|Home | About | Journals | Submit | Contact Us | Français|
In the past decade, tools derived from DNA transposons have made major contributions to vertebrate genetic studies from gene delivery to gene discovery. Multiple, highly complementary systems have been developed, and many more are in the pipeline. Judging which DNA transposon element will work the best in diverse uses from zebrafish genetic manipulation to human gene therapy is currently a complex task. We have summarized the major transposon vector systems active in vertebrates, comparing and contrasting known critical biochemical and in vivo properties, for future tool design and new genetic applications.
DNA transposons are mobile genetic elements that move or transpose using a ‘cut and paste’ mechanism. Use of transposons in vertebrates was effectively launched with the high profile publication of the Sleeping Beauty (SB) transposon system . In this study, a molecularly reconstructed system was developed, with the transposase catalyst provided in trans with the key DNA cargo tagged by flanking transposon end sequences. Based in part on the rich array of prior work with transposon elements from invertebrates and plants, this paradigm is still the primary transposon approach deployed today.
In the past decade, DNA transposons from many different gene families have now been shown to work well in diverse species. This review focuses on activities in vertebrate organisms or tissues, as such studies will benefit development and application of DNA transposon tools in gene discovery and gene delivery. With advantages of easy production, simpler handling and without pathogenicity, it is worthwhile to generate and optimize non-viral transposon vectors over the current preferentially used viral vectors . As the biochemical properties of elements within a family can be very similar, we have organized current tools accordingly, with each element (Table 1) and transposon family (Table 2) summarized and compared.
Genetic applications can differentially leverage biochemical diversity in these transposon systems. ‘Transposition efficiency’ has been a limitation for DNA transposon elements compared to highly active viral methods. To generate transposon systems with improved activity is a constant effort in this field. ‘Integration site preference’ is an important consideration, as gene discovery applications can be enhanced by intragenic insertional preference, whereas such preference would be a perceived weakness for gene therapy uses. The DNA ‘footprint’ left behind after transposon integration and remobilization differentially impacts the use of these elements. For gene discovery work, a small DNA tag can be valuable to distinguish an engineered chromosome from an otherwise wild-type allele in that study. However, gene therapy applications would prefer as little disruption in the genome should an element hop more than once before landing in the final locus. High ‘cargo capacity’ is generally preferred, and transposon systems show robust differences in this key area. ‘Host origin’ is also an important constraint, as transposon vectors are normally used in heterologous organisms to avoid mobilizing related endogenous transposon sequences. With 10 transposons from four different superfamilies, genetic researchers have a rich array of diverse options at their disposal for vertebrate applications.
The Tc1/mariner superfamily represents the most widespread DNA transposon family in nature, members having been described for all kingdoms of life . Canonical representatives of this family were isolated from nematodes (Tc1 ) and arthropods (mariner ). Since that time, numerous members of this family have been isolated, some having been harnessed for gene transfer in numerous species. However, only a selected few are able to efficiently transfer genes in vertebrate cells or embryos (see Minos  and Himar  subsequently), leading to the search for related family members in vertebrate genomes. Despite their being widespread, often at thousands of copies per genome, nearly all of the elements that have been characterized and recovered have been ‘evolutionarily inactivated’ . Phylogenetically informed reconstruction of an inactive piscine element resulted in the isolation of the first active vertebrate Tc1/mariner element, SB . From that time, multiple members of the Tc1/mariner family from vertebrates have been reconstructed using similar approaches, including Frog Prince  and Hsmar1  transposons. In addition, intact endogenous elements have been discovered by comparative genome sequence analysis, one of which is widely represented in pleuronectid fish and was recently demonstrated as functionally competent for gene transfer in mammalian cells .
Tc1/mariner elements are ~1300–2400 bp in length and contain a single gene encoding a transposase enzyme that is flanked by terminal inverted repeats (IRs) . Shared biochemical features of the Tc1/mariner family include a transposase that contains a ‘DDE’ motif . This motif (two aspartic acid residues and a glutamic acid residue, the latter two separated by 35 residues) was first identified in retroviral integrases and bacterial IS (insertion sequence) transposases [12, 13]. This acidic amino acid triad has been inferred to be critical to the catalytic active site and involved in contacting divalent ions . In the Tc1/mariner family, the second aspartic acid residue and the glutamic acid residue are separated by only 34 residues, and an aspartic acid residue is found at the position of the glutamic acid residue in the mariner-like elements.
These transposons are bracketed by inverted terminal repeats that bind the transposase. The length of these repeats vary; the simplest being <100 bp with a single transposase binding site per repeat [7, 15]. Tc3 elements have IRs of >400 bp in length, each of which contains two binding sites, although the internal pair is not required for transposition . A third group referred to as ‘IR–DR’ have a pair of binding sites containing short 15–20 bp direct repeats (DRs located at the ends of IRs that are 200–250 bp long, including Minos and S elements in flies [6, 17], Quetzal elements in mosquitos , Txr elements in frogs  and at least four Tc1-like transposon subfamilies in fish [10, 20]. Tc1/mariner elements invariably integrate into TA dinucleotides in a manner that has been generally described as random, although local sequence context and deformation apparently plays some role in target site selection [21, 22], and at least one family member displays some proclivity for landing in genes .
There is conflicting data regarding the sensitivity of transposition for various Tc1/mariner family members to cargo size. In vitro (cell free) and in vivo transposition of Tc1 was decreased by 10- to 15-fold by simply increasing cargo size from ~2 kb to ~8 kb . This limitation in cargo has also been observed for Himar1  and for the SB transposon. Transfection of cultured mouse and human cells with SB transposons with different sized cargos has revealed a dramatic inverse relationship between transposon size and transposition [24, 25], although total plasmid size, not cargo size, has been hypothesized as responsible . Interpretation of these experiments is in part confounded by size-dependent, differential transfection and resulting gene expression as described by Yin W et al., 2005. In addition, transposition from genome-resident SB transposons [27, 28], or from DNA injected into the mouse pronucleus (S. Fahrenkrug, submitted for publication) does not seem to be as constrained by cargo size. Although the reason for these differences is unknown, it may be that nuclear localization of transposon–transposase complexes is size limiting, a barrier not experienced by transposons already resident in, or physically introduced to the nucleus.
Where it has been examined, members of this superfamily have commonly been found to be sensitive to a phenomenon referred to as ‘overproduction-’ or ‘overexpression inhibition’, first described for the mariner element , wherein excessive transposase expression was found to inhibit transpositional transgenesis. The same phenomenon has been described for the Himar1 , SB  and Passport  transposons, although Izsvak et al.  found the frequency of SB transposition to be directly proportional to the level of transposase expression. The mechanism of ‘overproduction inhibition’ is still poorly understood. One working model suggests that an optimal ratio of transposase molecule to its binding terminals in an effective complex falls into a small concentration window and excess transposase will quench the formation of the complex . In numerous studies, the optimal concentration and molar ratio of transposon and transposase has been found to vary between cell types, suggesting either differing abundance of co-factors, or a difference in the expression or stability, of the transposase. Thus, application of these transposons as gene delivery vehicles requires a careful assessment of the conditions best suited for the particular application and cell type.
SB was the first widely deployed transposon for vertebrate applications and was molecularly reconstructed from ancient inactivated mariner-like elements . SB has become the de facto reference system and has been used for diverse applications, including: cellular transgenesis (piscine, avian and mammalian), animal transgenesis (fish, frogs, mice, rats and pigs) and gene therapy, with over 175 publications describing its use at the time of this publication.
Some milestones of SB applications are highlighted here. SB was the first DNA transposon with high efficiency gene transfer into human cells , the first non-viral integrating gene therapy vector , used for the first non-viral somatic mutagenesis system , and deployed for the first transposon-based transgenic mammal . SB was the first effective transposon-based system for trangenic fish  and insertional mutants  in zebrafish. Finally, SB was the basic system for the initial attempts at generating a chimeric transposase for targeting transposon integration .
Components of the SB system have undergone extensive engineering with the resulting elements showing increased activity from improvements to the transposon [37–39] and to the transposase ([25, 39–41]). Most assays for these hyperactive elements are focused on measurements using a HeLa tissue culture gene transfer method . SB activity, like other Tc1/mariner family members, is subject to overexpression inhibition; as a result, the selection of transposase dose for maximal gene transfer is a critical consideration in using SB for best gene transfer results.
Methylation studies have demonstrated that chemical modification of SB transposons by cytosine-phosphodiester-guanidine (CpG) hypermethylation can significantly enhance their transposition rates . The mechanism for this enhancement of transposition is currently unknown. Identifying the experimental scenarios where methylation can make a significant improvement in gene transfer by SB is an ongoing research area for this element and will potentially have implications for work in other transposons.
In a series of yet-unexplained observations in Xenopus, two research groups have noted a positive gene transfer rate for SB that did not appear to be due to standard transposition reactions, instead displaying non-canonical footprints [43, 44]. Given the abundance of related Tc1/mariner transposons in frog genomes, this may reflect interference by endogenous transposable elements, or this may be due to some other aspect of SB biology that is limited in frogs.
The Frog Prince (FP) transposon was ‘kissed’ to life by Csaba Miskey in the laboratories of Zsuzsanna Izsvak and Zoltan Ivics  using the approach deployed for SB. For generating the FP transposon system, 10 inactive putative elements with complete open reading frames were identified from the frog Rana pipiens and used to produce a consensus transposase sequence. A single element with two amino acid substitutions different than the consensus (mFP) was used to produce a consensus transposase (FP). Interestingly, the mFP version of the transposase was inactive until correction to the consensus in FP . FP is closely related to Txr elements found within Xenopus but differs significantly from other members of the Tc1 family, notably SB. FP is capable of mobilizing Txr elements albeit at a reduced rate to FP transposons. In a limited analysis, the FP IRs allowed splicing to internal splice-acceptors suggesting the 214 bp IRs do not significantly interfere with splicing, making FP a potential transposon for gene discovery using gene-trapping technologies .
Like SB, FP appears to be active in a wide range of vertebrate cells, including HeLa, Cho-KI (hamster), A6 (Xenopus laevis), FHM (fat-head minnow) and PAC2 (zebrafish). The level of transposition in these cell lines was comparable or slightly improved relative to transposition of the original SB system.
Miskey et al.  reactivated the human mariner 1 transposon (Hsmar1) from about 200 copies of full-length transposons that contained inactivating mutations. The consensus Hsmar1 transposase sequence was not capable of promoting transposition in human cells. Therefore, 51 sequences from the human genome as well as mariner sequences from the cecropia transposase subfamily were aligned to determine the likely ancestral protein sequence for site-specific mutagenesis. An additional four amino acid substitutions were needed to make the Hsmar1 reconstructed ancestral protein (HsMar1-Ra). HsMar1-Ra was capable of excising transposons in HeLa cells and zebrafish embryos. Integration was shown by colony formation as well as examination of the integration sites in HeLa cells. Hsmar1 integrates into TA dinucleotides and appears to distribute randomly throughout the genome. The levels of excision or transposition of Hsmar1-Ra was not directly compared to any other transposon during this study, although there was an ~20-fold increase in stable colony formation in HeLa cells.
Hsmar1-Ra is unlikely to be used for human applications like gene therapy because of its ability to mobilize both non-autonomous full HsMar1 elements as well as the more prevalent (approximately 2500) Hsmar1-related mini-inverted repeat transposable elements (MITEs) found within the human genome . How valuable HsMar1 will be in non-human work is still unclear. It seems likely that Hsmar1 will function in other systems including the zebrafish (where Miskey et al. demonstrated the ability of Hsmar1-Ra to cut or excise Hsmar1 elements after injection into zebrafish embryos), but whether the transposition achieved is at a high enough rate for practical use remains to be tested. One advantage of Hsmar1-Ra is the short IRs of this class of mariner-type elements, as their relatively simple sequence requirements may facilitate the production of transposon vectors through the addition of IR sequences to a particular genetic cargo via a PCR reaction.
Minos, named after the legendary Greek King, is a DNA transposon belonging to the Tc1/mariner superfamily . Minos elements isolated from Drosophila hydei show high sequence homogeneity with characteristic features: about 1.8 kb in length; perfect inverted terminal repeats (IRs) of 254 bp; encode a putative single transposase gene with two exons and a 60 bp intron in between, which has been shown to be spliced out in the D. melanogaster germline. Southern blot analyses of various D. hydei strains strongly suggested that Minos transposons are active in the germline  and that Minos-like elements are widely spread among Drosophila species .
In Drosophila, the P-element is the most widely used transposon and has generated thousands of valuable insertion lines for various research purposes. However, P-element activity is very host-restricted and appears to be inactive in non-Drosophilids . In contrast, Minos elements have demonstrated a much wider host spectrum, including Drosophila and non-Drosophilid insects (for review, see ).
In human HeLa cells, researchers were able to achieve high frequency of Minos transposition in stably transformed cells. They proposed that Minos can be used to potentially tag all genes of the human genome . Minos elements are active in mouse lymphatic tissues and embryonic fibroblast cell lines . Minos transposase was effective in mediating marker gene transposition in the mouse germline .
Minos does not prefer any sequence beyond the di-nucleotide TA and tends to insert more frequently into introns than exons in both fly and mouse . In the fruit fly, the excision of Minos could leave precise conversion at the insertion sites or 6 bp characteristic footprints . However, in mammalian systems, the footprints appear more complex, potentially due to differences in the host chromatid-repair machinery in these organisms [49, 51].
Himar1 is a natural element of the Tc1/mariner superfamily from the horn fly, Haematobia irritans. This element has been shown to be active in a number of animals, including mammals [52, 53]. Hyperactive variants have been developed as well . How this element quantitatively compares to other members of this family in vertebrate applications is still largely unknown.
The Passport transposon was initially identified within the plaice (Pleuronectes platessa) genome by Leaver et al. in 2001 . In that study, a complete transposase open reading frame was described from within the plaice genome with a high degree of similarity among isolated transposon sequences. This potentially active transposon was dubbed PPTN.
Recently, Clark et al. [56, 10] dissected the transposase and transposon IRs of the PPTN transposon and showed that it was a functional transposase active with native IR and transposase sequences in human cells. The transposon system was renamed Passport referring to its ability to give the right or privilege of passage, entry or acceptance—in this case into the genome. Passport is a Tc1-type element, but it differs significantly from both SB and FP in IR and transposase sequence. Like other Tc1-type elements, Passport has IRs of around 200 bp and exhibits transposition into a wide variety of cell types, including human, hamster, pig, monkey, turkey and chicken. Passport was directly compared to SB in human cells through a range of transposase activity. Passport was about 40% as active as an improved SB at their respective peaks. Passport was shown to exhibit overexpression inhibition dynamics that were shaped similar to SB; delivery of too little or too much transposase results in 40% or less peak activity.
As compared to other Tc1 elements used in vertebrates, early data suggests that target-site selection is less random with a preference for integration into transcription units (~60%) and 6 of 27 integrations into human chromosome 12 (distributed) and 0 of 27 integrations into human chromosome 1. This is unexpected since the size of human chromosome 1 is nearly twice as long as chromosome 12. Passport could likely function as a new transposon system for the manipulation of vertebrate genomes.
Transposition from hAT family members in maize was the first visible mobile genetic element activity noted in science . Diverse active native elements have been identified in vertebrates, including two from the fish medaka (Tol1 and Tol2; [58, 59]). Unlike studies with reconstructed SB, most hAT family work to date has been with autonomously active elements. Native hAT transposons are about 5 kbp in size, and common shared properties include high cargo capacity, no obvious primary sequence insertion site requirement and less influence of overexpression inhibition compared to SB and related Tc1/mariner elements. A common 8 bp duplication is noted at the insertion site for each of these elements, with no primary sequence constraint for integration.
Most hAT transposon family work in vertebrates to date has been with derivatives of the native element from the medaka fish (Oryzias latipe), Tol2 [58, 60]. The native Tol2 transposon is ~5 kbp in size and encodes an hAT family transposase . Minimized Tol2 IR transposon ends sufficient for full transposition in vivo have been developed (called ‘miniTol2’; [61–63]). Good cargo capacity has been described for Tol2, with little or no major decrease in activity with transposons >10 kbp compared to smaller, 2 kbp elements [62, 63]; larger elements can be robustly mobilized with Tol2 in zebrafish (S. Ekker, submitted for publication). Tol2 demonstrates perhaps the least functional constraint due to overexpression inhibition of the elements describe here, with a large range of the transposase resulting in near peak gene transfer activity due to transposition .
Tol2 is active in all vertebrates tested to date, including robust activity in zebrafish , Xenopus , mice and human tissue culture cells . Favorable kinetics over SB makes this a current favorite element for the generation of transgenic zebrafish , and notably a range of excellent pre-assembled vectors for work in this model organism (called the ‘Tol2kit’) is freely available [67, 68]. Tol2 represents the current standard for hAT transposon family functionality and utility in vertebrates.
Tol1 was identified as an active DNA transposon from the medaka genome [58, 69]. The autonomous Tol1 element is ~4.4 kbp, encoding a transposase gene with three exons. The transcribed mRNA is ~2.9 kbp coding for a putative functional transposase of 851 amino acids with sequence similarity to hAT superfamily .
The activity of Tol1 transposase to deliver cargos flanking by Tol1 IRs has been analyzed in detail. Minimized Tol1 IRs with 157 bp from the left and 106 bp from the right termini maintain full transposition efficiency . Cargo capacity was also tested, with DNA fragments up to 22.1 kbp in size were transposed without internal deletion or aberration , consistent with previous observation that transposons from hAT family have demonstrated large cargo delivery capacity in general compared to those from other families . Tol1, like Tol2, has a wide range of activity in a variety of host species, including good activity in human (HeLa) and mouse cells (NIH/3T3)  and Caenorhabditis elegans . Excision activity of Tol1 has been reported in the frog X. laevis . Most recently, Tol1-derived vectors have been shown to actively transpose in zebrafish germline and pass to the next generation .
So far, Tol1 biochemistry parallels work with the widely used Tol2 transposon in zebrafish. Tol1 and Tol2 do not cross react to the cis-sequences of each other, indicating Tol1 as a good candidate for developing genetic tools in vertebrates and further mutagenesis of insertional mutants generated by Tol2 .
Over 50 years ago, the first transposable system Ac/Ds (Activator/Dissociation) was discovered by Barbara McClintock from Zea mays (maize) . Being one of the founding members of the hAT superfamily, the Ac element is 4.5 kb in length and displays all the common molecular features of this family, including short terminal IRs (11 bp), encoding a putative DDE-domain containing transposase and causing 8 bp host duplication upon insertion . Ds elements are the natural non-autonomous versions of Ac, containing the IRs and cis-required sequences but lacking the ability to produce a functional transposase . Their transpositions depend on the presence of the Ac element or Ac transposase.
The transposition of Ac/Ds system in heterologous plants has been demonstrated in both dicot and monocot (for review see ). Later, Saccharomyces cerevisiae became an even more distant species that can host Ac/Ds ‘cut and paste’ activity . Most recently, the possibility of employing a plant transposon in a vertebrate was exploited. A chimeric Ac transposase (NLS-Tpase), composed of an animal-origin nuclear localization signal (NLS) fused to the N-terminal of the truncated maize Ac lacking the fist 102 amino acids (Tpase 103–807), demonstrated Ds element transposition in zebrafish in both somatic and germline cells . This modified Ac (NLS K5E-Tpase) effectively inserts Ds element DNA in a human embryonic kidney cell line as well .
The Ac transposase behaves similarly in zebrafish and plant species in various transposition assays [76, 78]. However, it is unknown whether the regulation of Ac observed in the plant kingdom still holds in vertebrates, such as different dosage effects in different species [79, 80], methylation of transposase binding sites [81, 82] and certain host factors effects . More careful analysis of Ac activity in more diverse vertebrates should provide more information on the mechanisms of transposition of the hAT family.
The PIF/Harbinger transposon family includes elements from plants and animals, and the Harbinger3_DR is a molecularly reconstructed element of this family from the zebrafish (Danio rerio) genome . Besides containing a gene encoding transposase, Harbinger_DR has a second gene encoding a Myb-like protein involved in either DNA binding or protein–protein interactions. This transposon has short 12 bp IRs. Insertion site preference indicates a strong requirement for CAG or CTG trinucleotides, with an even larger, 15 bp total consensus sequence noted . This element does not appear to leave a footprint upon excision .
piggyBac  is the founding member of its class of transposon and was first described as a DNA hitchhiker in an invertebrate culture system. This element has been effectively harnessed, with minimized sequences identified . piggyBac is active in many vertebrates tested to date [87–90]. Key biochemical properties include a sequence requirement at integration (‘TTAA’; ), and an apparent preference for integration into transcriptionally active genomic DNA . Remobilization of this element often results in a chromosome with no residual genetic change yielding no observable footprint . A more modest overexpression inhibition profile compared to SB has been described .
DNA transposons are extremely effective tools for genetic modifications of vertebrates. With molecular engineering skills activating evolutionarily quiescent genes, or minimizing autonomously active elements, we now have an array of vectors from diverse organisms showing robust activity in a variety of model systems. The complementary biochemical properties encoded by these vector systems open the door to many new vertebrate gene discovery and gene transfer applications. Table 2 has summarized the positive and negative features of each transposon superfamily for prospective applications.
Mayo Foundation and National Institutes of Health grant (DA14546 to S.C.E.).
Jun Ni's research interests are on the biochemistry of transposons for vertebrate genetic applications.
Karl J. Clark has a long-standing interest in the development of vertebrate transposons including key work on Sleeping Beauty and Passport.
Scott C. Fahrenkrug's science focus includes advanced applications for the use of transposons for application to animal biotechnology.
Stephen C. Ekker is developing transposons for genetic applications in the zebrafish, Danio rerio.