Exchange of Genetic Material Between Cells in Plant Tissue Grafts

See allHide authors and affiliations

Science  01 May 2009:
Vol. 324, Issue 5927, pp. 649-651
DOI: 10.1126/science.1170397


Tissue grafting includes applications ranging from plant breeding to animal organ transplantation. Donor and recipient are generally believed to maintain their genetic integrity, in that the grafted tissues are joined but their genetic materials do not mix. We grafted tobacco plants from two transgenic lines carrying different marker and reporter genes in different cellular compartments, the nucleus and the plastid. Analysis of the graft sites revealed the frequent occurrence of cells harboring both antibiotic resistances and both fluorescent reporters. Our data demonstrate that plant grafting can result in the exchange of genetic information via either large DNA pieces or entire plastid genomes. This observation of novel combinations of genetic material has implications for grafting techniques and also provides a possible path for horizontal gene transfer.

Grafting is widely used in plant breeding programs in order to modify plant architecture, improve vigor, or increase disease resistance. Grafting also occurs naturally; for example, when the stems or roots of trees contact each other (1). Although the grafted tissues fuse and establish vascular connections, the stock (the lower part of the graft) and scion (the upper part, usually supplying solely aerial parts to the graft) are thought not to exchange their genetic materials (2).

To test this assumption, we generated two transgenic tobacco lines carrying different selection markers and reporters. One line, Nuc-kan:yfp, harbors a kanamycin resistance gene (nptII) and the yellow fluorescent protein gene (yfp) in its nuclear genome, whereas the other, Pt-spec:gfp, possesses a spectinomycin resistance gene (aadA) and the green fluorescent protein gene (gfp) in its plastid (chloroplast) genome (Fig. 1A and fig. S1).

Fig. 1

Genetic screen for intercellular gene transfer. (A) Maps of the plastid genome in Pt-spec:gfp plants and the transgenic locus in Nuc-kan:yfp plants. PpsbA and TpsbA, promoter and terminator from the plastid psbA gene; Prrn, promoter from the plastid rRNA operon; Trps16, terminator from the plastid rps16 gene; Pnos and Tnos, promoter and terminator from the nopaline synthase gene from Agrobacterium tumefaciens; P35S and T35S, promoter and terminator from the cauliflower mosaic virus 35S transcript; LB and RB, left and right borders of the T-DNA region; Eco RV and Xho I, restriction sites used for restriction fragment length polymorphism analysis (fig. S3). (B) Selection experiments. The grafted stem region was either sectioned (horizontal lines) or directly exposed to selection (bracket). The middle panel shows the arrangement of tissue explants, the right panel a selection plate (right half, stem sections from the graft site; upper left quarter, three stem sections and three leaf explants from Nuc-kan:yfp; lower left quarter, the corresponding explants from Pt-spec:gfp). After 4 weeks on medium with spectinomycin and kanamycin, some explants from the graft site developed growing callus tissue or regenerating shoots (arrows). (C) Expression and subcellular localization of the fluorescent reporters. The wild type, the two grafting partners, and a YG line were assayed for GFP, chlorophyll (Chl), and YFP fluorescence.

We performed grafting experiments in which Pt-spec:gfp scions were grafted onto Nuc-kan:yfp stocks and vice versa. After the establishment of a physical connection, the graft site was excised and analyzed for gene flow between scion and stock by testing for the presence of cells that harbor both the kanamycin resistance gene (from Nuc-kan:yfp) and the spectinomycin resistance gene (from Pt-spec:gfp). Exposure of stem sections to double selection for kanamycin and spectinomycin resistance frequently yielded resistant calli (that is, mounds of undifferentiated cells) and regenerating shoots (Fig. 1B). Cell lines isolated from grafts with Pt-spec:gfp as scion and Nuc-kan:yfp as stock are referred to as GY; lines from the reciprocal grafting are referred to as YG.

We next assayed GY and YG lines for expression of the fluorescent reporters. All cells in the regenerated plants showed GFP fluorescence in chloroplasts and YFP fluorescence in the cytosol, indicating that the two reporter proteins were present in the same cell (Fig. 1C and fig. S2). It is known that some proteins and RNAs can travel between cells and across graft junctions (3). We therefore performed Northern blot analyses to confirm transcription of all four transgenes in GY and YG lines (Fig. 2A) and also demonstrated the presence of all transgenes at the DNA level (Fig. 2B and fig. S3).

Fig. 2

Analysis of transgenes and molecular markers in gene transfer lines. (A) Northern blot analyses of transgene expression. (B) Transgene detection at the DNA level by polymerase chain reaction (PCR). (C) PCR analysis of plastid and nuclear markers to determine the direction of intercellular gene transfer. Pt, assay of a length polymorphism in the plastid genomes of the two grafted varieties PH [Pt-spec:gfp, 125–base pair (bp) product] and SNN (Nuc-kan:yfp, 191-bp product); PH, a PH-specific nuclear marker (189-bp product); SNN, an SNN-specific nuclear marker (1569-bp product). WT, wild type; M, marker (sizes in kilobase pairs); c, buffer control.

We recovered doubly resistant lines at high frequency (94 events from 74 grafted plantlets, table S1). The frequency of intercellular gene transfer was independent of the orientation of the graft (table S1). When we subjected leaf explants and distant stem sections to selection, no resistant lines were obtained (table S2), suggesting that gene transfer is confined to the graft site and no long-distance transfer may occur. Whether the gene transfer is strictly dependent on direct cell-to-cell contact remains to be investigated.

Experiments in which whole graft sites were subjected to selection (Fig. 1B) also produced resistant cell lines, suggesting that tissue injury (by cross-sectioning) was not involved in the gene transfer process. In contrast, when stock and scion were separated before fusion, no resistant lines were obtained (table S1). Karyotype analyses excluded the possibility that YG and GY plants arose through fusion of Nuc-kan:yfp cells with Pt-spec:gfp cells (fig. S4). Finally, genetic crosses demonstrated stable inheritance of the transferred genes and additionally confirmed the absence of polyploidy (fig. S5).

Two directions of gene transfer are possible: movement of plastid genes from Pt-spec:gfp cells into Nuc-kan:yfp cells or transfer of nuclear genes from Nuc-kan:yfp to Pt-spec:gfp cells. To distinguish between these possibilities by tracking molecular markers in the plastid and nuclear genomes, we grafted two different tobacco varieties, Petit Havana (PH; Pt-spec:gfp) and Samsun NN (SNN; Nuc-kan:yfp). Analysis of a polymorphism in the plastid genomes of the two cultivars [>15 kb away from the transgene insertion site (4)] revealed that all GY and YG plants analyzed carried the PH marker (Fig. 2C), indicating that large DNA pieces or even entire plastid genomes are transferred. Polymorphisms in pathogen resistance genes were used as nuclear markers (5). All GY and YG plants tested positive for the SNN-specific marker and negative for the PH-specific marker (Fig. 2C), suggesting that the plastid transgenes moved from PH cells into SNN cells. Plant cells are connected via plasmatic bridges called plasmodesmata, but the passage of large macromolecules requires the action of specific plasmodesmata-widening proteins (3). Whether large DNA pieces or even entire organelles can travel through plasmodesmata requires further investigation.

Our discovery of grafting-mediated gene transfer further blurs the boundary between natural gene transfer and genetic engineering and suggests that grafting provides an avenue for genes to cross species barriers. Phylogenetic evidence suggests that DNA can be transferred horizontally between reproductively isolated species (6). We propose that grafting (whether natural or assisted) provides a path for horizontal gene transfer.

Finally, although our data demonstrate the exchange of genetic material between grafted plants, they do not lend support to the tenet of Lysenkoism that “graft hybridization” would be analogous to sexual hybridization. Instead, our finding that gene transfer is restricted to the contact zone between scion and stock indicates that the changes can become heritable only via lateral shoot formation from the graft site. However, there is some reported evidence for heritable alterations induced by grafting (7) and, in light of our findings, these cases certainly warrant detailed molecular investigation.

Supporting Online Material

Materials and Methods

Figs. S1 to S5

Tables S1 and S2


References and Notes

  1. We thank M. Lohse and D. Karcher for providing transgenic lines, Y. Weber for technical assistance, and D. Karcher for help with identifying markers. This research was financed by the Bundesministerium für Bildung und Forschung.

Stay Connected to Science

Navigate This Article