questions revisited with new tools. Heredity 110,99-104 (2013). 4. Leitch, A. R. & Leitch, I. J. Genomic plasticity and the diversity of polyploid plants. Science 320, 481-483 (2008). 5. Paterson, A. H. Polyploidy, evolutionary opportunity, and crop adaptation. Genetica 123,191-196(2005). 6. Hegarty, M. J. & Hiscock, S. J. Genomic clues to the evolutionary success of polyploid plants. Curr. Biol. 18, R435-R444 (2008). 7. Soltis, P. S. & Soltis, D. E. The role of hybridization in plant speciation. Annu. Rev. Plant Biol. 60, 561-588 (2009). 8. Joly, S., Heenan, P. B.& Lockhart, P. J. A Pleistocene inter-tribal allopolyploidization event precedes the species radiation of Pachycladon (Brassicaceae) in New Zealand. Mol. Phylogenet Evol. 51,365-372 (2009). 9. Seidel, C. F. Ueber Verwachsungen von Stämmen und Zweigen von Holzgewächsen und ihren Einfluss auf das Dickenwachsthum der betreffenden Theile. Naturwiss. Ges. Isis Dresden Sitzber. 161-168 (1879). 10. Küster, E. Über Stammverwachsungen. Jahrb. Wiss. Bot. 33,487-512 (1899). 11. Beddie, A. D. Natural root grafts in New Zealand trees. Transact. Proc. R. Soc. New Zeal. 71,199-203(1942). 12. Larson, P. R. The vascularcambium:developmentand structure. Springer Series in Wood Science 1-725 (Springer-Verlag, 1994). 13. Bock, R. The give-and-take of DNA: horizontal gene transfer in plants. Trends Plant Sei. 15,11-22 (2010). 14. Mudge, K., Janick, J., Scofield, S. & Goldschmidt, E. E. A history of grafting. Hortic. Rev. (Am. Soc. Hortic. Sei.) 35,437-493 (2009). 15. Stegemann, S. & Bock, R. Exchange of genetic material between cells in plant tissue grafts. Science 324, 649-651 (2009). 16. Stegemann, S., Keuthe, M., Greiner, S. & Bock, R. Horizontal transfer of chloroplast genomes between plant species. Proc. Natl Acad. Sei. USA 109, 2434-2438 (2012) . 17. Thyssen, G., Svab, Z. & Maliga, P. Cell-to-cell movement of plastids in plants. Proc. Natl Acad. Sei. USA 109, 2439-2443 (2012). 18. Zhang, W. C, Yan, W. M. & Lou, C. H. Intercellular movement of protoplasm in vivo in developing endosperm of wheat caryopses. Protoplasma 153, 193-203 (1990). 19. Zhang, W.-C. Progress in research on intercellular movement of protoplasm in higher plants. Acta Bot. Sin. 44, 1068-1074 (2002). 20. Mursalimov, S. R., Sidorchuk, Y. V. & Deineko, E. V. New insights into cytomixis: specific cellular features and prevalence in higher plants. Planta 238, 415-423 (2013) . 21. Melaragno, J. E., Mehrotra, B. & Coleman, A. W. Relationship between endopolyploidy and cell size in epidermal tissue of Arabidopsis. Plant Cell 5, 1661-1668(1993). 22. Madlung, A. etal. Genomicchanges in synthetic Arabidopsis polyploids. PlantJ. 41, 221-230 (2005). 23. Pontes, 0. etal. Chromosomal locus rearrangements are a rapid response to formation of the allotetraploid Arabidopsis suecica genome. Proc. Natl Acad. Sei. USA 101,18240-18245 (2004). 24. Chester, M. et al. Extensive chromosomal variation in a recently formed natural allopolyploid species, Tragopogon miscellus (Asteraceae). Proc. Natl Acad. Sei. USA 109,1176-1181 (2012). 25. Xiong, Z., Gaeta, R. T. & Pires, J. C. Homoeologous shuffling and chromosome compensation maintain genome balance in resynthesized allopolyploid Brassica napus. Proc. Natl Acad. Sei. USA 108, 7908-7913 (2011). 26. Yu, H.-S. & Russell, S. D. Occurrence of mitochondria in the nuclei of tobacco sperm cells. Plant Cell 6, 1477-1484 (1994). 27. Shepard, J. F., Bidney, D., Barsby, T. & Kemble, R. Genetic transfer in plants through interspecific protoplast fusion. Science 219, 683-688 (1983). 28. Evans, D. A., Wetter, L. R.& Gam borg, 0. L. Somatic hybrid plants of Nicotiana glauca and Nicotiana tabacum obtained by protoplast fusion. Physiol. Plant. 48,225-230 (1980). 29. Trojak-Goluch, A. & Berbec, A. Cytological investigations of the interspecific hybrids of Nicotiana tabacum L. X N. gla uca G ra h. J. Appl. Genet. 44,45-54 (2003). 30. Greiner, S. & Bock, R. Tuninga menage ä trois: co-evolution and co-adaptation of nuclear and organellar genomes in plants. Bioessays 35,354-365 (2013). Acknowledgements We thank the MPI-MP Green Team for help with plant transformation, K. Köhl for providing the hptvector and S. Ruf (MPI-MP) for discussions. We are grateful to J. Fuchs (IPK Gatersleben) for advice on flow cytometry measurements. This research was financed by the Max Planck Society. Author Contributions I.F., S.S. and H.G. performed the experiments. All authors participated in data evaluation and experimental design. R.B. conceived the study, and wrote the manuscript. Author Information Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome to comment on the online version of the paper. Correspondence and requests for materials should be addressed to R.B. (rbock@mpimp-golm.mpg.de). 10 JULY 2014 I VOL 511 I NATURE I 235 ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH METHODS Plant material and growth conditions. Sterile Nicotiana tabacum cv. SNN and Nicotiana glauca plants were grown under aseptic conditions by germinating surface-sterilized seeds in Petri dishes with MS medium31 supplemented with 3% sucrose and transferring seedlings into magenta boxes containing the same medium. Plants were raised in controlled environment chambers under a diurnal cycle of 16 h light (50uEm 2s l) at 24 °C and 8 h dark at 22 °C. For cultivation in soil, plants were either germinated in soil or transferred from tissue culture into soil and grown under standard greenhouse conditions in a diurnal cycle of 16 h light at25°Cand8h darkness at 20 °C. Generation of transgenic plants. The vector containing the nptll aadyfp expression cassettes and the generation of transgenic N. tabacum and N. glauca plants have been described previously16. A binary vector containing the hygromycin phosphotransferase {hpt) cassette (Fig. la) was provided by Dr. Karin Kohl (MPI-MP). Nuclear transformation was performed by Agrobacterium tumefaciens mediated leaf-disc transformation using standard protocols. Nuclear transgenic lines were selected on antibiotic-containing MS-based regeneration medium, rooted in sterile boxes with MS medium containing 2% sucrose and then transferred to the greenhouse. Grafting and selection for horizontal gene transfer. Grafting experiments were performed with sterile plants under aseptic conditions or with plants in the greenhouse. For grafting under aseptic conditions, plants were raised on MS medium containing 3% sucrose. Plants with a similar stem diameter were selected as graft partners. Stems were cut at an angle of approximately 45°, the graft partners were joined and the scion was fixed with a sleeve produced from a silicon tube. Grafting experiments were performed reciprocally in that each plant was used as both stock and scion, giving rise to two grafts. Successful grafting was evidenced by establishment of a physical connection between scion and stock and continued growth of the scion. After 2-4 weeks, the graft site was excised, cut with a sharp scalpel in cross sections and exposed to regeneration medium containing 250 or 500 mg 1 1 kanamycin and 50 or 150 mgl 1 hygromycin. Selection was performed at 50 uE m 2s 1 under 16 h light at 24 °C and 8 h dark at 22 °C. Grafting in the greenhouse was performed by attaching surface-wounded stems to each other (by fixing them with tape), to mimic the natural grafting process (Fig. 3a). Following establishment of the graft, the graft site was surface-sterilized with Plant Preservative Mixture (Plant Cell Technology) and exposed to selection for DNA transfer as described above. Preparation of nuclei from plants tissues. Young leaves from N. tabacum, N. gla uca, N. tabauca and S. lycopersicum were collected and nuclei were prepared according to published protocols32,33. To determine the absolute genome size of the material, S. lycopersicum leaves were used as an internal standard. Immediately after removing the leaves from the plant, leaf pieces (sample plus internal standard in a 1:1 ratio) were chopped with a disposable blade in ice-cold Galbraith's buffer32 with 10 mM DTT and supplemented with 50 u.g ml 1 of RNase A solution and 50 ug ml 1 propidium iodide solution. The suspension of cell nuclei was then filtered through a 20 urn filter (CellTrics, Partec, Germany) and kept on ice in the dark for 10 min before determining the DNA content. Measurement of nuclear DNA contents. DNA contents were determined by flow cytometry using the FACSAria II cell sorter (BD Bioscience). Propidium iodide fluorescence was measured using a blue laser (488 nm), a 616/23 nm band-pass filter and a 610 LP mirror. A minimum of 2,000 gated nuclei per sample were recorded. DNA content was calculated based on the ratio between the mean value for the 4C peak of the internal standard (S. lycopersicum, In = 2X =24 chromosomes = 2C = 2.05 pg DNA) and the mean value of the 2C sample peak {N. tabacum, In = 4X = 48 chromosomes = 2C = 10.35 pg DNA; N. glauca, 2n = 2X =48 chromosomes = 2C = 10.65 pg DNA). C-values were obtained from (http://dataiew.org/cvalues/). Changes in the ploidy level measured by flow cytometry were further verified by chromosome counting (see below). Chromosome staining in pollen mother cells (PMCs). Aceto orcein solution (2% w/v orcein in 45% v/v acetic acid) and Vectashield Mounting Medium with DAPI (Linaris B.P., Germany) were used to stain chromosomes from meiotic PMCs. Flower buds were fixed overnight in ethanolracetic acid (3:1) solution at room temperature. Afterwards, the buds were washed with H20 for 10 min, followed by incubation in 1M HC1 at 60 °C for 8 min. The material was then rinsed three times with H20 and the anthers were placed onto a glass slide. One drop of aceto-orcein solution or Vectashield Mounting Medium was added and the anthers were covered by a coverslip. PMCs were released from the anthers by gently squashing the sample with the forefinger. PMCs with meiotic chromosomes were analysed by light microscopy or ultraviolet (UV) microscopy. Preparation of mitotic chromosomes. For mitotic chromosome preparations, root tips from freshly germinated seedlings were collected, transferred to a saturated solution of 1-bromonaphthalene and incubated at 4°C in the dark for 24 h. The samples were then washed several times with distilled water, transferred to freshly prepared fixative solution (3:1 v/v mix of ethanol and glacial acetic acid) and softened by enzymatic maceration (5% pectinase, 2% cellulase (Sigma-Aldrich) and 1% cel-lulase Onozuka RS (Serva), dissolved in citric acid/sodium citrate buffer, pH 4.6) for 15 min at 37 °C or, alternatively, by incubation in 1N HC1 at 60 °C for 10 min. Afterwards, the tissue was washed several times with distilled water and gently squashed under a coverslip in a drop of 45% acetic acid. The samples were shortly frozen in liquid nitrogen, the cover glass was removed and the tissue was air dried. Preparations from enzymatically digested material were mounted in DAPI (containing antifade solution (Vectashield, Vector Laboratories)). Those obtained by hot HC1 hydrolysis were stained with a drop of 0.1% aqueous toluidine blue solution, washed in distilled water, air-dried and mounted in Entellan (Merck). Metaphase plates were subsequently analysed by light microscopy or epifluorescence microscopy. Isolation of plant genomic DNA, polymerase chain reactions (PCR) and DNA sequencing. Genomic DNA was extracted from fresh leaf tissue using a CTAB-based method34. PCR amplification was performed with synthetic oligonucleotides as primers (Extended Data Table 2) and GoTaq Flexi DNA polymerase (Promega) in a Mastercycler EPgradient (Eppendorf Germany). The standard reaction involved an initial denaturation step at 94 °C for 1 min, followed by 30 cycles of 30 s at 94 °C, 30 s at 50 to 60°C and 60 to 90 s at 72 °C, and a final elongation for 2 min at 72 °C PCR analysis of NGT lines was performed using the following oligonucleotide combinations: PHygFor + PHygRev for hptll (658 bp), PnptllFor + PnptllRev for nptll (795 bp) and P35SInternRev + PyfpRev ioryjp (772 bp; Extended Data Table 2). PCR analysis of Nicotiana tabauca plants was performed using the following oligonucleotide combinations (Extended Data Table 2): PHygFor + PHygRev for hptll (658 bp), PnptllFor + PnptllRev for nptll (795 bp), P35SInternRev + PyfpRev ioryjp (772 bp), PT30188F + PT30188R for a polymorphic microsatellite marker35 (156 bp in N. tabacum and 106 bp in N. glauca) and Pqpt2F + PqptR for the quinolinate phosphori-bosyltransferase 2 gene {qpt2). The qpt2 gene occurs in two copies in N. tabacum (yielding PCR fragments of 1,121 bp and 1,172 bp, respectively) and one copy in N. glauca (giving rise to a 1,162 bp product). The amplification products were digested with the restriction enzyme Dral resulting in fragments of 199 bp and 922 bp (for the N. tabacum 1,121 bp PCR product); 200 bp, 289 bp and 683 bp (for the N. tabacum 1,172bp PCR product); and 54bp, 199bp, 295bp and 614bp (for the N. glauca 1,162 bp PCR product; note that the 54 bp fragment is not visible in the gel in Fig. 3a because of its small size). The fragments were separated on 4% agarose gels. For DNA sequencing, amplified PCR products were purified by agarose gel electrophoresis and recovered from excised gel slices with the NucleoSpin Extract II kit (Macherey-Nagel). Purified PCR products were sequenced with specific primers using the chain termination method (Eurofins MWG Operon). Crosses and inheritance tests. Crosses were performed by hand-pollinating emasculated plants grown in the greenhouse. To accelerate flowering of the tree tobacco (N. glauca) and the tree-like new allopolyploid species N. tabauca, plants were kept in small pots. Inheritance assays were performed by sowing surface-sterilized seeds on MS medium containing the appropriate antibiotics. Kanamycin was used at a concentration of 250 ug ml l, hygromycin at 25 u.g ml l. 31. Murashige, T. & Skoog, F. A revised medium for rapid growth and bio assays with tobacco tissue culture. Physiol. Plant. 15, 473^197 (1962). 32. Galbraith, D. W. etal. Rapid flowcytometric analysis of the cell cycle in intact plant tissues. Science 220, 1049-1051 (1983). 33. Doležel, J., Greilhuber, J. & Suda, J. Estimation of nuclear DNA content in plants using flow cytometry. Nature Protocols 2, 2233-2244 (2007). 34. Doyle, J. J. & Doyle, J. L. Isolation of plant DNA from fresh tissue. Focus 12,13-15 (1990). 35. Moon, H. S., Nicholson, J. S. & Lewis, R. S. Use of transferable Nicotiana tabacum L. microsatellite markers for investigating genetic diversity in the genus Nicotiana. Genome 51, 547-559 (2008). 36. Shinozaki, K. etal. The complete nucleotide sequence of the tobacco chloroplast genome: its gene organization and expression. EMBOJ. 5,2043-2049 (1986). 37. Sugiyama, Y. ef al. The complete nucleotide sequence and multipartite organization of the tobacco mitochondrial genome: comparative analysis of mitochondrial genomes in higher plants. Mol. Genet. Genomics 272, 603-615 (2005). ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH Nt-kan:yfp (48 Chr.) Nt-hyg (48 Chr.) NGT-2-1 (60 Chr.) NGT-2-2 (76 Chr.) NGT-2-3 (84 Chr.) NGT-2-4 (96 Chr.) 30 w OJ w ra o. 20 re 5 1& I Nt-kan:yfp I Nt-hyg INGT J_l_I I....... ill 48 60 72 Number of chromosomes per cell 84 ._ 300 150 ja SI |z390 300 150 50 100 150 Relative PI fluorescence (x1000) 50 100 150 Relative PI fluorescence (x1000) 600' 300- J3 E SI u 250 ■S 150 SI i 4.939 A 50 100 150 Relative PI fluorescence (x1000) Extended Data Figure 1 Autotetraploidy and chromosome loss in NGT plants, a, Chromosome preparations of mitotic cells. Along with preparations from each of the two graft partners (Nt-kan:yfp and Nt-hyg), four examples of mitotic cells from four individual progeny plants of the self-pollinated line NGT-2 are shown. Chromosomes are visualized by hot tissue hydrolysis in HC1 and staining with toluidine blue. Chromosome numbers are given in parentheses. Chr, chromosomes, b, Chromosome counts for individual seedlings from the two graft partners and the selfed line NGT-2. Mitotic cells from the root tips were analysed. The total number of metaphases investigated was 32 for Nt-kan:yfp (blue), 35 for Nt-hyg (purple) and 86 for NGT-2 (pink). 50 100 150 Relative PI fluorescence (x1000) c, Absolute genome size determination by flow cytometry. Leaf samples were mixed with tomato leaves (Solatium lycopersicum, ST) that served as internal standard, nuclei were isolated and the relative fluorescence intensity of propidium iodide (PI) was measured. Each peak corresponds to a population of nuclei. For each sample, the ratio between the peak of the analysed tobacco line and the peak of the internal standard (4C tomato nuclei) was calculated to determine the absolute genome size. Sample 1: Nt-hyg; sample 2: Nt-kan:yfp; sample 3: seedling from a cross between NGT-1 and wild type; sample 4: seedling from a self-pollinated NGT-2 plant. ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH Extended Data Figure 2 Meiotic chromosome missegregation in allopolyploid NGT tobacco lines, a-c,Young flower buds were fixed and stained, either with aceto-orceine (a, b) or with DAPI (c). The anthers were collected and the pollen mother cells (PMC) were examined under the microscope. Two plants from the Fl generation of self-pollinated NGT plants were used, a, Four representative PMCs of a wild-type tobacco plant. Scale bar, 50 urn. b, Six PMC from NGT plants. The upper three cells are from line NGT-2, the lower three from line NGT-3. Scale bar, 20 um. c, Two PMCs from line NGT-3 stained with DAPI. Scale bar, 5 urn. Mis-segregating chromosomes are indicated by arrowheads in b and c. ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH NGT progeny Nt-wt V la Nt-hyg im DD El Nt-kan:yfp * * •f b NGT progeny Nt-wt a □ n Nt-hyg D n Nt-kan:yfp * NGT-2 NGT-3 NGT-1 NGT-2 NGT-3 Extended Data Figure 3 Phenotypes of NGT progeny plants, a, b, Wild-type tobacco (Nt-wt), the two transgenic graft partners (Nt-hyg and Nt-kan:yfp) and the second generation of three NGT lines (NGT-1, NGT-2 and NGT-3) were grown under greenhouse conditions. A total of 21 NGT progeny plants were investigated, 14 of them resulted from self-pollinated lines (NGT-2 and NGT-3) and the remaining 7 from the cross of the NGT-1 line with a wild-type plant (which was male sterile and could not be selfed). Pictures were taken 30 (a) and 45 (b) days after sowing. ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH Nicotiana tabauca F1 progeny Nt-wl Nt-hyg Ng-wl Ng-kan Nt-wl Nt-hyg Ng-wl Ng-kan ' J^jgjf 'i W " ^ f. V<4l *• 0m Nicotiana tabauca F1 progeny □ □ □ □ Pi • * □ □ ft E3 Extended Data Figure 4 | Phenotypes of Nicotiana tabauca progeny plants. and grown under greenhouse conditions. Pictures were taken 28 (a) and 47 a, b, Wild-type tobacco (Nt-wt), the two transgenic graft partners (Nt-hyg (b) days after sowing, and Ng-kan) and the Fl generation of an N. tabauca line were raised from seeds ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH 72 chromosomes Extended Data Figure 5 | Detection of allopolyploid and aberrant on hot tissue hydrolysis in HC1 and staining with toluidine blue was used (see karyotypes in Nicotiana tabauca Ft progeny plants, a, Two examples of Methods). Shown is a Nicotiana tabauca Fl plant with the full allopolyploid DAPI-stained metaphases with fewer than 72 chromosomes, b, As an chromosome set of 72 chromosomes, alternative to the D API staining shown in Fig. 4c and in panel a, a method based ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH Polymorphism Plant line ptl pt2 pt3 pt4 Nt-wt Nt Nt Nt Nt Nt-hyg Nt Nt Nt Nt Ng-wt Ng Ng Ng Ng Ng-kan Ng Ng Ng Ng Ntca-1 Nt Nt Nt Nt Ntca-2A Ng Ng Ng Ng Ntca-2B Ng Ng Ng Ng Ntca-3A Ng Ng Ng Ng Ntca-3B Ng Ng Ng Ng Ntca-4A Ng Ng Ng Ng Ntca-4B Ng Ng Ng Ng Ntca-5A Ng Ng Ng Ng Ntca-5B Ng Ng Ng Ng Extended Data Figure 6 Molecular analysis of four polymorphic regions in the plastid genome of Nicotiana tabauca lines, a, Physical map of the Nicotiana tabacum plastid genome showing the four plastid polymorphic regions analysed (ptl, pt2, pt3 and pt4). ptl: polymorphism amplified with oligonucleotides PptlF and PptlR (Extended Data Table 2) resulting in a 211 bp fragment in N. tabacum cv. SNN and a 203 bp fragment in N. glauca; pt2: polymorphism amplified with oligonucleotides Ppt2F and Ppt2R resulting in a 221 bp fragment in N. tabacum cv. SNN and a 229 bp fragment in N. glauca; pt3: polymorphic region of ~3kb (containing altogether 25 polymorphisms) amplified using the primer pairs Ppt3-1F + Ppt3-1R, 229 bp 221 bp Ppt3-2F + Ppt3-2RandPpt3-3F + Ppt3-3R; pt4: polymorphic region of ~3kb (containing altogether 32 polymorphisms) amplified with the primer pairs Ppt4-1F + Ppt4-1R, Ppt4-2F + Ppt4-2R and Ppt4-3F + Ppt4-3R. The polymorphisms were selected based on published sequence information of the plastid genomes of N. tabacum and N. glauca16'36. b, Overview of the plant lines and the polymorphic regions analysed. Nicotiana tabacum sequence is represented with Nt on yellow background, Nicotiana glauca sequence is represented with Ng on pink background, c, A 4% agarose gel showing the size difference of the PCR fragments for the ptl and pt2 polymorphisms between N. tabacum, N. glauca and five N. tabauca lines. ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH Polymorphism Plant line mtl mt2 mt3 mt4 mt5 Nt-wt Nt Nt Nt Nt Nt Nt-hyg Nt Nt Nt Nt Nt Ng-wt Ng Ng Ng Ng Ng Ng-kan Ng Ng Ng Ng Ng Ntca-1 Nt Nt Nt Nt Nt Ntca-2A Nt Nt Nt Nt Nt Ntca-2B Nt Nt Nt Nt Nt Ntca-3A Nt Ng Ng Ng Ng>Nt Ntca-3B Nt/Ng Nt Ng Ng Ng>Nt Ntca-4A Nt Nt Nt Nt Ng>Nt Ntca-4B Nt Nt Nt Nt Ng>Nt Ntca-5A Nt/Ng Nt Ng Ng Ng>Nt Ntca-5B Nt/Ng Nt Ng Ng Ng>Nt Extended Data Figure 7 Molecular analysis of five mitochondrial genome polymorphisms in Nicotiana tabauca lines by sequencing of amplified PCR products, a, Physical map of the Nicotiana tabacum mitochondrial genome showing the five mitochondrial polymorphic regions analysed (mtl, mt2, mt3, mt4 and mt5). mtl: polymorphic region of 693bp (containing 5 polymorphisms) amplified with oligonucleotides PmtlF and PmtlR (Extended Data Table 2); mt2: polymorphism (SNP) amplified with oligonucleotides Pmt2F and Pmt2R; mt3: polymorphism (SNP) amplified with oligonucleotides Pmt3F and Pmt3R; mt4: polymorphism (SNP) amplified with oligonucleotides Pmt4F and Pmt4R; mt5: polymorphic region amplified with oligonucleotides Pmt5F and Pmt5R and resulting in a 298 bp fragment in N. tabacum and a 288 bp fragment in N. glauca. The polymorphisms were selected based on published sequence information of the mitochondrial genomesofN. tabacum andN. glauca1637.b, Overview of the plant lines and the polymorphic regions analysed. Nicotiana tabacum sequence is represented with Nt on yellow background, Nicotiana glauca sequence is represented with Ng on dark pink background. Heteroplasmy (that is, detectability of both the N. tabacum and the N. glauca sequence in an N. tabauca plant) is indicated by Nt/Ng on light pink background. A and B denote two different plants from the same N. tabauca line. ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH 37,012 GTAGAGC 37,048 A A A C G A A 37,177 37,184 GCTAGTTTTTATAA Nt-hyg GTAAAGC A A A T G A A GCTCGTTTTTTTAA Ng-kan GTAGAGC A A A C G A A GCTCGTTTTTtTAA Ntca-5B mt4 Extended Data Figure 8 Detection of mitochondrial heteroplasmy in N. tabauca plants, a, Example of sequences amplified from the mtl polymorphic region in Nt-hyg, Ng-kan and in an N. tabauca (Ntca) line. The mtl polymorphic region was amplified with the oligonucleotide combination PmtlF and PmtlR (Extended Data Table 2) resulting in a 693 bp fragment (corresponding to nucleotide positions 36,746 to 37,438 in the Nicotiana tabacum mitochondrial genome, accession number: NC 006581.1). The fragment contains five single nucleotide polymorphisms (SNPs) that are denoted with an arrow in the sequence chromatograms and the nucleotide position in the N. tabacum mitochondrial genome. A mixed nucleotide position indicating heteroplasmy is denoted by two letters above each other, b, A 4% agarose gel analysing the mt4 polymorphism in N. tabacum, N. glauca and nine N. tabauca plants (representing five different lines). A 310bp 100 bp fragment was amplified by PCR and digested with the two restriction enzymes Ncol and BsaBI. Digestion of the PCR product amplified from the N. tabacum mitochondrial genome yields three restriction fragments (25bp, 185bp and 100 bp), whereas digestion of the PCR product amplified from the N. glauca mitochondrial genome results in two fragments (25 bp and 285 bp). This difference between the two species is due to a T to G substitution in N. glauca (relative to N. tabacum; position in the N. tabacum mitochondrial genome: 305,402), resulting in the loss of a BsaBI site. The 25 bp restriction fragment is not detectable in the gel because of its small size. Note that the low level of heteroplasmy detected by restriction-fragment length polymorphism (RFLP) analysis was not reliably detectable by DNA sequencing (compare with Extended Data Fig. 7b). A and B denote two different plants from the same N. tabauca line. ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH Extended Data Table 1 | Segregation ratios of three autopolyploid NGT lines Plant line Self-pollinated Cross NGT xwt Hr: Hs Kr: Ks HKr: HKs Hr: Hs Kr: Ks HKr: HKs NGT-1 326 : 45 (87.9%) 154 : 127 (54.8%) 195 : 223 (46.7%)* NGT-2 396:15 (96.4%) 383:125 (75.4%) 317 : 130 (70.9%) 262 : 48 (84.5%) 223:121 (64.8%) 107:153 (41.2%)* NGT-3 36 : 2 (94.7%) 55 : 20 (73.3%) 105 : 28 (78.9%) 165 : 32 (83.8%) 56 : 68 (45.2%) 85 : 109 (43.8%) Hr, hygromycin resistant; Hs, hygromycin sensitive; Ks, kanamycin sensitive; Kr, kanamycin resistant; HKr, resistant to hygromycin + kanamycin; HKs, sensitive to hygromycin + kanamycin; -, not analysed (due to pollen sterility); *, segregation ratio significantly different from the expected segregation of a tetraploid plant (Pearson's chi squared, P< 0.01). ©2014 Macmillan Publishers Limited. All rights reserved RESEARCH Extended Data Table 2 | List of oligonucleotides used in this study Oligonucleotide Sequence (5' to 3') PHygFor GACGTCTGTCGAGAAGTTTCTGATCG PHygRev GTATTGGGAATCCCCGAACATCGCCTC PnptllFor ATGATTGAACAAGATGGATTGCAC PnptllRev TCAGAAGAACTCGTCAAGAAGGCG P35Slntern Rev AAGGTGGCTCCTACAAATGCCAT PyfpRev GCCGTTCTTCTGCTTGTCGGCC Pqpt2F CTCTTTAAGAAGAAAAAAAGATTC PqtpR CGCGAATATCATCTCAGCAA PT30188F AACCATACGCCTTCAGATCG PT30188R TGGTTTGAGTAAAGAAATGTTGTGA Ppt1 F CTTGATCCACTTGGCTACATCC Ppt1 R GCTAATGTTACTATATCTTTTTGAT Ppt2F TTCTTTATAGGAGAGGACAAATC Ppt2R GCATAGAAATCCAATCACTAGG Ppt3-1 F CAGGTATTGTAGATATTCCCTC Ppt3-1 R AGGCGACTCCCGGATTTGAAC Ppt3-2F GAGTG GTAAG GCAGAG GAC Ppt3-2R CTAGAGTCCACTTCTTCCCC Ppt3-3F ATAGTAAGTCTTGCTTGGGC Ppt3-3R AAACAGTCAGTCAAAACGATTAA Ppt4-1 F CAATTGGCCGAAATGAATTTCTA Ppt4-1 R ATGGCCGATACTACTGGAAG Ppt4-2F TAGAGGGATGAACCCAATCC Ppt4-2R ATGTCTGGAAGCACAGGAGA Ppt4-3F GTAATGCTATGAATGACCCAGT Ppt4-3R GCCGCTAATAGAAAACCGAAATA Pmt1 F TTCGACTGAACGACGGAATTC Pmt1 R TCTTTGCTTGTTTCGTGGTATG Pmt2F GGTATTACACTATCGCGGAG Pmt2R ACTTTCGTCCCCAGTATTCTC Pmt3F AGGCGGTGATGAAGACAAAAG Pmt3R CTTTGGACTGCTTTTTTCAAAAG Pmt4F TCTTTTATGAACTCCACTGGTC Pmt4R ATG ACTAAG GTAATTG CCAATAG Pmt5F TGTAGTTTAGGGGTGCGCAG Pmt5R TTAG GG GAAACAAAG CTTCAAC ©2014 Macmillan Publishers Limited. All rights reserved