Horizontal gene transfer: a universal phenomenon

Gene transmission usually occurs from one generation to another (i.e., from parents to their progeny). However, horizontal gene transfer (HGT), although relatively less frequent, can also allow transmission of genetic material between phylogenetically distant species or between cytoplasmic organelles (such as chloroplastids and mitochondria) and the nucleus without the mode of parent-to-offspring inheritance. HGT is a universal phenomenon, observed in the genomes of bacteria, fungi, viruses, and eukaryotes (Bansal and Meyer 2002). However, the possible pathways for HGT in plants and the contribution of HGTs to plant genomes have been less investigated. Here we summarize what is known about HGT between plants and other organisms and systemically proposed all possible pathways for HGTs involving plants.

Numerous cases of HGT in prokaryotic genomes have been reported, and HGT appears to be more frequent in prokaryotes than in eukaryotes (Mc Ginty et al. 2011; Rolland et al. 2009; Hotopp 2011). However, with the increasing availability of large numbers of eukaryotic genome sequences, HGTs are being found with increasing frequency in eukaryotic genomes (Deschavanne et al. 2010). Within the Kingdom Fungi, HGTs have been observed between Aspergillus and Podospora (Rokas and Slot 2011), Stagonospora and Pyrenophora, and Fusarium and Alternaria (de Wit et al. 2011). Within the Animal Kingdom, HGT has been observed between insects, nematodes, amoeba (Fournier et al. 2009), and asexual animals (Hotopp 2011; Bird and Scholl 2011; Sommer et al. 2011). Between-kingdom transfers have also been observed: from fungus to aphids (Moran and Jarvik 2010), from human host to bacterial pathogen (Seifert and Anderson 2011), and from plants to silkworms (Bombyx mori) (Li et al. 2011). Genetic information transfer to plants can also occur from viruses, arthropods, fungi, nematodes, and protozoa (Liu et al. 2010).

Different mechanisms for horizontal gene transfer

In terms of genetic material involved, HGTs can be divided into three types: organelle transfers, nuclear DNA transfers, and RNA transfers. Whole organelles have been observed to be involved in HGTs between different tobacco species (Thyssen et al. 2012) and even between sexually incompatible species such as oak and birch (Sandra Stegemann et al. 2012). Also, organelle-to-nucleus HGTs have been documented within a single species such as Arabidopsis (Archibald and Richards 2010), including chloroplast-to-nucleus transfers in tobacco (Wang et al. 2011) and mitochondria-to-nucleus transfers in Arabidopsis (Bonen 2006). At the DNA level, horizontally transferred genes include sterigmatocystin, observed to have transferred between the fungi Podospora anserina and Aspergillus (Rokas and Slot 2011), DNA transposon ITmD37D (MJ1), transferred between individual mosquitos (Anopheles sinensis) (Capy et al. 2008; Diao et al. 2011), and retrotransposons, transferred between individual Drosophila (Capy et al. 2008). RNA polymerase genes from double-stranded RNA viruses (totiviruses and partitiviruses) have also been demonstrated to integrate into a number of eukaryotic genomes, including plants (Jiang et al. 2010).

In prokaryotes, we have a relatively clear understanding of a number of HGT pathways, such as transformation, conjugation, and transduction. However, HGT processes in eukaryotes are much more complicated. HGT from plant to plant can occur through either a direct pathway or a vector-mediated pathway. Direct pathways allow genes to be horizontally transferred by direct DNA exchange, i.e., plant to plant. This may occur during grafting (Stegemann and Bock 2009), symbiosis (Finan 2002), parasitism (Shirasu et al. 2010), pathogenesis (Burger and Lang 2003; Loria et al. 2002), and epiphyte or entophyte (Bock 2009) interactions. Vector-mediated transfer pathways involve the transfer of genetic material via vectors, such as pollen, fungi, bacteria, viruses, viroids, plasmids, transposons, and insects (Capy et al. 2008; Trobridge 2009; Danchin et al. 2011), with no direct contact between HGT partners (Fig. 1).

Fig. 1
figure 1

Pathways for horizontal genetic information exchange between plants and other organisms

Full size image

DNA transfer between the plastid and nuclear genomes

In plants, DNA within a single cell may be part of the mitochondrial genome, chloroplast genome, or nuclear genome. The presence of these three genomes within a cell provides three possible types of intracellular HGT: between mitochondria and the nucleus, between chloroplasts and the nucleus, and between mitochondria and chloroplasts. Each HGT type may be bidirectional: for example, both mitochondria-to-nucleus and nucleus-to-mitochondria HGTs can occur.

HGTs between the nuclear and plastid genomes have been postulated to have occurred ever since the initial endosymbiosis events by which plastids became part of the eukaryotic cell makeup (Burger and Lang 2003; Andersson 2005; Thyssen et al. 2012). In plants, the exchange of gene sequences between the chloroplast and the nucleus is very common. These gene transfers from the organelles to the host nucleus can enhance the ability of the host cell to regulate gene expression in the organelles. Functional genes transferred from the chloroplasts to the nucleus have been observed to regulate plant response to stress signaling, e.g., in Arabidopsis and rice (Cullis et al. 2009). This form of gene interaction also exists between mitochondria and the nucleus, and between chloroplasts and mitochondria (Woodson and Chory 2008).

Recent studies indicate that plant mitochondrial genomes are unusually active and remarkably adept at HGTs compared to other organelle and nuclear genomes (Archibald and Richards 2010). The mitochondrial genome in plants is composed of both native and immigrant genetic materials: for example, approximately 13.4 % of the rice mitochondrial genome came from the nuclear genome (Notsu et al. 2002), and numerous integrated mitochondrial fragments in Arabidopsis originated from alien genomes, including 16 sequences transferred from other plastids, 41 fragments from nuclear transposons or retrotransposons, and 2 fragments from fungal viruses (Marienfeld et al. 1999). Hence, mitochondrial genomes can be thought of as undergoing secondary expansion as a result of integration and propagation of external genetic material through HGTs.

The transfer of genetic material from mitochondria and chloroplasts to the nucleus seems to be an ongoing evolutionary process and a recurrent and consistent feature of eukaryotic genome evolution (Giulotto et al. 2010), which is also a continuing process of endosymbiotic evolution of the prokaryotic ancestors (Timmis et al. 2004). Numerous mitochondrial DNA sequences have been identified in the nuclear genome of Arabidopsis (Brennicke et al. 1993). Presumably, these mitochondrial genes comprise an intermediate stage of the mitochondria-to-nucleus gene transfer process and suggest that most of the genes originally present in ancient endosymbiont organelle ancestors have been gradually transferred to the nucleus (Brennicke et al. 1993; Bonen 2006; Thyssen et al. 2012).

Direct plant-to-plant HGT

Accumulating evidence suggests that two organisms are likely to undergo HGT when they are in intimate contact (Bock 2009). In plants, HGTs have been observed in association with grafting or parasitism, where plants were not only the donors but also the recipients of horizontally transferred genes (Fig. 2; Mower et al. 2010; Archibald and Richards 2010; Feschotte et al. 2010). Grafting creates an opportunity for cell-to-cell contact between distantly related plant species (Mower et al. 2004). HGTs are facilitated by particular processes during grafting. Firstly, grafting causes the rootstock and scion to come in contact with each other. After this, calli with numerous sensitive competent cells generate in the graft sites. These competent cells from the scion and root stock are adjacent, allowing naked genomic or organellar DNA fragments or messenger RNA (Liang et al. 2012; Lucas et al. 2009; Roney et al. 2007) or even the entire plastids to transfer between them via plasmodesmata channels (Stegemann and Bock 2009; Lucas et al. 2009) or vesicle transport (Bock 2009) (Fig. 2). Following this, transferred genes might directly integrate into the nuclear or plastid genomes of the scion or the rootstock plant. Previously, genetic material has been observed to be exchanged between the rootstock and scion during experimental grafting between a maple tree and a poplar tree (Bock 2009). Large DNA fragments or entire plastid genomes can travel across the graft junction, which facilitates the exchange of large quantities of genetic material between distantly related species (Fig. 2; Sandra Stegemann et al. 2012; Thyssen et al. 2012). This transferred genetic material might result in chimeric genes in the recipient cells. However, how long this gene expression can be maintained and whether these foreign genes are integrated into the host genome and inherited in the next generation of the plants require further investigation.

Fig. 2
figure 2

Genetic information transfers in plant grafting (between rootstock and scion) and in parasitic plants (between parasite and host)

Full size image

Direct contact between plant parasites and host plants has been proposed as an possible mechanism underlying many HGT events involving different DNA sequences, genes and transposable elements in many species (Bock 2009). Recently, evidence has also been obtained that both DNA fragments and complete chloroplasts can be horizontally transferred between different plants (even sexually incompatible species) by grafting (Sandra Stegemann et al. 2012; Stegemann and Bock 2009). Genetic material can also be conveyed between the plant parasite (including bacteria, fungi, insect, or parasitic plants) and its host plants via cell plasmodesmata channels, which have recently been discovered to act as a cell-to-cell communication pathway in plants (Lucas et al. 2009; Zambryski and Crawford 2000). Most parasitic plants have an organ (e.g., haustoria in Cuscuta sp. (dodder)) which allows penetration of the vascular bundles of the host plant. This long term persistent contact between the penetrative organs of the plant parasite and the plant host would lead to the connection of plasmodesmata between the parasite and the host surface cells. Subsequently, large DNA fragments or organelle DNA from the apoptotic cells of both the parasitic plant and its host could be mutually transferred through the plasmodesmata channels. As well, transfer of mRNAs in the phloem has also been documented in dodder, allowing transfer of genetic material from the host to the parasitic plant (Roney et al. 2007).

There are more than 4,000 parasitic plants which provide plenty of opportunities for HGTs between parasitic plants and host plants of different species (Davis et al. 2005; Shirasu et al. 2010; Mower et al. 2004; Davis and Wurdack 2004). In one example, a nuclear monocot gene was observed to be transferred into the genome of its parasitic weed (Striga hermonthica) (Shirasu et al. 2010). In another, phylogenetic evidence in the order Malpighiales showed host-to-parasite HGT: the mitochondrial genome of family Rafflesiaceae (an endophytic parasite) was derived from its host plants by HGT (Davis and Wurdack 2004). Schneeweiss et al. (2007) also found evidence for HGT of a plastid gene between Orobanche (broomrape) and closely related genus Phelipanche (both non-photosynthetic flowering plants in the Orobanchaceae family) (Schneeweiss et al. 2007). However, we still lack substantive evidence for the integration of foreign genetic material in plants as it may still be discarded by the native genome following mitosis or meiosis. If genomic integration is successful, these genes might participate in gene expression or regulation to change the characteristics of individuals and to shape the evolution of species as a whole.

Vector-mediated horizontal gene transfers between plants and other genomes

HGT between plants and other genomes can be mediated by the following possible vectors: parasitic plants, pathogenic organisms, transgenic vectors such as viruses and bacteria (e.g., Agrobacterium tumefaciens and Escherichia coli) (Broothaerts et al. 2005), insects (Fortune et al. 2008), fungi, pollen (Bergthorsson et al. 2003), plant parasites including nematodes (Sommer et al. 2011), and even transposable elements.

Virus-mediated horizontal gene transfers

Viruses have an extremely strong ability to capture genes from host genomes, including bacteria, fungi, and eukaryotic genomes. Thus, viruses are an important vector for HGT. The horizontal transfer of host genes to the virus is relatively common, whereas the horizontal transfer of virus genes to the host cell is less frequent but can also occur between retroviruses and some DNA viruses and their host cells (Jiang et al. 2010). DNA viruses can, for instance, capture host genes and integrate these genes into their own DNA (Ogata et al. 2007), subsequently infecting other organisms such as plant parasitic insects and transferring this captured DNA. In nature, virus-mediated horizontal gene transfer pathways may include nematode–virus–plant, nematode–virus–insect–plant, or insect–virus–plant (Fig. 1), where an example of an insect would be a mosquito. Plant/other host organisms–virus–aphid–plant may be another important HGT pathway (Fig. 1).

Bacteria-mediated horizontal gene transfers

Bacteria comprise a vector for gene transfer between plants, with a well-known example of this found in bacteria species Agrobacterium tumefaciens (Vogel et al. 2009). Evidence has also been provided in transplastomic tobacco that naked DNA from degraded plant tissue in the environment can be transferred to bacteria (Vogel et al. 2009). Many genes from diverse species of bacteria (also possibly including horizontally transferred genes from other plants) can be transferred to plants by natural bacterial transformation or transgenic technology (Schluter et al. 1995; Broothaerts et al. 2005; Richards et al. 2006).

Fungi-mediated horizontal gene transfers

Fungi often have the ability to infect other organisms, providing a possible pathway for horizontal gene transfer from the fungus to the host plant (Fig. 1). HGT via this “fungi–plant” pathway has previously been demonstrated in rice (Oryza sativa) (Richards et al. 2009). Many land plants have obtained the shikimate pathway, which links the biosynthesis of aromatic compounds to carbohydrate metabolism, from symbiotic fungi via horizontal gene transfer (Richards et al. 2006; Gribaldo et al. 2009). Genes transferred from plants to fungi may also be subsequently transferred to other plants infected by the same fungus. Plant–fungus–insect (e.g., aphid or nematode)–plant also comprises a possible HGT pathway (Fig. 1).

Parasitic insect-mediated horizontal gene transfers

Parasitic insects can also be a vector for horizontal gene transfer. For instance, cellulase genes were found to have been transferred to nematodes from bacteria, fungi, and other plant parasites, including other nematode species (Sommer et al. 2011). Carotenoid genes in aphid were also found to have been obtained from fungi (Moran and Jarvik 2010), and HGT from endosymbiotic bacteria to their host mosquitoes (Wolbachia pipientis) have also been documented (Woolfit et al. 2009). These HGT pathways involving infection by parasitic insects may also operate in plants. Bees not only carry pollen from one plant to another, potentially promoting HGTs between related or even distantly related species, but also allow the opportunity for HGTs due to interactions between pollen and the bee gut bacteria (Fig. 1; Mohr and Tebbe 2007).

Transposable element-mediated horizontal gene transfers

Transposable elements (TEs) are important vectors for the horizontal movement of genes between eukaryotic genomes (Diao et al. 2011; Boissinot et al. 2010; Panaud et al. 2009; Sormacheva et al. 2012). Transposons, with their inherent ability to mobilize, proliferate and integrate into genomic DNA and generate HGTs with ease (Capy et al. 2008). Transposons have been demonstrated to capture and transduce genomic DNA sequences in both Daphnia pulex (Schaack et al. 2010) and Drosophila species (Capy et al. 2008). The TE-mediated HGT process involves activation of the TE in a species and integration/transformation into a viral genome, infection of another species by the virus, then TE activation again, and integration into the new species host genome, completing the HGT.

HGT via transposable elements in plants may occur with an even higher frequency than in other organisms. Many TEs in rice carry genomic DNA gene fragments, which can be transferred directly or by other mediated vectors. Long terminal repeat (LTR) retrotransposons can produce viruslike particles that may work as vectors for gene transfer (Capy et al. 2008). Horizontal transposable element transfer has also been observed in the genus Oryza for LTR retrotransposon RIRE1 (Panaud et al. 2008). The transfer of Mu-like transposons between Setaria and rice has also been documented (Diao et al. 2006; Diao et al. 2011). Hence, TEs may mediate frequent horizontal gene transfer between plants. The frequency of horizontal transfer differs between DNA transposons and retrotransposons: retrotransposons are more frequent vectors for HGT (Silva et al. 2004).

In addition, in the photosphere, plant DNA released after the degradation of plant tissues can persist and remain biologically active for significant periods of time (Vogel et al. 2009; van Elsas et al. 2003). This mechanism may also provide a chance for direct contact between soil- or plant-associated bacteria and the naked plant DNA, leading to the integration of this plant DNA into the bacteria. Hence, decayed plant tissues in the soil provide a major DNA source for genetic material transfer between plants and bacteria (van Elsas et al. 2003; Vogel et al. 2009).