Minos
The first germ-line transformation of a non-drosophilid insect mediated by a transposon-based vector system was achieved with the Minos element. Minos was originally isolated as a fortuitous discovery in D. hydei during the sequencing of the non-coding region of a ribosomal gene (Franz and Savakis, 1991). Minos was found to be a 1.4-kb element having, unlike the other Class II trans-posons used as vectors, relatively long inverted terminal repeats of 255 bp, with its transcriptional unit consisting of two exons (see Figure 2). Additional Minos elements were isolated from D. hydei, having small variations of one or two nucleotides, though the new elements had a transition change that restored the normal reading frame, allowing translation of a functional transposase. The sequence homology, general structure, and TA insertion-site specificity placed Minos within the Tc transposon family (Franz et al., 1994). Minos was first used to transform D. mela-nogaster, with Minos-mediated events demonstrated by sequencing insertion sites and remobilization of integrations (Loukeris et al., 1995b). The first non-drosophilid transformation with Minos was achieved in a medfly white eye host strain using a cDNA clone for the medfly white gene as a marker (Zwiebel et al., 1995), at an approximate frequency of 1-3% per fertile G0 (Loukeris et al., 1995a). Minos has since been used with fluorescent protein markers to transform another tephritid pest species, the olive fruit fly Bactrocera oleae (Koukidou et al., 2006); the silkworm moth Bombyx mori (Uchino et al., 2007); An. stephensi (Catteruccia et al., 2000); and T. castaneum (Pavlopoulos et al., 2004). Recently, transformation frequencies have been substantially increased in Drosophila and medfly by use of in vitro synthesized transposase mRNA as a helper (Kapetanaki et al., 2002).
Although Minos has been only occasionally used for insect transformation, embryonic and cell-line mobility assays in several insect species in the Diptera, Lepidoptera, and Orthoptera have indicated a broad range of functions.
Notably, Minos transposition in the cricket Gryllus bimac-ulatus was driven by transposase regulated by a Gryllus actin gene promoter, and not by the Drosophila hsp70 promoter that has been widely used in dipterans (Zhang et al., 2002). The broad function of the Minos vector is further supported by its ability to transpose in a mouse germ-line and in invertebrates (Drabek et al., 2003; Pav-lopoulos et al. , 2007).
Minos structure places it within the mariner/Tc transposon superfamily, though knowledge of the distribution of Minos is thus far limited to the genus Drosophila (Arca and Savakis, 2000). In Drosophila, Minos is clearly widely distributed in the Drosophila and Sophophora subgenus, though discontinuously in the Sophophora. As noted for the hAT, mariner, and piggyBac elements, Minos may have also undergone horizontal transfer between Drosophila species.
Tn5
Tn5 is one of a number of very well-characterized trans-posable elements from prokaryotes. Recently, hyperactive forms of this element have been created in the laboratory that have proven to be the basis for the development of a number of commercially useful genomics tools (Epicentre, Madison, WI; http://www.epicentre.com) (Goryshin and Reznikoff, 1998). Tn5-based genomics tools can be used in a wide variety of bacterial species, and, given the system’s independence from host encoded factors, might be applicable to eukaryotic systems as well (Goryshin et al., 2000). Efforts to use Tn5 as an insect gene vector have been successful.
Tn5 is a prokaryotic transposon, 5.8 kb in length, that is often referred to as a composite transposon because it consists of five independently functional units (for review, see Reznikoff, 2000; see also Figure 2). It contains three antibiotic resistance genes that are flanked by 1.5-kb inverted repeat sequences. Each inverted repeat is actually a copy of an IS50 insertion sequence, which are themselves functional transposons. Each IS50 element contains 19-bp terminal sequences known as OE (outside end) and IE (inside end), and while OE and IE are very similar, they are not identical. IS50 also encodes for two proteins: transposase (Tnp) is 476 amino acids long and catalyzes transposition, while the second protein is an inhibitor of transposition (Inh). The IS50 elements present at each end of Tn5 are not identical, and only IS50R is fully functional. IS50L contains an ochre codon that prematurely terminates the Tnp and Inh proteins, resulting in a loss of function of both proteins.
The transposition reaction and all of the components involved in the reaction have been studied in great detail (Reznikoff et al., 1999). Transposition proceeds by a cut-and-paste process involving binding of Tnp to the end sequences followed by dimerization of the bound Tnp to form a synaptic complex. Cleavage at the ends of the element results in an excised transposon with bound trans-posase that interacts with a target DNA molecule. Strand transfer results in the integration of the element into the target, and, in vitro, this reaction requires only a donor element, a target DNA molecule, transposase, and Mg2+ (Goryshin and Reznikoff, 1998). Modifications of both the transposase and the terminal 19-bp sequences have led to the creation of Tn5 elements consisting of little more than two copies of end sequences that can be mobilized 1000-fold more efficiently than an unmodified Tn5 element. This hyperactive Tn5 system has been developed into a powerful tool for genetic analysis of a variety of organisms. Tn5 has been attractive as a broad host-range genomics tool because its pattern of integration is random and its biochemical requirements very simple. Tn5 has been shown to function in a variety of bacterial and non-bacterial systems.
Current insect transformation protocols consist of microinjecting a mixture of two plasmids into preblasto-derm embryos (see section 4.6.3). One plasmid contains a non-autonomous transposable element with the trans-genes and genetic markers of interest, while the second plasmid contains a copy of the transposase gene. Transient expression of the transposase gene is required post-injection, and is followed by element excision and integration. Previous experiments examining the frequency of element excision in Hermes, mariner, Minos, and piggyBac from plasmids injected into insect embryos along with "helper" plasmids showed that only one plasmid per 1000 injected underwent an excision event. Therefore, 99.9% of the donor plasmids introduced into insect embryos contributed nothing to the transformation efforts. The introduction of pre-excised elements configured as active intermediates, such as synaptic complexes, was considered a means to permit higher integration rates and overall efficiency of transformation.
Transgenic Ae. aegypti were created using a Tn5 vector containing DsRed under 3xP3 regulatory control (Rowan et al., 2004). Pre-excised vectors in the form of synaptic complexes were injected into preblastoderm embryos; 900 adults were obtained from the injected embryos, and families consisting of approximately 10 G0 individuals were established. Two families of G0 individuals produced transgenic progeny for an estimated transformation frequency of 0.22% (2/900). Analysis of the transgenic progeny showed that multiple integrations of Tn5 occurred in each line. The patterns of integrations were complex, with evidence of the Tn5 vector integrating into Tn5 vector sequences. The integration of the vector into copies of itself, followed by the integration of the resulting con-catamers, was very unusual, and in no case was a simple cut-and-paste integration of the Tn5 vector found with characteristic 9-bp direct duplications flanking the element. The complex pattern of Tn5 integration was thought to be a direct consequence of injecting pre-assembled intermediates that were inactive in the absence of Mg2+. Therefore, as soon as the synaptic complexes were injected they became activated, and the first target sequences the elements were likely to encounter were other Tn5 synaptic complexes. At the time of injection, Ae. aegypti embryos only contained approximately four to eight nuclei, making genomic target DNA relatively rare. Furthermore, the synaptic complexes injected were expected to have a very short half-life. Therefore, although active intermediates were being introduced, a number of factors contributed to the inefficiency observed with this system, including a short half-life of the active intermediate and low numbers of genomic target sequences. Injecting binary plas-mid systems (as is done with Hermes, mariner, Minos, and piggyBac), while relatively inefficient in producing active transposition intermediates, achieves persistence over an extended period of time. Consequently, more target genomes are exposed to active vectors over a longer period of time, resulting in higher transformation rates. The limitation of injecting synaptic complexes is unlikely to be specific to the Tn5 system, and similar approaches with other insect gene vectors are likely to encounter similar problems. It should be noted, however, that the results of Rowan and O’Brochta demonstrate that Tn5 is functional in insects, and, while injecting active intermediates is not recommended, using Tn5 in a more conventional binary plasmid system consisting of a donor and helper plasmids is likely to be a viable option for creating trans-genic insects.
Improved Transposon Vectors for Basic and Applied Transformation
Stabilization of transposon vectors Unlike transformation systems, in which genomic integrations result from the recombination of introduced DNA, transposon-based vectors are subject to remobilization by the intended or unintended presence of functional transposase. Vector remobilization is a desired result for insertional mutagenesis protocols that require the repetitive insertion, or "hopping," of the vector after injection of helper plasmid or mating to a jumpstarter strain having a genomic source of transposase (Brand et al., 1994; Horn et al., 2003). However, for the development of transgenic strains for most applied purposes, and especially for programs requiring their field release, stable vector insertions are required, if not essential, for maintaining stable lines and ensuring that transgenes do not move between species by horizontal interspecies transfer. While these may be rare events, horizontal transmission is thought to be a natural phenomenon responsible for the movement and spread of transposable elements among species (Robertson and Lampe, 1995b; Lampe et al., 2003). For non-autonomous vectors whose transposase coding region is deleted or interrupted, such mobilization could be catalyzed in trans by an unintended source. This could be from the undetected presence of a functional transposon in the genome of the host insect, but, more dauntingly, by a similar transposon that is less likely to be detected, whose transposase can cross-mobilize the vector. Such cross-mobilization has indeed been experimentally demonstrated for the hobo and Hermes elements within the hAT transposon family (Sundararajan et al., 1999).
While cross-mobilization of a defective non-autonomous vector may be rare in nature, the large population scale and pressures of mass-reared transgenics for biocon-trol release programs could be favorable for the selection of such events (Robinson et al., 2004), raising serious ecological concerns for transgenic insect release. Thus, the post-integration stabilization of transposon-based vectors was thought to be a critical need for the applied use of transgenic strains (Handler, 2004). Two similar approaches have taken advantage of mobilization requiring the transposon 5′ (or left arm) and 3′ (or right arm) inverted terminal repeat sequences. If either one or both of the termini are deleted, transposon mobility – or remo-bilization – is eliminated. This was first achieved in Dro-sophila with a piggyBac vector having an internal tandem duplication of the left (L) 5′ terminus, and a single right (R) 3′ terminus in an L1-L2-R1 orientation, with distinguishable marker genes in between L1-L2 and L2-R1. This resulted in a vector, pBac{L1-PUbDsRed1-L2-3xP3-ECFP-R1}, that could have the L1-PUbDsRed1 trans-gene sequence stabilized by remobilization, or deletion, of the L2-3xP3-ECFP-R1 sequences (Handler et al., 2004). This was possible by first integrating the entire vector into a host genome, with subsequent remobilization of the L2-3xP3-ECFP-R1 subvector by mating to a jumpstarter strain (having phspBac helper integrated in a Minos vector). This deleted the ECFP marker along with L2, and the only 3′ (R1) terminus, thus stabilizing the remaining L1-PUbDsRed1 sequences with respect to any unintended source of transposase. Extensive efforts to remobilize the L1-PUbDsRed1sequences proved their stability, and the same stabilization method was also achieved in the teph-ritid pest species A ludens (Meza et al., 2010).
The single terminus deletion method was subsequently modified so that none of the original vector sequences remained, by essentially creating a dual vector with intervening DNA sequences in an L1-marker1-R1-marker2-L2-marker3-R2 orientation (Dafa’alla et al., 2006). For this vector, remobilization of both the L1-marker1-R1 and L2-marker3-R2 subvectors resulted in only the marker3 DNA left integrated in the genome, with no transposon sequences remaining to facilitate mobilization. While theoretically an improvement over the single-arm deletion method, the dual-arm deletion method is more cumbersome, and likely offers only a minor advantage in terms of transgene stability.
Vectors for gene targeting Transposon-based vector insertions in a host genome are typically random, though different transposons have varying insertion site sequence biases. This lack of specificity is highly useful for insertional mutagenesis screens such as transposon-tagging and enhancer traps, making this type of transformation a valuable tool for functional genomic analysis. However, there are negative effects of random integrations for both basic and applied studies. Gene expression from different insertion sites often varies due to genomic position and enhancer effects that result in abnormal, often suppressed, gene expression or misexpression with respect to tissue or development. This makes comparative gene expression and functional analyses problematic, if not unreliable, and makes creation of optimal strains for applied use difficult. Furthermore, insertional mutations often have deleterious, if not lethal, effects on transgenic strain viability and fitness, compromising their use for biological control programs (Catteruccia et al., 2003; Irvin et al., 2004).
The disadvantages of random vector insertions can be minimized by targeting transgene integrations to predefined genomic sites devoid of known coding or regulatory function that are minimally affected by position and enhancer effects. Targeting transgene insertions also allows true allelic comparisons of gene expression when studying gene structure-function relationships. The need for targeting motivated the development of site-specific targeting systems in insects based upon the phiC31 inte-grase (Groth et al., 2004), and the FRT/FLP (Horn and Handler, 2005) and loxP/Cre (Oberstein et al., 2005) recombinase systems. All of these systems have been tested, initially in Drosophila, by introducing recombination sites into the genome with transposon vectors. Secondary insertions were then made, with plasmids having the appropriate recombination sites co-injected with their respective integrase or recombinase helpers. For the applied use of transgenic strains, the strategy would be to create several strains with differing target sites that would be characterized in terms of transgene expression, in addition to strain viability and fitness. Optimal target site strains would then be used for subsequent integrations and genetic manipulations.
Initially, target sites were created by using the same dual heterospecific FRT (Senecoff et al., 1985) or loxP (Hoess et al., 1985) recombination sites in the target site and donor plasmid. Both systems have recombination sites that consist of two 13-bp inverted repeats separated by an 8-bp spacer that specifically recombine with an identical site in the presence of recombinase. Heterospecific sites have sequence variations in the spacer regions, and since only identical sites can recombine, use of the dual sites allows only reliable double-recombination. This results in the exchange of DNA between the sites in the donor plasmid and the sites in genomic target, also known as recombinase-mediated cassette exchange (RMCE). This was assessed in the FRT system by interconvertible fluorescent protein markers (Horn and Handler, 2005) that could also be used in non-drosophilids, while the loxP system used Drosophila mutant-rescue markers (Oberstein et al., 2005). DNA introduced by FRTRMCE contained a terminal piggyBac sequence that allowed subsequent stabilization of the target site piggyBac vector. While RMCE is highly effective in Drosophila, thus far it has not been successfully tested in other insect species, though there is no theoretical reason for it not to be equally effective.
The phiC31 integrase-based system has also been tested in Drosophila (Groth et al., 2004), where genomic attP sites, introduced by P element transformaton, were targeted in a unidirectional fashion in the presence integrase by plasmid vectors having an attB site. Targeting occurred at relatively high frequencies, though drawbacks of this system are that the entire attB-containing plasmid is inserted into the attP-target site, additional heterospecific attachment sites do not exist that might allow re-targeting of the same site, and occasionally genomic "pseudo" attP sites may be targeted. Still, using dual attB and attP sites in the target and donor sequences, an RMCE system has been created (Bateman et al., 2006). However, the sequence exchange is irreversible, though further targeting may be possible using the FRT or loxP systems. Nevertheless, the fC31 single recombination system has been successfully tested in mosquito species (Nimmo et al., 2006; Labbe et al., 2010) and a tephritid species (Schetelig et al., 2010). In addition, medfly targeting has been used to introduce a piggyBac terminus, allowing target site stabilization.
Transformation Methodology
The technical methodology for insect transformation has largely remained the same or has only been slightly modified from the techniques originally used to transform Drosophila. The references cited for P transformation are relevant to this, as well as several articles that focus on methods for non-drosophilid transformation (Handler and O’Brochta, 1991; Morris et al, 1997; Handler, 2000). The most variable aspect of this method is the preparation of embryos for DNA microinjection, though, arguably, the lack of new techniques for DNA introduction has been the primary limitation in the more widespread use of the technology. While all successful insect transformations have utilized microinjection, variations on this method have been necessary for different types of embryos, and most of the procedures must be tested empirically and modified for particular insect species. This may be extended to different strains and for a variety of local ambient conditions, including temperature and humidity. The apparatus for microinjection is usually the same for all species, though a wide variety of variations and modifications are possible and sometimes required.
The basic equipment includes an inverted microscope or a stereozoom microscope with a mechanical stage having a magnification up to 60-80x, a micromanipulator that is adjustable in three axes with an appropriate needle holder, and a means to transmit the DNA into the egg. For dechorionated eggs, transmitted light allows precise positioning of the needle within the egg posterior, while direct illumination is needed for non-dechorionated eggs, which typically include those of mosquitoes and moths.
The standard for gene transfer methodology in general, and embryo microinjection in particular, was originally developed for Drosophila. The standard method involves collecting preblastoderm embryos within 30 minutes of oviposition, and dechorionating them either manually or chemically. The timing of egg collection and DNA injection is related to the need to inject into pre-blasto-derm embryos during a phase of nuclear divisions previous to cellularization. This allows the injected DNA to be taken up into the nuclei, and specifically into the primordial germ cell nuclei that are the gamete progenitors. For Drosophila, cellularization of the pole cells begins at approximately 90 minutes after fertilization at 25°C, with blastoderm formation occurring about 30 minutes afterwards. The timing of these events and the location of the pole cells varies among insects, and thus some knowledge of early embryogenesis in the desired host insect is highly advantageous. In the absence of this information for a particular species, the most prudent time of injection would be the earliest time after oviposition that does not compromise viability.
Embryo Preparation
Manual dechorionation of Drosophila eggs is achieved by gently rolling the eggs on double-stick tape with a forcep until they pop out. While gentle on the eggs and requiring little desiccation time, manual dechorionation is tedious, and has not been applicable to any other insect. Chemical dechorionation is typically achieved by soaking eggs in a 50% bleach solution (2.5% hypochlorite) for 2-4 minutes and washing them at least three times in 0.02% Triton-X 100. Tephritid fruit fly eggs usually have thinner chorions that can be dechorionated in 30% bleach (1.25% hypo-chlorite) in 2-3 minutes, but this must be determined empirically, since they are easily over-bleached, resulting in death, either directly or after injection. Some species, such as M. domestica, can be only partially dechorionated, but bleached eggs can be released from the chorion by agitation. We have found the simplest and most precise method for bleach dechorionation with rapid washes is by using a 42-mm Buchner funnel with a filter flask attached to a water vacuum. Eggs can be washed into the funnel on filter paper and swirled within the funnel, with the solution gently sucked out by regulating the water flow or the seal between the funnel and flask. The last wash is done on black filter paper that allows the eggs to be easily detected, which facilitates their mounting for injection (see below).
Many insect eggs cannot be dechorionated without a high level of lethality, and must be injected without dechorionation. These include those of most moth and mosquito species. Drosophila and tephritid flies can, similarly, be injected without dechorionation, and while embryo viability after injection is often lower than for dechorionated eggs, the frequency of transformation in surviving embryos is often higher. It is more difficult to determine a precise site for injection in non-dechorion-ated eggs, though this can be aided by adding food coloring to the DNA injection mix.
After dechorionation, fruit fly embryos are typically placed on a thin strip (~ 1 mm) of double-stick tape placed on a microscope slide or 22 x 30 mm cover slip, though use of a cover slip is more versatile for subsequent operations. A thin strip of tape is suggested due to anecdotal reports of toxic solvents from the tape affecting survival, though some particular tapes are considered to be non-toxic (3M Double Coated Tape 415; 3M, St Paul, MN) and some are useful for particular applications, such as the aqueous conditions needed for mosquito eggs. Adhesives resistant to moisture include Toupee tape (TopStick™, Vapon Inc.) and Tegaderm (3M). When eggs are injected under oil, the tape strip is placed within a thick rectangle, created with a wax pencil, that can retain the oil. It is important that the wax fence is not breached by oil when overlaying the eggs, since the loss of oil will result in embryo death.
Where possible, eggs are placed on the tape in an orientation with their posterior ends facing outwards towards the needle, but at a slight angle. All fruit fly eggs must be desiccated to some extent before injection. The interior of the egg is normally under positive pressure, and yolk and injected DNA will invariably flow out after injection without desiccation. This will result in lethality, sterility (from loss of pole plasm), or the lack of transformation if the plasmid DNA is lost. The time and type of desiccation, however, must be evaluated empirically, and sometimes varied during the course of an injection period. A major factor for dechorionated eggs is the length of time they are kept on moist filter paper before being placed on the tape. Typically, we desiccate embryos on one strip of tape (15-20 embryos) for 8-10 minutes. Depending on the ease of injection, the time can be varied by 1-2 minutes. In ambient conditions that are humid, it may be necessary to desiccate in a closed chamber with a drying agent (e.g., drierite), with or without a gentle vacuum. An important consideration is that a very short variation in the time for desiccation can be the difference between perfect desiccation and over-desiccation resulting in death, and that the optimal desiccation time will vary for different eggs on the tape. Thus, it is unlikely that all the eggs will respond well to the set conditions, which must be modified so that the majority of eggs can be injected with DNA at a high level of survival and fertility. After the determined time for desiccation, the eggs must be immediately placed under Halocarbon 700 oil, or oil of similar density, to stop the desiccation process. Desiccation of most non-dechorionated eggs is more challenging, and one approach is to soak eggs in 4-M NaCl for several minutes. In contrast, for non-dechorionated mosquito eggs desiccation can occur within 1-2 minutes after removal from water, which is evidenced by slight dimpling of the egg surface, and this must be observed to avoid over-desiccation. Due to the rapidity of desiccation, mosquito eggs are typically arranged on moist filter paper and blotted together onto a taped cover slip from above, and after desiccation the eggs are submerged in Halocarbon oil. Non-dechorionated eggs from many species do not require oil, and it may be lethal for some insects such as moths, yet we find oil submersion helpful for survival of Drosophila and tephritid flies.
Needles
The type of needle and its preparation is possibly the most important component of successful embryo injections. Most dechorionated fruit fly eggs can be injected easily with borosilicate needles that are drawn out to a fine tip and broken off to a 1- to 2-^m opening. Opening the tip is typically achieved by scraping the needle against the edge of the slide carrying the eggs to be injected. Opening the needle by beveling, however, creates consistently sharp tips that are much more important for non-decho-rionated eggs, and stronger alumina-silicate and quartz needles also provide an improvement to easily pierce chorions or tough vitelline membranes. Beveled needles are also critical when a large tip-opening is required for large plasmids that are susceptible to shearing. Preparation of borosilicate needles, pulled from 25-^l capillary stock that has been silanized, can be achieved with several types of vertical or horizontal needle pullers, and we find the Sutter Model P-30 (Sutter Instruments, Novato, CA) vertical micropipette puller to be highly effective. Alumina-silicate needles, and certainly quartz needles, require more sophisticated pullers that allow for fine programmable adjustment of high filament temperatures and pulling force, and the Sutter Models P-97 and P-2000 fulfill this need. Several needle bevelers are available, with the Sutter BV-10 used by many labs.
DNA Preparation and Injection
A mixture of highly purified vector and helper plasmid DNA is essential to embryo survival. This is achieved most optimally by purifying plasmid twice through cesium chloride gradients or a solid-phase anion exchange chromatography column. These have the advantage of high yields of DNA, but the disadvantage of specialized equipment and long preparation times. Successful transformation has been achieved with plasmids prepared with silica-gel membrane kits from Qiagen Corp. (Valencia, CA), but their successful use has been inconsistent, with failures possibly related to the type of host bacteria and its growth conditions. The Qiagen Endotoxin-free plas-mid preparation systems allow additional purity, and we routinely use this system for successful plasmid injection.
Purified plasmid concentration must be accurately titered and verified by gel electrophoresis previous to injection-mix preparation. Appropriate amounts of vector and helper plasmid are ethanol precipitated, washed several times in 70% ethanol, and resuspended in injection buffer. Injection buffer has typically been the same as that originally used for Drosophila (5 mM KCl, 0.1 mM sodium phosphate, pH 6.8), though this may not be optimal for other insects, and embryo survival should be assessed by control injections. Total DNA concentration for injection should not exceed 1 mg/ml, using two- to four-fold higher concentration of vector to helper (e.g., 600 ng/pl vector to 200 ng/pl helper). Higher DNA concentrations are inadvisable, since they are subject to shearing during injection and may clog the needle, and the nucleic acids and/or contaminants can be toxic to the embryo. High transposase levels may also have a negative effect on transposition, as with the overproduction-inhibition phenomenon observed with mariner (Lohe and Hartl, 1996b).
Previous to injection the DNA mixture should be filtered through a 0.45-pm membrane, or centrifuged, before loading into the injection needle. Typically, DNA is back-filled into the injection needle using a drawn-out silanized 100-pl microcapillary, and a microliter of DNA should be sufficient for injecting hundreds of eggs.
DNA injection The microinjection of DNA into embryos requires a system that forces a minute amount of DNA through the needle in a highly controllable fashion. Remarkably, many Drosophila labs simply use a mounted syringe and tubing filled with oil connected to a needle holder, with manual pressure applied. This system is successful due to accumulated expertise and the efficiency of transformation in the species, but would probably be less useful for injecting more sensitive embryos that transform less easily. Regulated air-pressure systems are available that are economical and allow highly controlled and rapid DNA injection. We use the PicoPump from WPI, which is most versatile in allowing positive and negative (with vacuum) pressure, and a hold capability that prevents backflow into the needle resulting in clogging (especially by yolk). A less expensive system can be constructed from Clippard components (Clippard Instrument Laboratory, Inc., Cincinnati, OH) that uses a simple air-pressure regulator and electronic valve and switch (see Handler, 2000). Needle holders from WPI can be used with both systems (MPH-3 and MPH-1, respectively).
All embryo injections are performed on a microscope with a mechanical stage, with the injection needle mounted on a micromanipulator. Microscopes first used for Drosophila transformation were inverted or compound microscopes, but the availability of a useful mechanical stage and stage adaptor for the Olympus SZ stereozoom microscope makes this the most versatile choice (the Olympus stage can be mounted on most stereomicro-scopes). The micromanipulator can be free-standing next to the stage, or mounted on the microscope base. It allows the precise positioning of the needle at the desired point of entry into the egg, while the actual injection occurs by using the mechanical stage to push the egg into the needle. Piezo Translators, which were developed for rapid and automatic intracellular injection, may be more efficient for some embryos, and will obviate the need for a mechanical stage (Peloquin et al., 1997). The WPI MPM20 translator used with the PV820 PicoPump allows a fully automated system for egg penetration, DNA injection, and needle withdrawal.
Post-Injection Treatment
After injection, the cover slip can be placed in a covered petri dish (but not sealed) with moist filter paper. We find use of square dishes with black filter paper to the easiest for up to six cover slips and simple observation of the embryos and hatched larvae. For injected embryos submerged in oil, oxygen concentration may be a limiting factor for development, if not viability. This can be ameliorated by reducing crowding of eggs on the cover slip, or by incubation in a portable hat-box tissue culture chamber that is humidified and under slight positive pressure with oxygen. For eggs without oil, oxygen saturation without pressure is advisable.
Most helper constructs have the transposase gene under heat shock regulation. The Drosophila hsp70 promoter is a constitutive promoter that is active in the absence of heat shock (but also responds to anoxia, which may occur in embryos under oil), and transformation is possible with most vectors with or without heat shock treatment. If heat shock is desirable, it should be noted that the optimal temperature varies for different species. For example, hsp70 responds optimally at 37°C in Drosophila, but at 39°C in medfly (Papadimitriou et al., 1998). Injected embryos should be incubated for at least 4-6 hours after injection before heat shock, or after overnight incubation. Optimal temperatures for insect development vary, but the lowest temperatures possible can be beneficial to survival, and the injection process can slow development by 50% or more. Thus, larval hatching may be delayed considerably, and hatching should be monitored for several days after the expected time before discarding embryos.
Hatched larvae can be placed on normal culture media, though they may be weak and require careful handling and soft diet. Rearing of putative transgenic lines is typically achieved by back-crossing to the parental line in small group matings, or individual mating if a determination of transformation frequency is required. Inbreeding of G0s can minimize rearing efforts, but this may be complicated by a high rate of infertility, which is typically close to 50% after fruit fly injections.
Improvements for Transformation Methodology
Dramatic progress has been made in transformation technology for non-drosophilid insects, and it appears that the vectors and markers in use should be widely applicable. Nevertheless, transformation of many other insect species will be highly challenging, primarily due to limitations in delivery of DNA into pre-blastoderm embryos. As noted, to date all successful non-drosophilid transformations have resulted from embryonic microinjection of DNA, but for many species current injection techniques are likely to result in high levels of lethality or sterility. Experimentation with alternative methods has been reported, though, arguably, none has been tested exhaustively for germ-line transformation, or vector systems were used that are now known to be ineffective. The most promising method is biolistics, where eggs are bombarded with micropellets encapsulated by DNA, which was first developed as a ballistics method to transform plant cells (Klein et al., 1987). Ballistics is based upon a "shotgun" technique for bombardment, and it is the only non-injection method successfully used to transform an insect. This was a P transformation of Drosophila, though only a single transformant line (at an unknown frequency) was created, and the technique never gained wide applicability (Baldarelli and Lengyel, 1990). This was most likely due to the high efficiency of P transformation of Drosophila by microinjection, eliminating the need for an alternative technique.
Mosquito eggs are considerably more difficult to inject, and a significant effort was made to modify a biolistics approach to DNA delivery in An. gambiae, using a burst of pressurized helium for bombardments (Miahle and Miller, 1994). This technique was effective in introducing plasmid DNA into mosquito eggs, yielding high levels of transient expression of a reporter gene. Biolistics was subsequently used for transient expression in specific tissues, allowing the testing of fibroin gene promoters in the B. mori silk gland (Horard et al., 1994; Kravariti et al., 2001). Recent advances have included the use of a rigid macro-carrier in the Bio-Rad PDS/1000-Helium biolistics apparatus that minimizes the blast effect in soft tissue (Thomas et al., 2001). This allows greater micropellet penetration into insect tissues, with improved survival. Despite these advances in delivering DNA into eggs and tissue, biolistics has yet to yield a germ-line transformant. A more recent modification of biolistics using the PDS-1000/He system with the Hepta adapter (BioRad) has, for the first time, yielded repeatable 3-4% transformation frequencies in Drosophila (Yuen et al., 2008). This is highly encouraging, though use of this system in non-drosophilid species has not been reported, and conditions may require modification for other types of embryos.
The only other method reported for DNA delivery is electroporation, which, similar to biolistics, has resulted in high levels of transient expression of plasmid-encoded genes in Drosophila (Kamdar et al., 1992), as well as in Helicoverpa zea and M. domestica (Leopold et al., 1996). Though transformation has not been reported, as with biolistics, it is not apparent that this was seriously tested, or if functional vector systems were used (certainly for non-drosophilids). Electroporation techniques have also advanced in recent years, with DNA transferred into many different tissue types from a variety of organisms using new electroporation chamber designs and electric field paramenters.
These recent advances with both biolistics and electro-poration are highly encouraging that new efforts will have greater chances for success, and they deserve high priority for testing. Both methods also have the advantage, if successful, of the ability to deliver DNA simultaneously to multiple embryos, ranging from hundreds to thousands, depending on the species. This would be highly beneficial to all transformation experiments, but especially so for species that transform at low frequencies. These methods could also be used in cellularized embryos after blastoderm formation in insects having embryos that cannot be easily handled or collected in the preblastoderm stage.
Other approaches to DNA delivery can include the incorporation of vector/helper DNA into bacterial or viral carriers, which may be delivered by maternal injection or feeding. Variations on microinjection that might be required for ovoviviparous insects include maternal injection into ovaries or abdominal hemocoel (Presnail and Hoy, 1994), and use of liposomes might allow injection into cellularized embryos (Felgner et al., 1987). All of these techniques should be re-evaluated with the use of vectors and markers now known to be highly efficient in non-drosophilid systems.
Summary
After concerted efforts for more than 30 years to achieve gene transfer in non-drosophilid insects, only in the past 15 years or so have these efforts been fruitful. Since 1995 the germ-lines of nearly 20 species in 4 orders of insects have been transformed, and this number may be only limited by the insects of current experimental and applied interest. Unlike plant and vertebrate animal systems that allow relatively efficient genomic integration of introduced DNA, insect systems have generally relied on vector-mediated integrations, and the only vectors found reliable for germ-line transformation are those based on transposable elements. Curiously, the two main vector systems developed for routine use in D. melanogaster, and originally discovered in that species, P and hobo, have not been applicable as vectors to any other species. Yet four other transposons found in non-melanogaster or non-drosophilid species are widely functional in insects and, for some, other organisms. Their discovery has been of enormous importance to the wider use of transformation technology, since little progress would have been made if most vector systems were specific to a particular host. Equal in importance to the advancements in vector development has been concurrent progress in genetic marker discovery and development. This began with the finding that cloned eye-color genes from Drosoph-ila could complement existing mutations in other insects, and has continued with the more recent use of several fluorescent protein genes that are widely applicable as markers for transformation and reporters for gene expression.
The advancement of these techniques comes at a fortuitous time when genomics is providing a wealth of genetic infomation and resources that might be used to create transgenic strains of pest and beneficial insects to control their population size and behavior. As part of these efforts, genetic transformation is also critical to functional genomics studies that will provide information essential to understanding the biological function of genetic material, and relating specific genomic elements to those functions. Techniques such as enhancer traps and transposon tagging, which rely on remobilizable insertional mutagenesis, are only possible with transposon-based vector systems, and other techniques such as RNAi are greatly facilitated by these systems. Together, routine methods for trans-poson-mediated germ-line transformation and genomics analysis should provide the tools for dramatic progress in our understanding and control of insect species.
