Transposable Elements for Insect Transformation Part 4

piggyBac

Discovery of piggyBac and other TTAA-specific elements Similar to several other insect transposable element systems, the piggyBac element was discovered fortuitously in association with a mutant phenotype. However, unlike all the other transposons used for insect transformation, the mutant phenotype was the result of a functional element that had transposed into an infectious organism. Fraser and colleagues (see Fraser, 2000) discovered several FP (Few Polyhedra) mutations in the baculoviruses Autographa californica nucleopolyhedrovirus (AcNPV) and Galleria mellonella nucleopolyhedrovirus (GmNPV) after passage through the Trichoplusia ni cell line, TN-368 (Fraser et al., 1983, 1985). Among these elements that inserted specifically into tetranucleotide TTAA sites was piggyBac (then named IFP2), which transposed into AcNPV. Although it might be assumed that IFP2 was an autonomous functional element based on its mobility, another TTAA insertion-site element, tagalong (then called TFP3), discovered in AcNPV and GmNPV, was later found not to have an uninterrupted transposase coding region, and thus had to be mobilized by another functional TFP3, or related element. Autonomous functional elements have not yet been found for tagalong, though the original IFP2 piggyBac element was indeed functional (Wang et al., 1989; Wang and Fraser, 1993). All the piggyBac elements discovered in TN-368 were found to be identical to IFP2, having a length of 2472 kb with 13-bp perfect ITRs and 19-bp subterminal repeats located 31 bp from the 5′ ITR and 3 bp from the 3′ ITR (Cary et al., 1989; see Figure 2). Notably, five piggyBac-like elements have been isolated from T. ni larval genomic DNA, but thus far none have been found to be identical to IFP2 (Zimowska and Handler, 2006).


The IFP2 transposase coding region exists as a single reading frame of 2.1 kb that encodes a protein with a predicted molecular mass of 64 kDa. The functionality of piggyBac and the precise nature of its transposition was further verified by a series of viral and plasmid transposition and excision assays. A piggyBac indicator plasmid marked with polh/lacZ was used in Spodoptera frugiperda SF21AE cell line assays which showed that the original piggyBac element, within the p3E1.2 plasmid, could mobilize the marked element. These assays proved that the 3E1 piggyBac element encoded a functional transposase, and defined the element’s TTAA insertion-site specificity and the precise nature of its transposition. Importantly, these assays also showed directly that piggyBac could be mobilized in other lepidopteran species (Fraser et al., 1995), indicating that it might function similarly as a vector for germ-line transformation. This was a critical realization given the failure of P to be mobilized in non-drosophilids, which was consistent with its failure as a vector in these species.

The highly precise nature of piggyBac transposition is unique among known transposons, many of which excise in a fashion that leaves staggered ends at the donor site. The necessary filling-in of these ends for target joining and gap repair by DNA synthesis often results in mutations, and this can be a desired effect for transposon-induced mutagenesis. Recent in vitro tests in yeast, using purified transposase protein, indicate that piggyBac excision results in complementary TTAA overhangs at the donor site, which are restored precisely by ligation without the need for DNA synthesis (Mitra et al., 2008). The mechanism for this is similar to that used by the widespread DDD/DDE recombinase transposons, which was surprising, given their lack of sequence similarity to piggy-Bac. Mutations affecting the piggyBac D268 and D346 residues, however, suggest that they have Mg2+-depen-dent catalytic function. The actual mechanistic relationship between piggyBac and the DDD/DDE transposon family, as well as other aspects of its transpositional activity, await further study. Given the growing importance of piggyBac to genetic transformation of a wide array of organisms, and especially its potential for gene therapy in human stem cells (Feschotte, 2006), this knowledge should be rapidly forthcoming.

piggyBac transformation The failure of P vectors to transform non-drosophilid species made the testing of other available transposon systems a high priority. The other systems found to be functional in non-drosophilids, however, were first successfully tested for gene-transfer vector function in Drosophila. For piggyBac, germ-line transformation was first attempted in the Mediterranean fruit fly C. capitata. This was possible due to the availability of a marker system that had been previously tested by medfly transformation with the Minos transposon vector. The medfly white+ gene cDNA was linked to the Drosophila hsp70 promoter, and used as a mutant-rescue system in a white eye host strain (Loukeris et al., 1995a; Zwiebel et al., 1995). In the absence of data for the minimal sequence requirements for piggyBac mobility, the first piggyBac vector was constructed by insertion of the 3.6-kb hsp-white+ cDNA marker into the unique HpaI site within piggyBac in the p3E1.2 plasmid. None of the piggyBac sequences was deleted, although the insertion interrupted the coding region and eliminated the expression of functional transposase. Construction of the first piggyBac helper plasmid involved a simple deletion of the 5′ terminal inverted repeat resulting from a SacI digestion and religation of p3E1.2. There is some uncertainty as to whether the upstream SacI site cuts within the piggyBac promoter (Cary et al., 1989), yet transposase expression was indeed sufficient to support germ-line transpositions from the vector plasmid. The first experiment with this helper in medfly resulted in one transgenic line at a transformation frequency of 5% per fertile G0; however, sibling sublines exhibited two and three independent integrations (Handler et al., 1998). This experiment with a piggyBac-regulated helper was repeated, and five additional G1 lines were isolated at approximately the same frequency (5% per fertile G0). These attempts at piggyBac transformation yielded relatively low transformation frequencies, but it was notable that a lepidopteran transposon vector system had autonomous function in a dipteran species.

Subsequent to the medfly transformation, piggyBac transformation was tested in Drosophila using the mini-white marker from that species (Handler and Harrell, 1999). Using the self-regulated helper, transformants were isolated at a similar frequency of 1—3%, but tests with a hsp70-regulated transposase increased the frequency to above 25%, consistent with P and hobo transformations using heat shock promoted transposase.

Dipteran transformations With a more highly effective helper, the white+-marked piggyBac vector tested in medfly was subsequently tested in a white eye mutant strain of another tephrtid pest, the oriental fruit fly Bactrocera dorsalis (Handler and McCombs, 1998). Although the transformation frequency of B. dorsalis was somewhat lower than that observed in Drosophila using the same phspBac helper (26%), it was discovered that B. dorsalis genome contained multiple piggyBac-like elements that might have had a repressive effect on piggyBac transposition. Several other tephritid pest species have since been transformed with piggyBac vectors containing fluorescent protein expressing transgenes, including the Caribbean fruit fly Anastrepha suspensa and the Mexican fruit fly A. ludens. The transformation of A. suspensa was the first non-drosophilid transformation to use a fluorescent protein expression transgene as a marker (PUb-nls-EGFP; Handler and Harrell, 2000). The transformation of A. ludens was done with vectors allowing post-integration stabilization and sperm-specific marking using a beta2-tubulin-regulated DsRed marker (Condon et al., 2007; Zimowska et al, 2009; Meza et al, 2010). Other teph-ritids that have been transformed include the Queensland fruit fly Bactrocera tryoni (Raphael et al., 2010), and recent transformations of medfly and caribfly created insects with sperm-specific expression of DsRed (Scolari et al., 2008; Zimowska et al., 2008) or site-specific recombination sites, attP (Schetelig et al, 2009, 2010).

Other dipteran species transformed with piggyBac include several of medical and agricultural importance, such as the mosquitoes Ae. aegypti (Kokoza et al., 2001; Lobo et al., 2002), An. gambiae (Grossman et al., 2001), An. albimnaus (Perera et al., 2002), An. stephensi (Nolan et al., 2002), and Aedes fluviatilis (Rodrigues et al., 2006). An. gambiae was of particular interest, given its medical importance as a major malaria vector, and its relatively low transformation frequency of ~1% in initial experiments was discouraging (with many anecdotal reports of failure). A more recent transformation yielded several founder trans-formant lines at a frequency range of 4—18%, representing a considerable improvement (Lombardo et al., 2009). This experiment took advantage of improvements in mosquito transformation methodology discussed in detail by Lobo et al. (2006). Other successful piggyBac-mediated transformations of dipterans include M. domestica (Hediger et al., 2001), Lucilia cuprina (Heinrich et al., 2002), L. sericata (Concha et al., 2010), the New World screwworm Cochlio-myia hominivorax (Allen et al., 2004; Handler et al., 2009), and a number of Drosophila species (Holtzman et al., 2010).

Lepidopteran transformations Given that piggyBac was first isolated from a lepidopteran species, there was some optimism that it would be functional as a vector in other moth species that had not yet been transformed using the other transposon vectors originally discovered in dipteran species. Function was first tested by transposition assays in the pink bollworm P. gossypiella (Thibault et al., 1999), which then led to successful germ-line transformation of this species using the phspBac helper and a vector marked with EGFP regulated by the Bombyx actinA3 promoter (Peloquin et al., 2000). Concurrent experiments were also performed in the silkmoth B. mori using a similar actinA3-regulated EGFP marker, but for this species transformation was achieved with an actinA3-regulated transposase helper (Tamura et al., 2000). Bombyx mori has since been transformed routinely for a variety of studies, making it the most widely transformed species with piggyBac-based vectors. Some of the transformed lines include a UAS-Gal4 gene expression system (Imamura et al., 2003), an enhancer-trap system (Uchino et al., 2008), an inheritable heat shock inducible RNAi system (Dai et al., 2007), and the production of a recombinant Spider dragline silk (Wen et al., 2010). More recent lepidopteran piggyBac transformations include the codling moth Cydia pomonella (Ferguson et al., 2010), Plutella xylostella (Martins et al., 2010), and the first transformation of a butterfly, Bicyclus anynana (Marcus et al., 2004).

Coleopteran, Hymenopteran and Orthop-teran transformations Other insect species transformed with piggyBac vectors include the coleopterans T. castaneum (red flour beetle; Berghammer et al., 1999; Lorenzen et al., 2003), and Harmonia axyridis (ladybird beetle; Kuwayama et al., 2006). The only hymenopteran transformed thus far is the sawfly Athalia rosae (Sumitani et al., 2003). For Tribolium, piggyBac vectors have been used for large-scale enhancer-trap screens (Lorenzen et al., 2007) as well as large-scale insertional mutagenesis (Trauner et al., 2009), which should be invaluable to the functional genomic analysis of this species. In Harmonia, effective RNAi activity against a GFP marker was demonstrated.

General considerations for piggyBac transformations Notably, many of the species transformed with a piggyBac vector used a helper regulated by the Dro-sophila hsp70 promoter, and with vectors marked with EGFP, though other fluorescent proteins have since been used for some as well. Although most of these transformations occurred at frequencies of between 3% and 5% per fertile G0, dramatic differences between species have been observed, and in some of the same species performed by different laboratories. As noted, a single transformant line was reported for An. gambiae, at a frequency of approximately 1% (Grossman et al., 2001), while transformation in An. albimanus occurred at frequencies ranging from 20% to 40% (Perera et al, 2002). The first transformations of Tribolium occurred at the unusually high frequency of 60% (Berghammer et al., 1999).

Many of the transformations were preceded by testing piggyBac function by embryonic transposition assays, which were first developed for piggyBac mobility in the pink bollworm (Thibault et al., 1999). As discussed previously, these assays can rapidly assess the relative mobility ofpiggyBac in a specific host species in a few days. Positive results from these assays provided some assurance that more tedious and time-consuming transformation experiments had some likelihood of success. For some studies the assays were also used to test promoter function in helper plasmids, or provide insights into insertion site specificity, or determine the likelihood of a particular vector construct retaining function in the absence of specific sequences (Lobo et al., 2001). For example,piggyBac helper promoters were tested by transposition assays and germ-line transformation in D. melanogaster and L. cup-rina (Li et al., 2001a; Heinrich et al., 2002). It was found that in Drosophila an hsp70-regulated helper yielded the highest transposition frequency, while a constitutive a1-tubulin-regulated helper was more effective for germ-line transformation. By comparison, in Lucilia the hsp70 helper was most effective for both plasmid and germ-line transpositions, while the Drosophila a1-tub-helper failed to support transformation. More recent comparisons of promoter function in Lucilia have been based on germ-line transformation, showing that the L. cuprina hsp83 promoter is more effective relative to the Drosophila hsp70 for both transposase and ZsGreen marker activity (Concha et al., 2010). Transposition assays have also shown target site preferences among the TTAA sites within the pGDV1 target plasmid, and assays in Drosophila showed a bias for sites having A or T nucleotides at positions -3, -1, +1, and +3 relative to TTAA (Li et al, 2001a). However, a sequence analysis of 45 genomic integration sites in Tribolium, after piggyBac vector remobilization, failed to show this bias (Lorenzen et al., 2003), which may be an indication of variances between plasmid and chromosomal transpositions, or species specificity for insertion site preference.

Mobility assays also provide a rapid means of testing sequence requirements for vector mobility, and allow modifications for more efficient vector function. Since vector mobility is known to be negatively affected by increasing size, this information should allow minimal vectors to be created that retain optimal function. However, minimal sequence requirements for plasmid transpositions may differ from those for chromosomal transposition. For example, excision and transposition assays performed in T. ni embryos showed that the piggyBac inverted terminal repeat and subterminal repeat sequences were sufficient for transposition (35 bp from the 5′ terminus and 63 bp from the 3′ terminus), but that an outside spacer region between the ITRs of greater than 40 bp is necessary for optimal transposition from a plasmid (Li et al., 2001b). Use of similar vectors in Drosophila, however, did not result in germ-line transformants (A. Handler, unpublished). The minimal sequence requirements for piggyBac transformation verified thus far for Drosophila are 300 bp from the 5′ terminus and 250 bp from the 3′ terminus (Li et al., 2005).

Post-integration behavior of piggyBac The post-integration behavior of piggyBac vectors has been investigated in a number of species, including Drosophila melanogaster, Ceratitis capitata, Tribolium castaneum, Bombyx mori, and Aedes aegypti (Thibault et al., 2004; Lorenzen et al., 2007; Sethuraman et al., 2007; Uchino et al., 2008; Schetelig et al., 2009; Trauner et al., 2009). In Drosophila, piggyBac has been used extensively to generate insertions in a large number of genes throughout the genome (Thibault et al., 2004). These studies have shown that piggyBac can remobilize in this species and that it has integration site preferences that are complementary to the widely used P element, increasing its value as a functional genomics tool. Notably, piggyBac was a more effective gene-disruption tool than P elements because, unlike P elements, they did not preferentially insert into the 5′ region of genes. Remobilization in the germ-line of D. melanogaster was efficient when transposase was provided by a transposase open reading frame located on a chromosome and regulated by a promoter active in germ cells. New transposition events were recovered from 60-80% of the germ-lines tested. Consequently, these investigators were able to generate over 18,000 piggyBac insertions. Although piggyBac can be remobilized, excision always results in the perfect restoration of the chromosome to its pre-integration state. This is a unique aspect of the piggyBac system and quite unlike all other insect gene vectors, in which excision often results in small, and sometimes large, perturbations of the genomic sequences around the site of element excision. Although excision-mediated addition or deletion of sequences can be a useful way of creating allelic series, this is not an option with piggyBac because of its tendency to excise precisely. Although Thibault et al. (2004) reported efficient remobilization, they used a limited number of initial elements to generate the 18,000 transpositions recovered during their experiment. If a somewhat larger sample of integrated piggyBac elements is examined, one finds that the rates of piggyBac remobilization in D. melanogaster vary widely and depend greatly on where in the genome the element is located. Esnault et al. (2010) measured the remobilization activity of 20 identical piggyBac elements on the X-chromosome and found that excision/transposition activities varied over two orders of magnitude, though almost all of the variance observed was due to chromosomal position effects. The effects of the vector’s position in the genome also affected the levels of gene expression from genes within the vector, but these effects were not correlated with levels of vector remobilization. Esnault et al. (2011) also showed that no more than approximately 500 bp of flanking chromosomal DNA are responsible for the observed position-dependent variance in element activity. An element could be transplanted to other genomic positions and would retain its original remobilization activity as long as 500-1000 bp of the original flanking chromosomal DNA accompanied the transplanted element. Thus, these authors found that piggyBac was sensitive to its local DNA sequence context, and that this context-effect was portable within the genome. In addition, they found that the context effect was also portable to plasmids. piggyBac elements in high mobility contexts were more efficient gene vectors in D. melanogaster than identical elements in low mobility contexts (Esnault et al. in press).

Lorenzen et al. (2007) recovered transposition events from 97% of the germ-lines of transgenic T. castaneum containing a single piggyBac element and a chromosomal source of piggyBac transposase. Using a similar strategy, Trauner et al. (2009) produced over 6500 new piggyBac insertions in T. castaneum as part of a large-scale effort to identify genes involved in development. Although the phenomenon of local hopping was not reported for piggyBac transposition in D. melanogaster, it was observed in T. castaneum (Thibault et al., 2004; Trauner et al., 2009). While active and efficient gene vectors are essential for such large-scale gene finding efforts, equally important is the ability to rear and maintain large numbers of unique genetic lines. This is not possible for many insect species.

With similar interests in using piggyBac remobilization as a tool for identifying genes through enhancer trapping and insertional mutagenesis, Uchino et al. (2008) were able to create B. mori lines ubiquitously expressing piggy-Bac transposase, and lines containing enhancer-trap constructs. They found that the average maximum frequency of transposition was approximately 42%. Although they only generated a small number of lines (105) relative to similar studies with D. melanogaster and T. castaneum, it appeared that piggyBac did not prefer to transpose locally, and that it did appear to prefer intergenic regions and repetitive DNA over coding and genic sequences. Nonetheless, piggyBac proved to be an effective tool for identifying genes, based on enhancer trapping in this lepi-dopteran species.

In the Mediterranean fruit fly C. capitata, integrated piggyBac elements can also be remobilized when supplied with functional piggyBac transposase (Schetelig et al., 2009). These investigators were not remobilizing piggyBac for the purposes of gene-finding, but as part of a strategy for stabilizing integrated transgenes in which excision of the element left a previously integrated transgene that was no longer flanked by functional terminal inverted repeats of the piggyBac element (Schetelig et al., 2009). Remobili-zation was stimulated by injecting transposase expressing plasmids into presumptive germ cells and screening for element mobility in the next generation.

Although piggyBac elements appear to have great potential in insects for being used in applications requiring the remobilization of integrated elements (enhancer/ promoter trapping, mutagenesis, transgene stabilization, gene drive), their behavior in Aedes aegypti is notably different. Sethuraman et al. (2007) attempted to remobilize the piggyBac elements in five transgenic lines of A. aegypti by introducing, through genetic crosses, chromosom-ally-located piggyBac transposase genes. Testing multiple combinations of piggyBac reporter elements and trans-posase-expressing transgenes, these investigators failed to detect any evidence of piggyBac transposition. This unexpected stability of piggyBac following its integration into the genome of Ae. aegypti was confirmed in transgenic Ae. aegypti cell lines which contained integrated piggyBac elements and were transfected with plasmids containing the same piggyBac transposase gene that had been integrated into the genome and used by Sethuraman et al. (2007) (D. O’Brochta and Palavasam, unpublished data). Although there was no remobilization of chromosomally located piggyBac elements in Ae. aegypti cell lines in the presence of piggyBac transposase, plasmid-borne piggyBac elements could remobilize (excise) under these conditions, confirming the presence of functional transposase. These data suggest that the chromosomal context of integrated elements is playing an important role in determining their potential to remobilize in Ae. aegypti, which is consistent with the results of Esnault and colleagues (in press). Interestingly when piggyBac elements and 1 kb of flanking chromosomal DNA were transplanted from Ae. aegypti into the genome of D. melanogaster, they were now capable of high levels of remobilization activity (excision and transposition) (A. Palavasam, C. Esnault and D. O’Brochta, unpublished), confirming the functionality of the integrated elements that were formerly in Ae. aegypti, and suggesting that the local context effect described by Esnault et al. (in press) is species-specific.

Phylogenetic distribution of piggyBac and implications for transgene stability Similar to other transposons used for transformation, piggyBac is a member of a larger family (or superfamily) of related elements, such as the mariner/ Tc1 or hAT families. The piggyBac superfamily includes piggyBacAike elements that are highly similar to piggyBac, as well as more diverged piggyBac-related elements (though use of this terminology has not been consistent). piggyBac-related elements were first discovered in T. ni genomic DNA, where five different, though nearly identical, piggyBac elements were discovered and sequenced (Zimowska and Handler, 2006). One of these elements had a single amino acid change in the transposase open reading frame that did not affect the functionality of the protein (G. Zimowska and A. Handler, unpublished). Other elements nearly identical to piggyBac were originally identified in the tephritid species B. dorsalis sensu strictu, where Southern analysis of transgenic lines and the host strain revealed at least 8-10 piggyBac-like elements in the genome (Handler and McCombs, 2000). PCR analysis of these elements from B. dorsalis s. s. and more recent sequencing ofpiggyBac-like elements from 14 species throughout the B. dorsalis species complex (consisting in total of 70 or more species) have led to the finding that all have 94% or greater nucleotide sequence identity to the original T. ni piggyBac (Bonizzoni et al., 2007; Handler et al., 2008). Yet none were found to be identical to IFP2, and only one, from B. minuta, was found to have an intact transposase open reading frame that has yet to be proven functional. The isolation of some complete B. dorsalis s.s. piggyBac-like elements as genomic clones and by inverse PCR indicates that these are complete elements with conserved terminal and subterminal sequences that are integrated into duplicated TTAA insertion sites.

The extensive evolutionary distance between T. ni and Bactrocera strongly suggests that the transposon moved recently between these species by horizontal transmission, and the separation of their geographical habitats raises the possibility that this movement may have been mediated by intermediary organisms. The likelihod that piggy-Bac elements exist in a wide range of insects, if not other animals, was supported by a database search for related sequences. A Southern blot survey (using IFP2 sequences as probe) for closely related piggyBac-like elements in more than 50 species showed the most clear evidence for multiple piggyBac elements in the fall armyworm Spodop-tera frugiperda, but hybridization patterns suggested that most of the elements are defective and non-functional (A. Handler, unpublished; see Handler, 2002b). This was supported by isolation of highly similar piggyBac-like sequences in S. frugiperda, Helicoverpa zea, H. armigera, and Macdunnoughia crassisigna (Zimowska and Handler, 2006; Wu et al., 2008), with a potentially functional element in M. crassisigna. Interestingly, these sequences share more similarities with the piggyBac-like elements in Bactrocera than with IFP2 (and the other T. nipiggyBacs), suggesting that these elements arose from a distinct lineage (Handler et al., 2008). More highly diverged piggyBac-related elements have also been found in the moths Heliothis virescens (Wang et al., 2006), Helicoverpa armigera (Sun et al., 2008), and Pectinophora gos-sypiella (Wang et al., 2010), with some having elements with uninterrupted transposase open reading frames and intact terminal sequences leaving open the possibility that they are competent to transpose. Indeed, we have found that at least one piggyBac-like element discovered in a larval T. ni genome, having a single amino acid residue change relative to IFP2 (Zimowska and Handler, 2006), is functional, based on transformation helper function (G. Zimowska and A. Handler, unpublished). The discovery of functional transposable elements in vivo is uncommon, likely due to their creating a genetic load resulting in organismal lethality, and thus mechanisms may exist in T. ni to repress piggyBac mobility. However, if IFP2 or a predecessor does exist in vivo, its presence in a derivative cell line is not surprising.

It is also intriguing to consider how horizontal transmission of piggyBac may have occurred, considering that the element was originally discovered by virtue of its transposition into an infectious baculovirus. This could potentially explain a distribution among lepidopterans, but the movement between moths and flies remains a mystery, although baculoviruses are capable of infecting (although not replicating) a wide range of organisms (Laakkonen et al., 2008). Understanding the interspecies movement of piggyBac, as well as all other vectors used for practical application, will be critical to understanding and eliminating risk associated with the release of transgenic insects.

mariner

Discovery, description, characteristics The mariner element was first discovered as an insertion element responsible for the white-peach (wpch) mutant allele of D. mauritiana (Haymer and Marsh, 1986; Jacobson et al., 1986). This particular allele was interesting when discovered because it was highly unstable, with reversions to wild type occurring at a frequency of approximately 10-3 per gene per generation. white-peach individuals also had a high frequency of mosaic eyes, at an approximate frequency of 10-3, suggesting somatic instability. Molecular analysis of the wpch allele indicated that it was the result of a 1286-bp transposable element insertion into the 5′ untranslated leader region of the white gene (Jacobson et al., 1986; see Figure 2). The mariner element is a Class II type transposable element with 28-bp imperfect inverted repeats with four mismatches. The element recovered from wpch contained a single open reading frame capable of encoding a 346 amino acid polypeptide (Jacobson et al., 1986). While the original wpch was highly unstable, another strain of D. mauritiana was discovered in which mosaicism of the eyes occurred in every fly (Bryan et al., 1987). This mosacism factor was found to be hereditable, and was referred to as Mos1 (Mosaic eyes). Mos1 was a dominant autosomal factor on chromosome 3, and was subsequently found to be identical to mariner except for six amino acid differences in the putative transposase coding region (Medhora et al., 1988). Mos1 encodes for a functional transposase, while the 346 amino acid polypeptide of the wpch mariner element was not a functional transposase.

One of the most notable characteristics of mariner and mariner-like elements (MLEs) is their widespread distribution. MLEs are found not only in insects and invertebrates, but also in vertebrates and plants (Robertson, 2000). Not long after the D. mauritiana mariner elements were described, a related element was discovered in the cecropin gene of the moth Hyalophora cecropia (Lidholm et al., 1991). Based on the sequence comparison between the mariner elements from D. mauritiana and H. cecropia, Robertson (1993) designed degenerate PCR primers and surveyed 404 species of insects for the presence of related sequences (Robertson, 1993). He found that 64 of the genomes examined contained MLEs, and within this group are five subgroups referred to as the mauritiana, cecropia, mellifera, irritans, and capi-tata subgroups (Robertson and MacLeod, 1993). Since that original analysis insect MLEs have continued to be discovered, and currently there are two additional subgroups recognized, known as mori and briggsae (Lampe et al., 2000). Additional subgroups are likely to be recognized in the future as additional representatives of this family of elements are found. As genome sequence data have accumulated, MLEs continue to be discovered (Liu et al., 2004; Zakharkin et al., 2004; Coy and Tu, 2005; Rouleux-Bonnin et al., 2005; Wang et al., 2005; Mittapalli et al, 2006; Ren et al, 2006; Haine et al., 2007; Carr, 2008; Rezende-Teixeira et al., 2008, 2010; Subramanian et al., 2008; Rivera-Vega and Mittapalli, 2010), and a recent analysis resulted in the identification of 15 subgroups within the mariner family of transposable elements (Rouault et al., 2009). Elements from different subgroups are typically about 50% identical at the nucleotide sequence level, while the transposases encoded by elements from different subgroups are usually between 25 and 45% identical at the amino acid level. A notable feature of the phylogenetic relationships of the MLEs is their incongruence with the phylogenetic relationships of the insects from which they were isolated. The implication is that many of these elements were introduced into their host genome via a horizontal gene transfer event (Robertson and Lampe, 1995a). The abundant examples of horizontal transfer of mariner elements have led to the conclusion that such transfers occur relatively frequently. Hartl et al. (1997) estimated that the rate of horizontal transmission of MLEs is about the same as the rate of speciation, at least within the D. melanogaster species subgroup. The widespread occurrence of horizontal transmission of MLEs has been proposed to be critical for the long-term survival of these elements. Horizontal transmission provides a means for invading naive genomes, where element proliferation can occur before inactivating influences of mutation and host regulation can occur (Hartl et al., 1997). This model continues to gain support from data describing the distribution and evolution of MLEs in insects (Lampe et al., 2003).

Although hundreds of MLEs have been reported, only three (Mos1 from D. mauritiana, Himar1 from Haemato-bia irritans, and Famarl from Forficula auricularia) have been demonstrated to be functional or active. Haemotobia irritans contains approximately 17,000 copies of Himarl, although all the copies examined were highly defective. Functional elements could be reconstructed based on the consensus sequence of Himarl, and then constructed by modifying the closely related Cpmarl element from the green lacewing, Chrysoperla plorabunda, to match the Himar consensus sequence (Robertson and Lampe, 1995b; Lampe et al., 1998). Purification of the trans-posase from a bacterial expression system and its use in an in vitro mobility assay demonstrated the functionality of the Himarl protein and the inverted terminal repeats of the element, but the elements were not active in insect cells (Lampe et al., 1996). Taking advantage of the ability of MLEs from insects to excise and transpose in bacteria, Barry et al. (2004) were able to screen approximately 2000 MLE open reading frames in the Famarl group of elements from the earwig, Forficula auricularia. Famarl is an abundant MLE in F. auricularia, with over 40,000 copies per genome. A total of 45 functional transposase open reading frames were discovered, and determining the sequence of 20 revealed unexpected diversity. As many as nine amino acid changes separated the ancestral element and the most diverged functional transposase (Barry et al., 2004). There were also differences in the relative activity of the elements in E. coli, ranging over 10-fold. The most active Famarl element was twice as active as Himarl in E.coli (Barry et al., 2004). While F. auricularia contains a diverse set of functional Famarl transposase coding regions, those proteins appear to be evolving neutrally, consistent with current models of transposable element evolution (Eickbush and Malik, 2002; Barry et al., 2004). The ability of Famarl elements to function in insects remains untested.

Structure-function relationships The trans-posases of MLEs belong to a large group of integrases and transposases that share a significant feature of their catalytic domains. Specifically, MLEs contain the highly conserved D, D, 35E amino acid motif within the active site of the protein, which is part of a conserved protein structure referred to as the integrase fold (Robertson, 2000; Richardson, 2006, 2009). This part of the active site interacts with divalent cations that are essential for catalysis. Transposase binds to the ITRs of the element, and gel retardation assays were used to assess the binding activity of eight mutant transposases with deletions at the N- or C-termini. Mutational and structural studies have led to a detailed understanding of the functional organization of the transposase protein, and a number of important details concerning the mechanism of Mosl transposition (Auge-Gouillou et al., 2001a, 2005; Dawson and Finnegan, 2003; Lipkow et al., 2004; Butler et al., 2006; Richardson et al., 2009). They were able to show that amino acids 1-141 were sufficient for binding to the ITRs. The ITR-binding domain of Mos1 transposase differs somewhat from that of Tc1 elements in that it is composed of two different structural motifs, a helix-turn-helix motif and an a-helical region (Auge-Gouillou et al., 2001a).

The structural organization of Mos1 appears to be important in determining the level of activity of the element. For example, the ITRs of Mos1 are not identical and differ in sequence at four positions, which affects the activity of the element in vitro. Auge-Gouillou et al. (2001b) reported a 10-fold higher affinity of Mos1 transposase for the 3′ ITR compared to the 5′ ITR. In addition, modified 5′ ITRs that were made to resemble 3′ ITRs at one of the four variable positions resulted in an increase in transposase binding. These investigators also showed that a Mos1 element with two 3′ ITRs had 104 times the transposition activity of the native ITRs (Auge-Gouillou et al, 2001b; Sinzelle et al, 2008; Casteret et al., 2009). This hyperactive double-ended configuration has not been tested in vivo, and did not result in increased element activity when tested in insects (Pledger et al., 2004).

The detailed molecular understanding of MLEs and their movement is beginning to permit rationally designed variants to be created and tested (Germon et al., 2009). Hyperactive transposase mutants of the Himar1 trans-posase were reported (Lampe et al., 1999), and one of the mutants contained two amino acid changes (at positions 131 and 137) in the ITR-binding domain of the protein. Although not tested directly, it is possible that these hyperactive mutants result in increased binding of the transposase, and, consequently, higher rates of movement. Paradoxically, neither Himar1 nor any of the hyperactive mutants showed any transpositional activity in insects (Lampe et al., 2000). The organization of MLEs and the transgenes they carry impact the activity of the elements (Casteret et al., 2009). Like most other transposable elements that have been tested, the MLEs have preferred integration sites, resulting in their non-random distribution in DNA target molecules (Crenes et al., 2009, 2010).

Host range of mariner The widespread distribution of MLEs in nature and the frequent examples of their horizontal transfer between species would seem to indicate that these elements have a broad host range. Empirical studies in which Mos1 has been employed as a gene vector in a wide variety of organisms support this conclusion. Mos1 has been used successfully to create transgenic D. melanogaster, D. virilis, Ae. Aegypti, and M. domestica (Lidholm et al., 1993; Lohe and Hartl, 1996a; Coates et al, 1998; Yoshiyama et al., 2000). In each of these species, the frequency of transformation was approximately 5%. This element has also been used to create transgenic B. mori cells in culture (Wang et al., 2000). In addition to transgenic insects, Mos1 has been used to create transgenic Leishmania, Plasmodium, zebrafish, and chickens (Gueiros-Filhos and Beverley, 1997; Fadool et al, 1998; Sherman et al, 1998; Mamoun et al., 2000). Similarly, the Himar1 element has been shown to function in E. coli, Archaebacteria, and human cells (Zhang et al., 1998, 2000; Rubin et al., 1999). Himar1, however, has not been shown to be active in D. melanogaster or any other insect species, for reasons that are not at all clear (Lampe et al., 2000).

Post-integration behavior The post-integration behavior of Mos1 has been investigated in D. melanogaster and Ae. aegypti. Mariner gene vectors used to create transgenic D. melanogaster have been found to be uncommonly stable, even in the presence of functional transposase. Lidholm et al. (1993) created two lines of transgenic D. melanogaster with a mariner vector derived from Mos1 and containing the mini-white gene as a genetic marker. When these lines were crossed to Mos1 transposase-expressing lines, eye mosaicism was found in only 1% of the progeny, while these same Mos1 expressing lines resulted in 100% mosaicism of the wpch element. Similarly, germ-line transposition occurred at rates of less than 1% (Lidholm et al, 1993), and Lohe et al. (1995) reported similar evidence for post-integration stability of mariner vectors. Lozovsky et al. (2002) suggested, after investigating the post-integration mobility of a number of mariner vectors containing different genetic markers in different locations within the element, that mariner mobility is highly dependent upon critical spacing of subterminal sequences and the inverted repeats. They found that vectors with simple insertions of exogenous DNA of varying lengths and in varying positions showed levels of somatic and germ-line excision that were at least 100-fold lower than that observed with uninterrupted mariner elements. Only vectors consisting of two (almost complete) elements flanking the marker gene showed detectable levels of both somatic and germ-line mobility. Approximately 10% of the insects with these composite vectors had mosaic eyes when transposase was present. Germ-line excision rates of approximately 0.04% were observed in these same insects. Again, these values are considerably less than those reported for uninterrupted elements. In addition to the potential importance of subterminal sequence spacing (Lozovsky et al., 2002), Lohe and Hartl (2002) suggested that efficient mobilization of mariner in vivo also depends on the presence of critical sequences located quite distant from the inverted repeats. Based on the mobility characteristics of about 20 mariner elements with a wide range of internal deletions, they concluded that there are three regions within the element that play an important role in cis. Region I is approximately 350 bp in length, and is located 200 bp from the left 5′ inverted terminal repeat. Region II is approximately 50 bp in length, and located approximately 500 bp from the right 3′ ITR. Region III is about 125 nucleotides in length, and located approximately 200 bp from the right ITR (Lohe and Hartl, 2002). While the presence of subterminal sequences that play a critical role in the movement of many Class II transposable elements is not unusual, what is uncommon in the case of mariner is the location of these cis-critical sequences. Their dispersed distribution within the element is unique, and, consequently, manipulating the element for the purposes of creating gene vectors and associated tools without disrupting these important relationships may be difficult.

The post-integration mobility of Mosl can also be regulated by non-structural aspects of the system, including "overproduction inhibition" and "dominant-negative complementation." Increasing the copy number of Mosl in the genome resulted in a 25% decrease in the rate of germ-line excision. Copy number increases in Mosl presumably lead to increased transposase levels, and, by an unknown mechanism, to the inhibition of excision (Lohe and Hartl, 1996b). High concentrations of transposase may lead to non-specific associations of the protein resulting in inactive oligomers of transposase. In addition, the presence of mutated forms of Mosl transposase can repress the activity of functional transposase. Because the trans-posases of other transposable elements act as dimers or multimers, it is thought that mutated Mosl transposases may become incorporated into multimers with functional transposases, thereby inactivating the entire complex (Lohe and Hartl, 1996b).

The possibility that transposase overproduction may negatively affect its own activity is a highly important concept in terms of vector system development. Most systems have the helper transposase under strong promoter regulation to optimize transpositional activity, though this may, indeed, be counterproductive. For mariner vectors, and potentially other systems, optimal transformation may require testing various helper promoters and a range of plasmid concentrations.

The post-integration mobility properties of mariner were also examined in the yellow fever mosquito Ae. aegypti (Wilson et al., 2003). As part of an effort to create an enhancer trapping and gene discovery technology for Ae. aegypti, Wilson and colleagues created non-autonomous mariner-containing lines, and lines expressing Mosl transposase. By creating heterozygotes between these two lines they attempted to detect and recover germ-line transposition events, but only a single germ-line transposition event was recovered after screening 14,000 progeny. Somatic transpositions were detected, and while precise estimates of rates of somatic transposition were not possible because of the detection method, the authors observed fewer than one event per individual, which they estimated to be an indication of a very low rate of movement. The vectors used by Wilson and colleagues resembled the simple vectors reported by Lozovsky et al. (2002) that had apparently disrupted spacing of the ITRs, and partial deletions of cis-critical sequences described by Lohe and Hartl (2002).

While the post-integration stability of mariner has been described in two species, and appears to be a general mobility characteristic of this element and not a reflection of a species-specific host effect, paradoxical observations remain to be explained. First, the use of mariner as a primary germ-line transformation vector in non-drosophilid insects and in non-insect systems is an effective means for creating transgenic organisms. Indeed, the host range of mariner as a gene transformation vector is unrivaled by any of the other gene vectors currently employed for insect transformation. Mariner has been used as a gene vector in microbes, protozoans, insects, and vertebrates. The rate of germ-line transformation using mariner-based vectors in insects is approximately 10% or less, and is comparable to the efficiency of Hermes, Minos, and piggyBac gene vectors. This raises the question of whether mariner vectors present on plasmids behave the same as mariner vectors integrated into insect chromosomes. Given the rates of germ-line integration from plasmids, it appears that the mariner vectors being used are not suffering from "critical spacing/critical sequence" defects. In addition, the behavior of mariner in vitro also differs from the behavior of chromosomally integrated elements. Tosi and Beverly (2000) demonstrated that only 64 nucleotides from the left end and 33 nucleotides from the right end of mariner were essential for transposition of a 1.1-kb vector in vitro. The rate of transposition of a minimal mariner vector in vitro was only two-fold less than that of a vector containing essentially a complete mariner element. These results suggest that mariner mobility has relatively simple sequence requirements, and that the role of subterminal sequences is minimal in vitro. These apparently conflicting data suggest that host factors may play an important role in the transposition process in vivo, and may influence the relative importance of cis sequences in the mariner transposition process. The broad distribution of MLEs and the host range of mariner/Mosl suggest, however, that host factors play little role in the movement of these elements.

The post-integration behavior of mariner/Mosl seems to indicate that this element will not be a good candidate for developing gene-finding tools such as promoter/enhancer trapping and transposon tagging systems in Ae. aegypti and perhaps other insects. On the other hand, if a high level of post-integration stability is desired, then mariner is an appropriate element to consider in insects. The potential of this element to be lost through excision or transposition is low, even in the presence of functional mariner trans-posase. As currently configured and used, mariner vectors may be considered suicide vectors in insects, since they essentially become dysfunctional upon integration.

MLEs in other insects While hundreds of MLEs have been described, few have been shown to be functional. The original mariner element from the white-peach allele was transpositionally competent, although it did not produce a functional transposase. Mos1 is a functional autonomous element, and has been the basis for constructing all mariner gene vectors that function in insects. Himar1 is a functional element from the irritans subgroup that was reconstructed based on multiple sequence comparisons of elements within this group. It has not been shown to be functional in insects, despite significant efforts to do so. Lampe et al. (2000) report that at least eight other elements from the other subgroups are likely to be active, or to be made active by minor modifications.

Next post:

Previous post: