Detection methods for fluorescent proteins
Once heterologous expression of GFP in nematodes was discovered, it was realized that use of the marker for whole body analysis of gene expression would require an optical system allowing a large depth of field and a stage with working space for culture plates. Up to this time, most epifluorescence systems were linked to compound or inverted microscopes that had limited field depth and capability to manipulate organisms under observation. This led to the development of an epifluorescence module using a mercury lamp that could be attached to a Leica stereozoom microscope system. Most major microscope manufacturers now market integrated epifluorescent stereozoom microscopes with capabilities for several filter systems.
A lower-cost alternative for GFP screening is use of a lamp module using ultra-bright blue-light emitting diodes (LEDs) with barrier filters, which attaches to the objective lens of most stereozoom microscopes (BLS Ltd, Budapest, Hungary). It costs considerably less than a mercury lamp system, but at present only has capabilities for detecting GFP and YFP.
The use of fluorescent protein markers, and especially multiple markers, will be greatly aided by the use of fluorescence-activated embryo sorters. A device that was first developed to sort Drosophila embryos expressing GFP (Furlong et al., 2001) has been modified and commercially marketed for Drosophila and other organisms as the COPAS system by Union Biometrica (Somerville, MA). The latest sorting machines are highly sensitive, having the ability not only to distinguish different fluorescent proteins, but also to discriminate between levels of fluorescence from the same protein. Thus, these systems may have enormous importance to the straightforward screening for transgenics, and more sophisticated assays such as those for enhancer traps. Practical applications could include the screening of released transgenic insects caught in traps (in systems adapted for adults), or for genetic-sexing of embryos having a Y-linked or male-specific fluorescent marker.
Transposon Vectors
Hermes
Discovery, description, and characteristics
Hermes is a member of the hAT family of transposable elements, and is related to the hobo element of D. melanogaster, the Ac element from maize, and the Tam3 element from Antirrhinum majus (Warren et al., 1994). The initial interest in this family of elements by those concerned with creating new insect gene vectors stemmed from two observations. First, during the middle and late 1980s, the mobility characteristics of the Ac/Ds element system were being extensively studied because the element was recognized as having great potential to serve as a gene-analysis and gene-finding tool in maize and other plants. In addition, the mobility properties of Ac/ Ds were being extensively tested in species of plants other than maize, and in almost every case evidence for Ac/Ds mobility was obtained (Fedoroff, 1989). Ac/Ds appeared to be a transposable element with a very broad host range – unlike, for example, the P element from D. melanogaster, which only functions in closely related species (O’Brochta and Handler, 1988). Because transposable elements with broad host ranges were of interest to those attempting to develop insect transformation technology, Ac-like elements warranted attention. The second significant observation at this time was that the hobo element from D. melanogaster had notable DNA sequence similarity to Ac/Ds, suggesting that it was a distant relative of this broadly active element (Calvi et al., 1991). Investigation into the host range of hobo using plasmid-based mobility assays (as described above) ensued (O’Brochta et al., 1994). It was during the investigation of hobo that Hermes was discovered (Atkinson et al., 1993). Atkinson and colleagues performed plasmid-based hobo excision assays in embryos of the housefly M. domestica as part of an initial attempt to assess the host range of hobo. Assays were performed in the presence of hobo-encoded transposase and hobo excision events were recovered, suggesting that hobo, like Ac/Ds, would have a broad host range. However, when the assays were performed without providing hobo-encoded transposase, hobo excision events were still recovered in M. domestica. The movement of hobo in the absence of hobo-transposase was completely dependent upon the inverted terminal repeats of hobo, and the resulting excision events had all of the characteristics of a transposase-mediated process. It was proposed that M. domestica embryos contained a hobo transposase activity, and that this activity arose from the transposase gene of an endogenous hobo-like transposable element (Atkinson et al., 1993). These investigators were eventually able to confirm their hypothesis, and the element they discovered was called Hermes (Warren et al., 1994).
Hermes is 2749 bp in length, and is organized like other Class II transposable elements in that it contains inverted terminal repeats and a transposase-coding region (see Figure 2). It contains 17-bp ITRs, with 10 of the distal 12 nucleotides being identical to the 12-bp ITRs of hobo. Hermes encodes for a transposase with a predicted size of 72 kDa, and, based on the amino acid sequence, is 55% identical and 71% similar to hobo transposase (Warren et al., 1994). The cross-mobilization of hobo by Hermes transposase that was proposed by Atkinson et al. (1993) was tested directly by Sundararajan et al. (1999). These investigators used plasmid-based excision assays in D. melanogaster embryos to show that hobo transposase could mobilize Hermes elements, and that Hermes trans-posase could mobilize hobo elements (Sundararajan et al., 1999). The phenomenon of cross-mobilization has important implications for the future use of transposable element-based gene vectors in non-drosophilid insects, and will be discussed further on. As is typical of trans-posable elements, Hermes is present as a middle repetitive sequence within the genomes of multiple strains of M. domestica, and in all populations of M. domestica examined there appeared to be full-length copies of the element (Subramanian et al., 2009). Hermes appears to be active in some strains of M. domestica, since excised Hermes elements in the form of covalently closed circles (episomes) were readily detected (O’Brochta et al., 2009).
Patterns of integration The integration behavior of Hermes has been examined in a variety of contexts. Sarkar et al. (1997a, 1997b) tested the ability of Hermes to transpose, using a plasmid-based assay, in five species of Diptera. They recovered transpositions of Hermes in the target plasmid at a frequency of approximately 10-3 in all species tested. In addition, they examined the distribution of 127 independent transposition events into the 2.8-kb plasmid used as a target in their assay, and observed a distinctly non-random pattern of integrations. Most notable was the existence of 3 sites that were targets for Hermes integration 10 or more times each. In an experiment in which any site used twice or more was considered a hotspot for integration, the 3 sites used 10 or more times constitute sites with unusual characteristics. The precise nature of those characteristics, however, could not be defined. The sites shared four of eight nucleotides of the target site in common (GTNNNNAC); however, other sites with this nucleotide composition were not equally attractive as integration sites, indicating that other factors must be influencing target choice. Saville et al. (1999) demonstrated that sequences flanking hobo integration hotspots were critical for determining the targeting characteristics of a site. These investigators were able to move an 8-bp hobo target site from plasmid to plasmid without losing its target characteristics, as long as they included 20 bp of flanking sequence on each side of the target. It was suggested that proximity to a preferred integration site increased the likelihood of a site being used as a target (Sarkar et al., 1997a). They found that sites 80- and 160-bp flanking the integration hotspot were also preferred integration sites. The authors suggested that nucleosomal organization of the target contributes significantly to the target site selection process, and contributes to the local juxtaposition of hotspots and flanking DNA.
Structure-function relationships Many Class II transposable elements contain a distinct amino acid motif within their catalytic domains, consisting of two aspartate residues and a glutamate. This DD35E motif can be found in many, but not all, Class II transposable elements. The presence of this motif in Hermes transposase was initially unclear. Bigot et al. (1996) proposed the existence of a DDE motif among members of the hAT family; however, they proposed that the second aspartate was replaced by a serine in Ac, hobo, and Hermes. Capy et al. (1996) concluded that hATelements, like P elements from Drosophila, do not contain the DDE motif, based on sequence alignments; Lerat et al. (1999) supported this conclusion based on the lack of similarity in predicted secondary structure of the transposase of members of the mariner/ Tc superfamily and hobo transposase. Michel et al. (2003) examined experimentally the importance of D402, S535, and E572 to the proper functioning of Hermes transposase. They found that mutations D402N and E572Q abolished transposase activity, while the mutations S535A and S535D had no effect on transposase activity. The work of Michel et al. (2003) provided the first experimental data to support the hypothesis that the positive charges of residues D402 and E572 are required for transposition. The authors concluded, based on these data, that D402, S535, and E572 do not constitute the catalytic center of Hermes transposase, because one of the residues was not essential for activity.
Zhou et al. (2004), using purified Hermes transposase protein and in vitro transposition reactions, found that, as expected, the element underwent cut-and-paste transposition, but this led to the creation of hairpin structures at the ends of donor DNA following excision (Zhou et al., 2004). Based on the structure of the reaction products and an analysis of the amino acid sequence, Zhou et al. (2004) concluded that there were significant similarities among Hermes transposase, the V(D)J recombinase RAG, and retroviral integrases (DDE transposases). The successful crystalization of Hermes transposase and the determination of the protein’s structure confirmed the presence of a retroviral integrase fold, and clearly links this element, albeit distantly, to other transposable elements containing that protein fold (Hickman et al., 2005).
Because Hermes transposase acts within the nucleus, it is expected to contain a nuclear localization signal to direct the mature transposase from the ribosome to the nucleus. Deletion and site-directed mutagenesis analysis were performed and demonstrated that the Hermes nuclear localization signal is located at the amino acid end of the protein and divided among three domains (Michel and Atkinson, 2003).
The inverted terminal repeats of transposable elements play an essential role in their mobility. Altering the sequence of ITRs can, depending on the element, lead to loss-of-function, hyperactivity of the element, or switching of the mode of transposition from a cut-and-paste mechanism to a replicative mechanism. Hermes contains imperfect ITRs, with a two base-pair mismatch within the ITR (Warren et al., 1994). In addition, a naturally occurring polymorphism in the terminal nucleotide of the right 3′ ITR exists. Elements with a cytidine in the terminal position of the right ITR have no activity within D. mela-nogaster, but are capable of undergoing an aberrant form of transposition in mosquitoes. Small pentanucleotide motifs in the subterminal regions of both Hermes and hobo have been found to be important for the mobilization of Hermes and hobo. The sequences GTGGC and GTGAC are interspersed throughout the subterminal region of the element, and similar repeats are present in the subterminal regions of Ac and are known to be transposase-binding sites. In Hermes, altering a single repeat can eliminate transpositional activity (Atkinson et al., 2001).
Hermes transposase is capable of dimerizing and one region of the protein critical for dimerization is located in the C-terminus of the protein, including amino acids 551-569. This region is not only essential for dimeriza-tion, but is also required for transposition activity. A second region that affects dimerization is located in the N-terminus of the protein, within the first 252 amino acids of the transposase. However, this region apparently plays a non-specific role in dimerization (P. W. Atkinson and K. Michel, personal communication). More recently, Hickman et al. (2005) found that transposition of Hermes was only observed when the protein formed hexamers.
Host range of Hermes Hermes has a wide insect host range, and has been found to function (as measured by either plasmid-based mobility assays or germ-line transformation) in at least 13 species of insects, including 11 flies, 1 beetle and 1 moth (Atkinson et al., 2001). Hermes functions rather efficiently in D. melanogaster, and transforms this species at rates of 20-40% (O’Brochta et al., 1996). In all other species tested the efficiency of transformation was considerably lower, and tended to be less than 10%. For example, Tribolium casteneaum was transformed at a rate of 1%, Ae. aegypti at 5%, C. quinquefasciatus at 11%, C. capitata at 3%, S. calcitrans at 4%, and Bicyclus anynana at 10.2% (Atkinson et al., 2001; Marcus et al., 2004). In all insects except mosquitoes, Hermes appeared to use a standard cut-and-paste type mechanism, as is typical of most Class II transposable elements. Such integrations are characterized by the movement of only those sequences delimited by the inverted terminal repeats, and the integrated elements are flanked by direct duplications of 8 bp. Integration of Hermes into the germ-line of Ae. aegypti and C. quinquefasciatus appears to occur by a non-canonical mechanism resulting in the integration of DNA sequences originally flanking the element on the donor plasmid. The amount of flanking DNA that accompanies the integration of Hermes in these mosquito species varies. In some cases, two tandem copies of the Hermes element were transferred to the chromosome and each copy was separated by plasmid DNA sequences (Jasinskiene et al. , 2000). Although these transposition reactions are unusual they are dependent upon Hermes transposase, since the introduction of Hermes-containing plasmid DNA in the absence of Hermes transposase failed to yield transformation events. The germ-line integration behavior of Hermes in mosquitoes is not unique, however; other elements being used as gene vectors, such as mariner and piggyBac, have occasionally shown similar behavior in Ae. aegypti (D. O’Brochta, unpublished). Transposition assays performed with plasmids in developing mosquito embryos and in mosquito cell lines showed that Hermes could transpose via a canonical cut-and-paste type mechanism under these conditions (Sarkar et al., 1997b). The basis for the difference in types of integration events between plasmid-based transposition assays and chromosomal integrations is unknown, but may reflect differences in somatic and germ cells. In Aedes, canonical cut-and-paste transposition has been readily detected in the somatic tissues of insects containing an autonomous element. Germ-line transposition in these same insects has not been detected. It has been suggested that mosquitoes might contain endogenous hAT elements that affect the ability of Hermes elements to be integrated precisely. An alternative suggestion is that Hermes may have a second mode of transposition, as do the transposable elements Tn7, IS903, and Mu, which utilize a replicative mechanism of integration. Such a mechanism would result in integration products that resemble those observed in the germ-line of Ae. aegypti and C. quinquefasciutus. Replicative transposition of Hermes has not been demonstrated experimentally, and direct tests of the "alternate mechanism" hypothesis have not been reported.
Hermes’ activity is not limited to insects; the element has be shown to be active in yeast (Schizosaccharomyces pombe) and planaria (Girardia tigrina), suggesting that it will have broad utility, and not just as an insect gene vector (Gonzalez-Estevez et al., 2003; Evertts et al., 2007; Park et al., 2009).
Post-integration behavior Once integrated into the genome of D. melanogaster, Hermes maintains its ability to be remobilized and has shown mobility characteristics similar to those of other transposable elements. Following the introduction of an autonomous Hermes element in which the transposase gene was under hsp70 promoter regulation, and also contained an EGFP marker gene under constitutive regulatory control of the actin5C promoter, Guimond et al. (2003) found the element continued to transpose in the germ-line at a rate of 0.03 jumps per element per generation. The element used in this study was also active in the somatic tissue, and the authors used this as a means of collecting approximately 250 independent transposition events. Analysis of somatic integration events revealed a number of interesting patterns. First, it was found that transpositions were clustered around the original integration event. On average, 39% of the Hermes transpositions recovered were intrachromosomal and 17% were within the same numbered polytene chromosome division. Of the new insertions, 10% were at sites within 2 kb of the donor element, indicating that Hermes, like other transposable elements, shows the characteristic of local hopping. Local hopping refers to the tendency of some elements to preferentially integrate into closely linked sites. Local hopping has been described for a number of elements and is likely to be a general characteristic of Class II transposable elements, although the mechanistic basis for this behavior is unknown. Certain regions of the D. melanogaster genome, as defined by numbered divisions of the polytene chromosomes, are preferred as integration sites, with these regions being repeatedly targeted by Hermes. The observed clustering of independent transposition events in regions of the chromosome seems to reflect undefined aspects of the transposition process that might be influenced by the chromatin landscape. With one exception, the clustering observed by Guimond et al. (2003) was not correlated with any common feature of the chromosomes or the genes within a region. This type of non-random pattern of integration with regional differences has also been reported for other elements. Interestingly, there does not seem to be any strong correlation between the preferred insertion-site regions of the elements P, hobo, and Hermes, at least with respect to chromosome 3 of D. melanogaster (see Figure 7 in Guimond et al., 2003). Guimond et al. (2003) also observed a notable clustering of integrations in polytene chromosome division 5. Of the 11 integration events recovered from division 5 (3.2% of all the transposition events examined), 8 were within the 2.7-kb segment of DNA upstream of the cytoplasmic actin gene, actin5C. This same 2.7-kb segment of the 5′ regulatory region of actin5C was also present within the autonomous Hermes element, as a promoter for the EGFP marker that the investigators tracked as it jumped within the genome.
The strong clustering of transpositions in a target sequence that is homologous to a sequence contained within the vector has been referred to as "homing." This type of target-site selection bias was first described for P elements, and has been reported on a number of occasions. It was initially reported as a strong bias in the integration site distribution of a number of primary germ-line integration events in which a P element containing the engrailed gene preferentially integrated into the engrailed region of the host genome (Hama et al., 1990; Kassis et al., 1992). A similar biasing of integration site selection was also observed with P elements containing Antenna-pedia and Bithorax regulatory sequences (Engstrom et al., 1992; Bender and Hudson, 2000). More recently, Tail-lebourg and Dura (1999) reported a remarkable example of homing of a remobilized P element in D. melanogas-ter. This element contained either an 11-kb or a 1.6-kb fragment of the 5′ region of the linotte gene, and it was found that 20% of the remobilized elements integrated into the 5′ region of the linotte gene. Insertions in this case were highly localized, and most occurred within a 36-bp fragment of the linotte regulatory region. Hermes homing indicates that the phenomenon is not element-specific, but may be a general characteristic of Class II elements. Guimond et al. (2003) suggested that homing was a special case of local hopping, and the physical proximity between donor elements and target sites seems to underlie the phenomenon of local hopping. The presence of transgene regulatory sequences (e.g., actin5C 5′ region) may promote tethering of the donor elements to similar regulatory regions via proteins with common DNA-bind-ing sites. Deliberate tethering of transposable elements to selected sequences may be a means to regulate target-site selection and to minimize the detrimental mutagenic effects of transposable element integration (Bushman, 1994; Kaminski et al, 2002).
The post-integration behavior of the same autonomous Hermes element described above in Ae. aegypti had quite different characteristics. In this case germ-line transposition of the autonomous Hermes element was never detected, and it should be noted that the primary integration events in the germ-line involved the integration of DNA sequences flanking the element (Jasinskiene et al, 1998, 2000). Despite the fact that the element was intact and functional transposase was expressed, the element was immobile in the germ-line. This was not the case, however, in the soma of Ae. aegypti, where Hermes excision and cut-and-paste transpositions were readily detected. Transposition events in the soma had all the hallmarks of Class II cut-and-paste integration. Only those sequences precisely delimited by the ITRs moved, and integration resulted in the creation of 8-bp direct duplications at the target site. Excision of Hermes was imprecise, and led, in some cases, to the creation of small deletions. The basis for the difference in behavior of the Hermes element in the germ-line versus the somatic tissue of Ae. aegypti is unknown. Clearly the post-integration behavior of Hermes in this species will influence how this element will be employed, and in situations where germ-line stability is essential Hermes will be particularly useful. It will not be useful, in its present form, for constructing gene-finding tools such as enhancer and promoter traps that rely heavily on transposable element vector remobili-zation to be effective.
Extrachromosomal forms of Hermes Excision of Hermes in M. domestica, and autonomous Hermes elements in D. melanogaster and Ae. aegypti, leads to the formation of circularized Hermes elements in which the terminal inverted repeats are covalently jointed end-to-end in various ways following the excision reaction (O’Brochta et al., 2009). The most common configuration results in the ends being joined end-to-end with a short spacer sequence between them. The spacer sequence was most often 1, 3, or 4 bp, but could also be as much as 200 base pairs. The extrachromosomal Hermes elements found in M. domestica are particularly interesting, because they have been found in all populations tested and in great abundance in somatic tissue. These data provide evidence for the somatic activity of Hermes in the insects from which it was originally isolated. Circularized forms of excised transposable elements of a number of types have been reported in the past (Sundraresan and Freeling, 1987). For example, circularized forms of Ac/Ds have been described, as well as Minos (Arca et al., 1997; Gorbunova and Levy, 1997), yet the significance of extrachromosomal forms of transposable elements has remained unclear. In some cases the circularized elements do not contain intact terminal inverted repeats, and consequently the elements are not expected to be integration-competent. Based on rather limited data, it has generally been concluded that such forms represent byproducts of aborted or interrupted transposition reactions. A recent study of the extrachromosomal forms of Hermes suggests that these elements may have some biological significance. Some circularized forms of Hermes elements with intact inverted terminal repeats were found to be capable of integrating into the genome of D. melanogaster, indicating that they could contribute to forward transposition (O’Brochta et al., 2009). The ability of circularized forms of excised Hermes elements to reintegrate may impact the potential of this element to be transmitted both vertically and horizontally. Circular, extrachromosomal forms of Hermes were readily detected in unfertilized eggs of M. domestica that contain native genomic copies of the Hermes element, and in D. melanogaster that contain active autonomous Hermes elements, strongly suggesting that they are capable of being transmitted maternally. Clearly, maternal transmission of active, integration-competent extrachromosomal forms of Hermes has the potential to facilitate an increase in frequency of the transposable element within populations; however, to date there are no data regarding the significance of extrachromosomal forms on element dynamics in populations.
hAT elements in other insects hAT elements have been widely detected in insect genomes. The Queensland fruit fly Bactrocera tryoni contains members of at least two distinct hAT-like transposable elements (Pinkerton et al., 1999). Homer is a 3789-kb element whose sequence is 53% identical to Hermes and 54% identical to hobo. The transposase coding region is approximately 53% identical and 71% similar to the transposases of Hermes and hobo, respectively. Similarly, the ITRs of Homer, which are 12 bp in length, are identical to those of the hobo and Hermes elements at 10 of 12 positions. There are also Homer-like elements within B. tryoni. There are fewer than 10 copies per genome, and, while these elements have not been fully characterized, a conceptual translation of the transposase of this Homerlike element reveals 48% identity and 66% similarity to the transposase of Homer. These Homer-like elements are as similar to hobo as they are to Homer. Although Homer appears to be weakly functional in D. melanogaster, based on plasmid-based excision assays, all Homer-like elements contain inactivating frameshift mutations.
The Australian sheep blowfly L. cuprina contains a nonfunctional hAT element called hermit. Hermit was initially found by low stringency hybridization screening of an L. cuprina genomic library using a DNA probe homologous to hobo (Coates et al., 1996). Hermit is 2716 bp in length and contains perfect 15-bp ITR, the distal 12 of which are identical to the hobo ITRs at 10 of 12 positions. Although inactive because of frameshift mutations within the transposase coding region, its amino acid sequence is 42% identical and 64% similar to hobo transposase. Hermit is unusual in that it is present as a unique sequence within L. cuprina, in contrast to multiple copies that exist for most transposons. Although present only once within this species, it does appear to have arisen within the genome as a result of transposition, since the existing copy of the element is flanked by an 8-bp direct duplication of a sequence that is similar to the consensus target site duplication derived from other hAT elements. Hermit appears to have become inactivated soon after integrating into the L. cuprina genome.
Several hAT elements have been discovered in tephritid fruit flies, using a PCR approach similar to that used to discover Hermes (Handler and Gomez, 1996). Of these elements, a complete hAT transposon (hopper) was isolated from a genomic library of the wild Kahuku strain of the Oriental fruit fly B. dorsalis, using the Bd-HRE PCR product as a hybridization probe (Handler and Gomez, 1997). A complete 3120-kb element was isolated, having 19-bp ITRs; however, the putative transposase-coding region is frameshifted and does not have a duplicated 8-bp insertion site, suggesting that it had accumulated mutations and was non-functional. The Kahuku sequence was used to isolate additional hopper elements using an inverse and direct PCR approach, and a new 3131-bp hopper was isolated from the B. dorsalis white eye strain (Handler, 2003). This element has an uninterrupted coding region and an 8-bp duplicated insertion site consistent with possible function. Preliminary experiments in which transformants have been generated in D. melanogaster and A. suspensa using a hopperwe vector marked with DsRed, and an hsp-hopperwe helper, support autonomous function for the hopperwe element (A. Handler and R. Harrell, unpublished).
Notably, hopper is highly diverged from all other known insect hAT elements, and its transposase is distantly, yet equally, related to the coding regions of hobo and Ac. Of the terminal 12 nucleotides only 5 are identical to those of hobo, while 6 are identical to the ITRs of Homer. Handler and Gomez (1996) inferred the presence of an active hobo-like transposable element system in the Mediterranean fruit fly Ceratitis capitata, because non-autonomous hobo elements from D. melanogaster were active in excision assays performed in medfly embryos in the absence of any experimentally provided hobo trans-posase. The element Cchobo was subsequently isolated, and its transposase coding region was found to be 99.6% and 73.3% identical to hobo and another C. capitata hAT element, CcHRE, respectively (Gomulski et al., 2004). No subsequent tests of Cchobo’s functionality as an insect gene vector have been reported.
hAT elements have been also reported in the human malaria vector, An. gambiae. Approximately 25 sequences resembling hAT transposases were discovered, although none appeared to be part of an intact transposable element. More recently, however, genomic DNA database-search criteria were used based on unique aspects of hAT transposable elements such as length and spacing of inverted terminal repeats, and the characteristics of hAT element target sites. This search revealed a hAT element An. gambiae that contained perfect 12-bp ITRs flanked by 8-bp direct duplications and a 603 amino acid transposase open reading frame that appeared to contain no internal stop codons. This element (Herves) is most closely related to hopper and is transpositionally active in D. melanogas-ter (Arensburger et al., 2005). Subramanian et al. (2007) examined copy-numbers, integration-site polymorphisms, and nucleotide diversity of Herves in individual An. gam-biae collected largely in East Africa, and concluded that Herves appears to have been introduced into this mosquito lineage prior to the recent diversification of species that now form the An. gambiae species complex. Integration-site polymorphism data are consistent with the element having been active in the recent past, although the authors did not test whether the element is active in contemporary populations of An. gambiae.
hAT transposable elements are well-represented in the genus Drosophila; Oritz and colleagues reported finding multiple new hAT elements in the genomes of 10 species of Drosophila for which whole-genome DNA sequence was available (Ortiz and Loreto, 2009; Ortiz et al., 2010). Two of the elements discovered by these investigators were related to the Herves element from An. gambiae, and, based on the apparent structural integrity of these elements, they were thought to be active (Depra et al., 2010). More extensive surveys for the presence of these Herves-like sequences (called hosimary by Depra et al., 2010) in 52 species in the family Drosophilidae revealed the presence of hosimary in members of the melanogaster species group and in distantly related Zaprionus indianus. The high degree of sequence similarity among hosimary elements in Drosophila and Zaprionus suggests that horizontal transfer may have played a role in the history of these elements. Other examples of hAT elements within drosophilids having high sequence similarity and discontinuous interspecific distributions, suggesting horizontal transfer, have also been reported. For example, Mota et al. (2010) studied the evolution of hAT elements related to Homo3 from Drosophila mojavensis, and Howilli3 from Drosophila willistoni, in 65 species of drosophilids, and found a high degree of DNA sequence similarity among elements isolated from different subgenera. These authors suggested that horizontal transfer was the best explanation for their observations. Although much of the study of insect hAT elements has been performed in Diptera, other orders of insects also harbor hAT elements, but none have been used as a gene vector (Borsatti et al., 2003).
