|Home | About | Journals | Submit | Contact Us | Français|
We describe two replication incompetent retroviral vectors that co-express GFP and beta-galactosidase. These vectors incorporate either the avian reticuloendotheliosis (spleen necrosis virus; SNV) promoter or the chick beta-actin promoter, into the backbone of the murine leukemia (MLV) viral vector. The additional promoters drive transgene expression in avian tissue. The remainder of the vector is MLV-like, allowing high titer viral particle production via transient transfection. The SNV promoter produces high and early expression of introduced genes, enabling detection of the single copy integrated GFP gene in infected cells and their progeny in vivo. Substitution of the LacZ coding DNA with a relevant gene of interest will enable its co-expression with GFP, thus allowing visualization of the effect of specific and stable changes in gene expression throughout development. As the VSV-G pseudotyped viral vector is replication incompetent, changes in gene expression can be controlled temporally, by altering the timing of introduction.
In the study of development, it is often necessary to know the effect of exogenous gene expression on later developmental events. The chick model, although an excellent model for all stages of embryonic development presents challenges when used as a genetic model; the reproductive cycle is too long and the animal is too big. Introducing genes into developing embryos is easily accomplished with electroporation or retroviral-mediated gene transfer (Hughes, 2004; Krull, 2004). For the study of later developmental events however, electroporation does not suffice, as rapid cell division in the chick embryo dilutes out the plasmid as quickly as 24 hours, and usually by 72 hours (Nakamura and Funahashi, 2001; Timmer et al., 2001; Watanabe et al., 2007), although in some tissues, such as the lens, expression of electroporated DNA can be stably expressed (Chen et al., 2004). The lack of reliable long term expression has been overcome with the coupling of transposon-mediated gene transfer and electroporation (Sato et al., 2007) but there still remains variability in the amount of DNA integrated by individual transfected cells. Retroviral-mediated gene transfer then, is the method of choice for developmental questions requiring long-term gene expression. Although replication competent vectors of the Rous Sarcoma Virus (Fekete and Cepko, 1993; Bell and Brickell, 1997; Federspiel and Hughes, 1997) give excellent results and widespread infection, it is difficult to know when exactly during development a particular tissue has been infected. This is a complicating factor in interpretation, as many expressed factors have different effects at different stages: it becomes impossible to know the precise effect of overexpression.
For precise timing of ectopic gene expression, it is preferable to use replication-incompetent retroviral vectors and alter the timing of introduction. In addition, the removal of the essential genes from the viral genome means that there is more space available for the ectopic gene cassette; both a gene of interest and a lineage marker/ gene can be simultaneously expressed. Replication-incompetent retroviral vectors need to be created in packaging cell lines, which supply the structural genes in trans. In a typical protocol, a viral producing cell line is generated by clonally selecting a transfectant of the packaging cell line that has stably integrated the viral genome. A cell line with these two independent integration events will effectively produce infectious viral particles. This method is time consuming, but if a stable clone is identified, it is straightforward to produce high titers of virus for introduction into the embryo. For reviews of virology and additional protocols in vector construction and manipulation, these references are recommended: (Cepko and Pear, 2001a; Cepko and Pear, 2001b; Hughes, 2004; Ishii et al., 2004)
In creating a retroviral vector for expression of exogenous genes, the viral Long Terminal Repeat (LTR) functions as the endogenous promoter, to drive expression of the desired construct. These are typically strong promoters. The RNA produced is both packaged into virus particles and translated by the cell line. If the exogenous genes do not perturb the virus producing cell line, then it is straightforward to identify clones producing high titers of virus. However, vectors that contain toxic genes, particularly dominant-negative elements of growth factor signals or RNAi of essential genes, are difficult to make, as the packaging cells do not tolerate that expression (Itoh et al., 1996; Ishii et al., 2004). In general, it is inconvenient to produce such replication incompetent retroviruses and has not been widely reported.
A replication-incompetent retroviral system based on the Moloney Murine Leukemia Virus (MLV) circumvents these difficulties by creating virus after transient transfection. By using derivatives of the highly transfectable HEK 293 cells as the packaging cells, the pro-viral vectors (the retroviral genome in plasmid form) can be efficiently introduced into a packaging cell line. Virus genome is created from what may be several copies of the proviral vector existing as episomal plasmids. The virus particles are collected from the culture supernatant over the course of three days. This system can produce viral titers on the order of 106 approximately 24 hours after transfection of the packaging cell line. Unfortunately, the MLV promoter is not active in avian tissue (Chen et al., 1999; Mizuarai et al., 2001; Koo et al., 2004), and so the available retroviral vectors, such as pBabe (Morgenstern and Land, 1990; Swift et al., 2001) and the commercially available MLV based vectors (Pantropic packaging system, Clontech, USA) cannot be used. The lack of avian expression has previously been overcome by adding an internal promoter to drive expression of the gene of interest (Mizuarai et al., 2001; Koo et al., 2004), or by using the mammalian packaging cell system to package avian retroviruses (Chen et al., 1999). We sought to improve on these methods by using the MLV based virus and packaging systems and placing an avian specific internal promoter to drive expression of all the desired exogenous genes. By using packaging cell lines, we are able to insert up to 6 kb of exogenous DNA, for a total genome size of 8.8kb (Hu and Pathak, 2000). After accounting for the second promoter and an internal ribosomal entry site (IRES), approximately 4 kb of space is available for specific gene expression. We created two vectors, using two different internal promoters; the chicken beta-actin genomic promoter and the promoter from the avian reticuloendotheliosis virus (spleen necrosis virus/SNV). Each virus carries two reporter genes, GFP and LacZ, which together represent 3.7 kb of exogenous DNA. We find that these constructs are efficiently packaged and pseudotyped using the well established transient transfection into a HEK293 based packaging cell line (Swift et al., 2001). We have tested these new viruses and find that they can be created at sufficiently high titers for effective targeted injection into chick embryos in vivo. As inserted retroviral genomes, they are stably expressed throughout development, and do not show significant tissue preferences. Finally, we show that the viral vectors using the SNV promoter show particularly robust expression, and uniquely allow visualization of GFP following infection and single copy integration of the construct.
Retroviral vectors were created by altering an existing murine leukemia viral vector to include additional promoters that would be expressed in avian tissue. Downstream of the tandem promoters, a construct was inserted to express GFP (green fluorescent protein), an IRES (internal ribosomal entry site) and the LacZ gene. The LacZ gene, at 3kb in size, serves as a demonstration for the upper limit in the size of potential genes of interest and can be replaced as needed. The chick beta-actin promoter (Niwa et al., 1991) was used to create the proviral vector pACID-gfp (Fig. 1A). We isolated the minimal promoter elements from the LTR of the spleen necrosis virus (Dougherty and Temin, 1986; Mikawa et al., 1991), chosen because of its robust expression in a wide range of chick tissue (Mikawa et al., 1991; Mozdziak et al., 2003; Noden and Francis-West, 2006). Because of cloning constraints, the SNV promoter was included along with a portion of its packaging signal, and a portion of the packaging signal region of MLV was removed; the construct was named pSNID-gfp (Fig. 1B). Throughout, they will be referred to as simply pSNID or pACID. Tandem reporters could potentially interfere with each other, resulting in either loss of expression of the exogenous genes or reduction in titer, if the MLV LTR is silenced. Virus particles were pseudotyped with the vesicular stomatitis virus. We used pSNID and pACID to create infectious virus particles, and tested the titers on mouse fibroblasts (NIH-3T3) and dog fibroblasts (D17). For each vector the mouse fibroblasts indicated an effective titer approximately 10 fold less than the D17 fibroblasts (data not shown). This seemed surprising, since the MLV LTR is a strong promoter in murine cells. From these in vitro experiments we concluded that the chicken promoters were most likely interfering in mouse cells. In several subsequent experiments, average viral titer produced were 1.3×104 virions/ml for pACID and 8.7×105 for pSNID, in keeping with various published reports for viral titers produced from a simple MLV vector or a replication-defective RCAS-based vector (Table 1). Concentrating the viral supernatant through ultracentrifugation-produced titers of 3.6×106 and 6.3×107 for pACID and pSNID respectively, a 100-fold increase in titer, similar to the reported two-fold increase typically found with ultracentrifugation (Burns et al., 1993). We did not see a significant reduction in the titer of the virus produced, compared to published accounts, due to having tandem promoter elements (Morgenstern and Land, 1990; Chen et al., 1999; Swift et al., 2001).
Concentrated retroviral stocks were used to microinject chick embryos at Hamburger Hamilton (HH) stages 8–12 with approximately 1–5 nanoliters of virus. Injection volumes were calculated using a micrometer to define a 1 nanoliter drop (of a diameter of 130 microns and slightly smaller in size compared to the axial length of a the most recently formed somite, at 150 microns). Thus at an average virus concentration of 1×106 virions/ml, a one nanoliter drop should contain 2 virions and at a concentration of 1×107 the same size drop would contain 20 virions. pSNID and pACID virus at average concentrated titers of 3.1×105 virions/ml and 1.4×106 virions/ml respectively, were introduced using a pressure injector to consistently deliver drops of no more than 5nl. Injections were precisely targeted to the presomitic mesoderm, the cranial neural crest, head mesoderm, the neural tube, the somites, forelimb level lateral plate mesoderm, the telencephalon and the optic vesicle. The embryos were re-incubated, harvested at various timepoints, and the resulting infection visualized with X-gal staining. Given the titers and injection volumes, and assuming an 100 percent infectivity, the maximal number of infected clones per injection is approximately 1 for pACID and 7 for pSNID; individual titers leading to displayed embryos are provided in the figure legends. As expected, the pSNID virus was expressed in all targeted tissues (Fig. 2). Likewise, pACID was widely expressed, but the intensity of X-gal reaction product was not as robust, making it difficult to localize foci of infection in whole-mount embryos prior to clearing (Fig. 2A–2D). As replication incompetent retroviral vectors cannot produce infectious particles without the required packaging cell line components, including the envelope gene, viral driven expression is confined to the original infected cell and its progeny. In light of the chimeric nature of this retrovirus, helper virus production is unlikely, and in fact we did not detect any (Cepko and Pear, 2001c). Therefore, the typical appearance of infected tissue is of small foci of transgenic cells within a larger field of wild-type cells.
By altering the titer and volume of virus introduction, the majority of a given tissue can be infected if required. We used higher titer viral stocks, and larger volume injections in order to create embryos with highly infected tissues (Fig. 2I–L). Titers for these injections were on average 7.2×107 virions/ml and approximately 100nl were injected (Fig. 2I–L). This would introduce approximately 640 infectious particles per injection. The in vivo efficiency of infection compares favorably with an RCAS-based replication-incompetent retroviral vector, when adjusted for the differences in injection volumes used (100–500nl versus the 1–5nl and the 100nl used here; Chen et al., 1999).
We infected chick embryos as described, and examined them 48, and 24 hours after infection. We were able to visualize pSNID infected tissues through direct detection of GFP epiflourescence. Infections were found in all targeted tissues, including the brain, eye, and head mesoderm (Fig. 3A). Injection into the presomitic mesoderm resulted in GFP expression in the dermomyotome and myotome of formed somites (Fig. 3C) and targeting the developing heart similarly resulted in labeled heart cells (Fig. 3E). The GFP expression pattern was mirrored by the resulting X-gal expression pattern (Figs. 3B, 3D, and 3F). This indicated that the expression in vivo from the engineered SNV promoter was strong enough to accumulate GFP at a detectable level after 48 hours.
We also detected pSNID-driven GFP expression less than 24 hours post-infection, with varying amounts of overlapping detectable X-gal stain (Fig. 4A–4D). It was possible to detect overlapped LacZ expression at 18 hours, but not at 12 or 8 hours, implying that the cap-dependent translation (of GFP) may be more efficient than cap-independent translation from the IRES (LacZ) in very young chicken embryos (Hennecke et al., 2001). However, at all timepoints earlier than 48 hours post infection, it was not possible to detect either GFP or LacZ expression from the pACID infections. We had previously noted that although pACID infections after 2 days post infection were detectable with X-gal staining, the staining was less intense than that seen for pSNID (Fig. 2). The beta-actin promoter is widely used for overexpression of constructs in electroporation protocols; we hypothesized that as a single integrated copy, the promoter was not strong enough to produce detectable expression in the rapidly dividing cells of the early embryo. This was tested by electroporating both pACID and pSNID into stage 10 embryos. Strong expression of GFP using this technique was found with both pACID and pSNID transfection, supporting our hypothesis (Figs. 4E and 4F) and lending support to our interpretation of the relative strength of beta-actin versus SNV promoter. When pACID infections were then re-evaluated using antibodies against GFP and beta-galactosidase to amplify the signal, both markers were indeed detectable (Fig. 4G–4J).
To confirm that expression of the integrated constructs driven by the SNV and beta-actin promoters are maintained over several days of development, embryos were re-incubated for longer periods following infection. Expression of both vector constructs were visualized in targeted tissues at 5–6 days post-infection (Fig. 5). As at earlier time-points, no GFP was detectable from pACID (data not shown) but GFP was directly detectable following pSNID infection (Fig. 5F). Targeted infection (low virion introduction) allows accurate evaluation of the level of marker expression. Under these conditions we did not see any significant downregulation of expression in targeted tissues at 2 days post infection versus 5 days (compare Fig. 3F to Fig. 5C). Here we have looked solely at marker gene expression driven by the beta-actin and SNV promoter constructs. The strength of GFP expression achieved with the SNV promoter will be a useful tool for studying the fate of tissues where bioactive factors have been mis/over-expressed. Infected tissues can be identified early in vivo and followed through development, improving end-point misexpression analyses.
We have designed new retroviral vectors for the rapid production of high titer replication incompetent retroviral vectors. These vectors were created to enhance and refine the ability to genetically manipulate avian tissues in vivo. To date, the most efficient methods to achieve exogenous gene expression have been the use of replication competent retroviral vectors and in ovo electroporation. However, various questions of development still cannot be answered using these techniques. Replication competent retroviruses produce a widely infected tissue, but the timing of infection, and thus gene expression, is variable throughout the tissue. The resulting analysis of the effects of exogenous gene expression therefore, becomes difficult. Electroporation, in comparison, introduces exogenous genes to the primary cohort of electroporated cells and their progeny. The exogenous genes remain as episomal plasmids in the cytoplasm of the cells, and in the chick embryo, expression is lost through dilution, as development proceeds. Thus, it is difficult to target a particular subpopulation of cells for genetic change and then examine how that specific change affects organogenesis.
Variations of the above techniques can produce discreet and permanent genetic changes in a cohort of cells in vivo. Recently, transposon driven integration systems were combined with electroporation techniques, to describe a method where electroporated constructs could be stably integrated into the genome of chick cells (Sato et al., 2007). This represents an advantage in speed and efficiency of making transgenic tissues in the avian. Pseudo-typed lentiviral vectors are also experimentally appropriate for use in the avian and are both replication incompetent and rapidly produced through the transient transfection of mammalian packaging cells. Most reports detail their use in making germline transgenics, as lentiviral promoters are not usually silenced by the host during development (Lois et al., 2002; Pfeifer et al., 2002). Depending on the promoter used to drive the in vivo expression, GFP could be directly detected in avian embryos (McGrew et al., 2004; Chapman et al., 2005; Scott and Lois, 2005). As retroviral vectors are fully sufficient for introducing genes into developing embryos, we chose to continue and improve a retroviral vector approach, for its ease of use, the widespread availability of reagents and protocols, and the high survivability of injected versus electroporated embryos.
Our vector differs from other reported constructs in the novel use of the promoter from the Spleen Necrosis Virus (SNV). SNV based retroviral vectors are characterized by particularly robust expression in avian tissue, making them excellent tools for lineage analysis (Mikawa et al., 1991; Wei and Mikawa, 2000; Evans and Noden, 2006). The SNV promoter is active at very early stages of embryo development; expression from the inserted virus has been found at HH stage 4, after injection at stage HH Stage 3 (Wei and Mikawa, 2000). Germline transgenic birds have also been created using the SNV viral vectors, and expression of the transgene, in this case beta-galactosidase, is expressed early and ubiquitously throughout the animal (Mozdziak et al., 2003; Mozdziak et al., 2006). This is in comparison to expression of transgenes driven by other viral promoters, such as the RSV, CMV and MLV promoters, which do not express uniformly throughout tissues, even in a germline transgenic animal (Challita and Kohn, 1994; Mizuarai et al., 2001; Kwon et al., 2004; McGrew et al., 2004). To date, ubiquitous strong expression of transgenes in the avian embryo has only been seen with eukaryotic promoters, such as the phosphoglycerol promoter (Chapman et al., 2005). The SNV promoter may not be as subject to developmental silencing, proposed to affect other retroviral family members (Scott and Lois, 2005). SNV driven expression is also stable throughout development, with transgene expression detectable within 24 hours after infection and over the course of 5 additional days.
Part of the usefulness of the pSNID vector described here lies in its ability to express GFP at directly detectable levels. In several other engineered retroviral vectors, it is not possible to visualize the GFP marker with epiflourescence after infection, because the construct is expressed from a single inserted copy of the gene (Jackson et al., 2006; Yamashita et al., 2006). For example, RSV-driven GFP expression in early embryonic chicken tissues could not be seen without immunohistochemistry (Zhang et al., 2004; Chau et al., 2006); GFP fluorescence was detectable by embryonic day 8 in some cases (Koo et al., 2004). Likewise, the chick beta-actin promoter, as used in the pACID vector, does not provide sufficiently robust expression of downstream genes: Although we can detect beta-galactosidase, we cannot detect GFP without immunohistochemistry. It is possible to enhance the GFP protein, or stabilize the message, and this can result in sufficient expression. Koo et al. used a WPRE stabilizing element at the end of the GFP message and this engineering allowed enough protein to accumulate such that it was possible to see directly (Koo et al., 2006). A second group detailed mutants of the GFP protein designed to stabilize the protein structure at 37°C and could thus directly observe GFP expression driven by beta-actin (Okada et al., 1999). The pSNID vector also allows real time visualization of GFP labeled cells. However, whereas the previous reports altered the marker gene, we have altered the promoter. Therefore, in addition to expressing this marker gene at significant levels, we can also express a gene of interest at comparable levels, by placing it downstream of the IRES. This will make it possible to study cell behavior in real time after genetic manipulation in vivo. The nature of the packaging protocol, since it is based on transient transfection of packaging cells, allows for any ectopic gene to be easily transferred, including genes that would be disadvantageous or toxic for the propagation of a traditional packaging cell line. Size limitations for ectopic genes are on the order of 4 kb and inclusion of a large gene, LacZ as an example, do not lead to reduced virus titers. Finally, since the infection is a permanent change to the infected cell and its progeny, it is possible to follow the results of ectopic gene expression into the latest of developmental timepoints.
In summary, we have presented a set of replication-incompetent retroviral vectors that provide certain improvements over existing replication-incompetent retrovirus systems. By incorporating a novel isolated SNV retroviral promoter in the pSNID series we have created a vector that: 1) expresses GFP at levels suitable for in vivo imaging of cell behavior; 2) expresses within 24 hours after infection; 3) does not show tissue restriction and 4) is not affected by gene silencing. Since the vectors are basically mouse viruses, engineered to express in avian tissue, there is very little possibility of helper virus being produced, making this vector fitting for lineage studies. This also means that pathogen-free eggs are not required for experiments; all infections in this report used fertile eggs from a local producer. Finally, the production of high titer retrovirus can be accomplished without cloning and selection of packaging cells, and is thus appropriate for the mis-expression and in vivo analysis of the effects of a large variety of exogenous genes.
To facilitate cloning, the pBMN-I-GFP construct (Dr. G. Nolan, Stanford University, Palo Alto, CA; plasmid obtainable from Addgene.org, catalog #173) was split into two shuttle vectors: 1) A “Promoter Shuttle” was created by isolating a SexA1-Not1 fragment containing a portion of the MLV packaging signal region and the multiple cloning site, 2) A “IRES-Gene” shuttle was created by isolating a Not-Sal fragment containing the IRES and the GFP gene [pIRES-gfp].
The IRES-gene shuttle vector was used to replace the GFP protein with the beta-galactosidase gene, by swapping in a Nco1-Sal1 flanked LacZ sequence from pCXIZ [pIRES-LacZ] (Itoh et al., 1996).
To generate pSNID: First, a 1kb EcoR1 fragment, containing the 5’ LTR and the packaging signal, was isolated from the CXL proviral vector (Mikawa et al., 1991) and temporarily placed into pBluescript (Stratagene). The Promoter Shuttle was opened using Pml1 and Sbf1 and an EcoRV-Pst fragment from pSNV was ligated in [pPromoter-SNV]. The eGFP gene (Clontech) was introduced using Xho1-Not1 sites. To assemble the pSNID vector, the 3 kb Bgl-Not fragment from pPromoter-SNV:gfp, the 5 kb Not-Sal fragment from the pIRES-LacZ and the 3 kb Bgl2-Sal fragment from pBMN-I-GFP were ligated together.
To generate pACID: A 1.5 kb SnaB1-Pst1 fragment containing the chick beta-actin promoter was isolated from pMES (Swartz et al., 2001) and placed into the Promoter Shuttle using the Pml1 and Sbf1 sites [pPromoter-ACT]. The eGFP gene was introduced using Xho-Not1 sites. The proviral vector was assembled similarly as described for pSNID: the individual components were isolated and ligated back together using Bgl2, Not1, and Sal1 sites. All constructs are available upon request.
Reagents: Phoenix-GP packaging cells were created by Dr. G. Nolan, Stanford University, obtainable from the American Type Culture Collection Safe Deposit (catalog number 3514). They can be ordered by request: gro.ccta@pedefas. Further descriptions of this and other packaging cells can be found at Dr. Nolan’s website: http://www.stanford.edu/group/nolan/retroviral_systems/retsys.html.
The Phoenix packaging cell line contains the gag and pol structural genes required for retroviral propagation. Cells were maintained in DMEM media with 4.5g/L glucose, 0.58g/L L-Glutamine, and 10% fetal bovine serum. For best results with viral production, the serum used should not be heat inactivated. Packaging cells were periodically placed under hygromycin selection every three months for two weeks. The envelope was added on a separate plasmid using pCI-VSV-G (Addgene.org, catalog number 1733). The VSV-G envelope protein allows infection of a broad range of organisms and strengthens the virus structure for the purpose of ultracentrifugation and frozen storage (Burns et al., 1993).
Phoenix cells were plated on 10cm plates at 70–80% confluency in 7.5ml DMEM with 10% Fetal Calf Serum and 1% Penicillin/Streptomycin several hours before transfection. For each 10cm plate, 2.5ml calcium phosphate reaction mix was prepared: 156.3µl 2M CaPO4, 15µl VSV-G and 30µl proviral DNA each at 1mg/ml, and sterile water to 1.25ml. The DNA solution was mixed dropwise into 1.25ml 2×HBSS, pH7.0 (34.2mM NaCl, 9.9mM KCl, 1.5mM Na2HPO4.7H20, 5.5mM glucose, 19.2mM HEPES). A fine precipitate should form upon mixing. The reaction mix and 25µM chloroquine was added gently and cells incubated. Following overnight incubation, the cells were gently washed with PBS and 6.5ml DMEM containing 10% non heat-inactivated fetal calf serum and 1% penicillin/streptomycin added. We observed that heat-inactivation affects the serum proteins, such that they are pulled down with the viral pellet and cause variability in the titer. The technical support at the supplier (Gemini BioProducts) confirmed our observations and guidelines from many suppliers indicate that new filtering technologies have supplanted heat-inactivation. Virus supernatant was collected and media replaced after 24 hours for a second collection.
For embryo injections, the supernatants from 6 plates were collected into a single centrifuge tube and spun at 18K rpm for 2 hours at 4°C in an ultracentrifuge (Beckman; SW28 rotor). Supernatant was drained immediately upon stopping and pellet allowed to re-suspend in the remaining media (approx 50µl). Polybrene was added to the resulting virus concentrate to a final concentration of 100 µg/ml. All virus preparations were titered on dog fibroblasts (D17-catalog number CCL183, ATTC, ATTC.org).
Helper virus: To confirm that no replication-competent helper virus was present in the viral stocks used, concentrated viral supernatants were used to infect D17 fibroblasts plated at 80% confluency in media containing 10µg/ml Polybrene. Following 24 hours incubation, media was collected, spun at 5000rpm for 5 mins and serially tested by applying to fresh D17 fibroblasts. The serial reinfection was repeated 5 times for 6 individual virus preparations for each vector. Cells were processed for beta-galactosidase expression and the infection titer calculated. No infections were detected after the second passage of virus supernatant containing media (data not shown).
Fertilized chick eggs were purchased from a local breeder (Petaluma, CA) and incubated to between stages 9 and 11 (Hamburger and Hamilton, 1992). For injection, a small window was opened on the widest end of the egg, and the intervening membranes removed. Contrast ink (India Ink, non-waterproof, 10% in Tyrodes saline) was added to the yolk under the embryo with a 30-gauge needle. Approximately 1–5nl of viral solution was pressure injected Injection volumes were calculated using a micrometer to define a 1 nanoliter drop, which has a diameter of 130 microns. This is facilitated using a digital camera and monitor, where the micrometer and embryo can be photographed at the same magnification and directly measured. As an internal reference, a 1nl drop is slightly smaller in size compared to the axial length of the most recently formed somite at most stages of early development, which is 150 microns from anterior to posterior boundary. Given an virus concentration of 1×106 virions/ml, a one nanoliter drop should contain 1 virion and at a concentration of 1×107 the same size drop would contain 10 virions. Needles were prepared from glass capillaries (external diameter 1 mm) using a Kopf needle puller. We use a pressure injector for viral introduction, which provides uniformity across embryos, (Harvard Apparatus model PLI-100 pressure injector). Virus is injected into various targeted sites in the embryo, the embryos resealed with Parafilm (Pechiney Plastics, Menasha WI), and re-incubated for various times between 8 hours and 6 days.
Embryos were dissected free from extra-embryonic membranes and fixed in PBS containing 2% paraformaldehyde. LacZ expression was detected in wholemount embryos by incubation in X-gal staining solution: 35mM potassium ferrocyanide, 35mM potassium ferricyanide, 2mM MgCl2, 1 mg/ml X-gal (5-bromo-4-chloro-3-indolyl- beta-D-galactoside, Research Products International Corp.) in PBS containing 2mM MgCl2 overnight at 37°C. To clear, following X-gal staining and washing in PBS, embryos were incubated in increasing concentrations of glycerol (20%, 50%, 80%, 100%) in 1% KOH for 2–3 days at 37°C for each concentration (Schatz et al., 2005). Following fixation, embryos for whole mount immunohistochemistry were washed in PBS and incubated in immunodiluent (PBS containing 10% normal goat serum, 0.1% Triton X-100, and 1% BSA) containing primary antibodies against beta-galactosidase (Developmental Studies Hybridoma Bank, 40-1a) at 1:200, and GFP (Chemicon) at 1:500 overnight at 4°C. Antibody labeling was visualized with goat anti-mouse IgG-alexa546 and goat and anti-rabbit alexa488 secondary antibodies (Molecular Probes) at a concentration of 1:500 incubated in immunodiluent overnight at 4°C.
Embryos were visualized as described above. Proviral vector plasmids were pressure injected using pulled borosilicate glass capillaries (1mm diameter, World Precision Instruments, Sarasota, FL) into the target tissues. 0.5mm diameter platinum electrodes were made in house and insulated with nail polish to leave only the tip exposed. Electrodes were held in a commercially available holder (Computech Manufacturing Co., Kansas City, MO). For electroporation, electrodes were placed either side of the appropriate embryo level. Current was applied as 4 square wave 40 ms pulses of 25 volts at 70ms intervals using an Intracell TSS20 electroporator. Eggs were sealed with Parafilm and re-incubated for 24 hours.
LacZ or GFP expression was visualized and imaged on a LeicaMZ16F fluorescence stereomicroscope fitted with a Leica DFC500 camera. Images were collected using Leica Firecam software version 3.2 and collated in Adobe Photoshop v10.0.1.
Supported by grants from the National Institutes of Health (through grants EY015429 and EY002162), and from That Man May See (grant 06720). We thank Arturo Alvarez-Buylla and Charles Ordahl for sharing equipment and reagents and the Takashi Mikawa lab for helpful comments and discussions. Thank you to Camila Alvarez-Buylla for help with sections.
National Institutes of Health: EY01529