Transgenic zebrafish produced by retroviral infection of in vitro-cultured sperm

Abstract

Transgenic modification of sperm before fertilization has distinct advantages over conventional transgenic methods. The primary advantage is that the mosaicism inherent in those other techniques is avoided. A culture system using primary cultures of zebrafish male germ cells, in which the differentiation from spermatogonia to functional sperm can occur in vitro, provides the opportunity for genetic modification of sperm in vitro. Here, we report the production of transgenic zebrafish from cultured sperm. The sperm were differentiated from premeiotic germ cells infected with a pseudotyped retrovirus in vitro. The collected sperm were used to perform successful in vitro fertilizations, and transgenic embryos were identified. The transgenic fish transmitted the proviral integration to the next generation in a Mendelian fashion. We report the generation of a transgenic animal by cultured sperm and open the door to many exciting possibilities for the rapid generation of transgenic lines in model organisms such as zebrafish or other animal systems that are otherwise intractable to transgenesis.

Mosaicism is inherent in conventional transgenic techniques, such as microinjection of DNA into eggs (1), chimeras of transfected embryonic stem cells (2–4), viral infection of embryos (5–7), and sperm-mediated gene transfer into oocytes (8–11). The injection of DNA into the male pronucleus of fertilized eggs and the sperm-mediated gene transfer reduce mosaicism but do not exclude it (1, 10, 11). By using any of these techniques, heterozygous transgenic animals are obtained in the following generation. Introduction of a foreign gene or alteration of a target gene in the sperm genome before fertilization leads to specific genetic changes in all cells in the initial generation, including the germ cells of the organism created from that sperm. However, sperm chromatin is highly condensed, making genetic manipulation difficult, thus foreign DNA must be introduced into male germ cells before their differentiation into sperm.

Several reports (5, 12–19) in the past two decades concerning in vivo infection of male germ cells by viruses and in vivo transfection by nonviral techniques have suggested that delivery of genes to postnatal male germ cells or fertilized embryos is difficult. For example, when lentivirus was injected into seminiferous tubules of the mouse testis, Sertoli cells were infected but no transgenic mice were obtained from 390 pups (19). With in vivo transfection by nonviral techniques, the only success reported was by combining intracytoplasmic injection of sperm previously selected by fluorescence (20). Recently, viral transduction of spermatogonial stem cells that differentiated into functional sperm after transplantation into the testes (21, 22) was reported to generate transfected sperm and transgenic animals. However, the male germ-line stem cells are needed to be transplanted into the testes and differentiated into functional sperm in vivo. Therefore, genetic modification by these methods is limited before transplantation and requires a long time for the differentiation of the germ-line stem cells into sperm. In another approach, a spermatogonial cell line derived from telomerase-transformed spermatogonia was established, which differentiates into spermatids in response to stem cell factor (or Steel factor) (23). Although introduction of a foreign gene into these cells has been shown, it has not been demonstrated that the spermatid can make a normal mouse and that the foreign gene can be transmitted into their germ line. More recently, normal and fertile mice were produced by nuclear injection with cultured round spermatids from mouse primary spermatocytes, but introduction of foreign genes into the germ cells has not been reported (24).

A culture system using primary cultures of zebrafish male germ cells (25), in which the differentiation from spermatogonia to functional sperm can occur in vitro, provides an opportunity to produce sperm-based vectors as an alternate approach to germ cell gene modification using spermatogonial stem cells. This study is designed to determine whether it is possible to infect the male germ cells with retroviruses before differentiation. Here, we demonstrate the successful integration of the retrovirus in cultured sperm and the successful rapid generation of transgenic zebrafish lines through a simple in vitro fertilization using these infected sperm cultures. The transgenic fish generated by the in vitro fertilization are not mosaic and transmit the provirus to their offspring in a Mendelian fashion.

Materials and Methods

Zebrafish Strain. The WT Tubingen stain and the albino type (alb1/alb1) were used throughout the experiments. Use of these animals for experimental purposes was done according to the guidelines of Fukui Prefectural University.

Preparation of Retrovirus. Packaging cell lines (26, 27) were transfected with the plasmid pHCMV-G, which encodes the envelope protein of the vesicular stomatitis virus, by lipofection as described (27). Medium was changed at 24 h and collected at 48 or 72 h after transfection. The media were filter-sterilized (0.2 μm filter) and concentrated by centrifugation at 21,000 rpm with a SW28 rotor for 1.5 h at 4°C. Viral pellets were resuspended in 30 μl of PBS, titered, and frozen at –80°C for future use.

Cell Culture. Normal adult testes were dissociated with 500 units/ml of collagenase (ICN Biomedicals) in L-15 at 28°C for 2 h by pipetting every 20 min. The cell suspension was diluted seven times with L-15 containing BSA and Hepes (pH 7.9) at the final concentrations of 1% and 20 mM, respectively. After removal of undissociated fragments, the suspension was centrifuged at 35 × g for 10 min. The centrifugation pellet, which contained germ cells such as spermatogonia, spermatocytes and spermatids, and testicular somatic cells but few sperm, was resuspended in testicular cell culture medium (TCCM) supplemented with 3% FBS. Feeder cells for the male germ cell culture were prepared from ZtA6 cell-derived ZtA6–12 cell line (28), which exhibited features characteristic of Sertoli cells, and treated with mitomycin C (10 μg/ml in L-15, Sigma) for 3 h as described (25). Typically, dissociated testicular cells of four males were plated on the feeder cells of a 35-mm culture dish. The next day, half of the medium was replaced with fresh TCCM supplemented with 7% crucian carp (Carassius carassius) serum and 50 ng/ml 11-ketotestosterone (this steroid induces spermatogenesis in fish) (29). Then, the cells were incubated with 9 × 10⁷ colony-forming units (cfu) of pseudotyped retrovirus (the final concentration, 7 × 10⁷ cfu/ml) and 10 μg/ml poly-l-lysine. A second dose of retrovirus was added 24 h later. The medium was changed on days 4 and 8. After 12 days, cells were collected and used for fertilization as described (25). Alternatively, on day 8, the cells in the medium were transferred to culture dishes without feeder cells and incubated for an additional 4 days.

Detection of β-Galactosidase Activity. Cultured germ cells were fixed with 0.5% glutaraldehyde in PBS for 10 min. After three washes with PBS + 1 mM MgCl₂, the cells were subjected to staining with 5-bromo-4-chloro-3-indolyl-β-d-galactoside as described (30).

PCR Analysis. Genomic DNA was prepared from caudal fin clips by lysis at 55°C in 10 mM Tris buffer (pH 8.0), 10 mM EDTA, 100 mM NaCl, 0.4% SDS, and 200 μg/ml proteinase K, followed by ethanol precipitation. Two pairs of primers were used for PCR analysis of zebrafish genomic DNA, one specific to the proviral DNA that generates a 310-bp PCR product and a second specific to the zebrafish Wnt5a gene that generates a 387-bp PCR product. PCR primers and conditions were as described (6, 26).

Southern Hybridization. Genomic DNA samples were digested with BglII. Southern blot hybridization was performed with a digoxigenin-labeled (Roche Diagnostics) proviral sequence probe, pSFG-nlacZ DNA, at 65°C by standard methods as described (26).

Results

Retroviral Infection of in Vitro-Cultured Sperm. The first step in the generation of transgenic lines by means of in vitro-cultured sperm is to efficiently introduce a foreign gene into male germ cells before their differentiation into mature sperm. We used a pseudotyped retroviral vector containing a genome based on the Moloney murine leukemia virus (MoMLV) and packaged by the envelope glycoprotein (G protein) of the vesicular stomatitis virus (26), which allows the MoMLV core particles to infect zebrafish cells (31). Male germ cells were attached and cultured on a cell growth-arrested ZtA 6–12 feeder cell line (28) that was isolated from cells derived from a spontaneous tumor-like hypertrophied testis. Pseudotyped retrovirus, concentrated by ultracentrifugation, was added to the culture medium with 10 μg/ml poly-l-lysine after 1 day of culture. A similar dose of concentrated viral suspension was added again after the second day in culture, totaling 2 days of exposure to the active viruses. Using the NK-type retrovirus (26) with lacZ driven from a constitutively active promoter (32) for the infection, we were able to identify transgenic, mature, flagellated sperm, by the detection of β-galactosidase activity after 12 days in culture (Fig. 1). This process demonstrated that the proviral integration was occurring before differentiation.

Fig. 1.

Retroviral infection of sperm cultured in vitro. (A) Detection of lacZ expression in infected sperm after 12 days of culture and (B) normal control sperm subjected to 5-bromo-4-chloro-3-indolyl-β-d-galactoside staining. (Scale bar: 10 μm.)

Generation of Transgenic Lines. Despite several infections using the NK-type retrovirus, we were not able to obtain transgenic embryos, we believe because of the relatively low titer (≈1 × 10⁸ cfu/ml on mouse 3T3 after concentration) of this virus. To demonstrate transmission of proviral integrations in the sperm to generate transgenic lines, we used a packaging line known to produce virus at much higher titer (≈2 × 10⁹ cfu/ml on mouse 3T3 after concentration) (27). This virus did not contain an active reporter gene. After 2 days of infection with the new virus, the male germ cells were cultured on the ZtA6–12 feeder cells for a total of 12 days, and then used to fertilize extruded zebrafish eggs with a simple in vitro fertilization technique (25, 33).

By using the feeder cells, male germ cells can be cultured for ≈16 days. Longer cultures resulted in lower efficiency of fertilization; only a few embryos were obtained from sperm cultured for 16 days. We got significantly better fertilization rates when we used sperm cultured for 8–12 days on the feeders. It takes 8–9 days for sperm to differentiate from actively dividing cells in this culture system (25), thus culture medium changed on day 8 should contain spermatids from cells that differentiated after cell division in culture. We collected the sperm and spermatids on day 8 as cells detached from the feeder layer when we were exchanging the media. We waited an additional 4 days in the absence of feeders to ensure we were harvesting sperm from cells that differentiated after cell division, then the sperm were used to inseminate unfertilized eggs in vitro (25, 33).

From four fertilization attempts using the 12-day cultured sperm, a total of 104 embryos from 1,111 eggs was obtained (Table 1). From these 104 embryos, 89 fish developed to adulthood, and PCR experiments using specific primers for the provirus showed that 5 fish carried a proviral insertion (Fig. 2A). The efficiency of transgenic production is ≈5.6% (5 transgenics per 89 fish). Another attempt at the experiment with the sperm cultured on feeders for 8 days, then incubated an additional 4 days in the absence of feeders, yielded 142 embryos from 299 eggs (Table 1). From those embryos, 132 fish developed to adulthood, and one fish had an insertion (Fig. 2 A); the efficiency for this experiment was 0.8% (1/132). It is unclear whether the removal from the feeder layer reduced the efficiency or whether the general infection rates were simply lower in this experiment. We cultured dissociated testicular cells from WT zebrafish and used the in vitro-cultured sperm to fertilize extruded eggs from albino fish. All fertilized embryos were completely WT for stripe patterns (Fig. 2B).

Table 1. Transgenic zebrafish from cultured sperm with retroviral infection.

Experiment	No. of eggs exposed to culture	Fertilized eggs	Adult fish	Transgenic	F1 transmission rate
1*	258	35	30	4‡	N/D
2*	214	28	27	0
3*	421	15	9	1	13/30 (43%)
4*	218	26	23	0
5†	299	142	132	1	35/60 (58%)

N/D, not determined.

^*Cultured sperm with feeder cells for 12 days were used for in vitro fertilization.

^†Cultured sperm with feeder cells for 8 days plus additional 4 days of incubation were used.

^‡Fish died before transmission could be determined.

Fig. 2.

Detection of transgenesis in fish created from infected in vitro sperm cultures. (A) PCR analysis using DNA extracted from individual founder fish. Two pairs of primers were used for PCR analysis of genomic DNA, one specific to the proviral DNA and a second specific to the zebrafish Wnt5a. The numbered lanes show the PCR products generated with DNA from individuals in a part of each experiment of Table 1 to screen the positive. Lane PC shows the positive control PCR products generated with DNA from a zebrafish cell line that contains a proviral insertion. (B) Photos of transgenic zebrafish from infected in vitro sperm cultures. The WT cultured sperm were used to fertilize extruded eggs of the albino type.

Based on the nature of retroviral infection, the six fish with proviral insertions should be heterozygous transgenics. To confirm this idea, we used Southern blot analysis to demonstrate the proviral integration is in single copy (Fig. 3) and that the integration is transmitted in a Mendelian fashion (Fig. 4). LacZ expression in transgenic embryos could not be examined as the lacZ gene had been inactivated in this construct. When the transgenic zebrafish were mated with albino fish, half of their progeny were albino, again suggesting normal ploidy.

Fig. 3.

Southern blot analysis of genomic DNA from individual PCR-positive founder fish. Genomic DNA samples were digested with BglII. The number of each lane corresponds to the founder's number in the PCR experiment. The hybridization pattern in each lane shows a single copy insertion and different genomic locus for each lane, indicating independent integrations.

Fig. 4.

Analysis of transmission of the retroviral insertion into progenies by PCR and Southern blot hybridization. (A) PCR analysis with DNA extracted from progeny of each positive founder fish in experiments 3 and 5. Almost half of the progeny are positive for the provirus, indicating the transgene is not mosaic in the founder fish. (B) Southern blot analysis of genomic DNA from a positive founder fish and four individual PCR-positive progeny. Genomic DNA samples were digested with BglII. All PCR-positive progenies show the identical integration site to the parent.

Discussion

The results of the present study show that transgenic zebrafish can be produced by cultured sperm modified genetically in vitro with retroviruses. Because we know that the retrovirus decreases in titer ≈10-fold in an 8-h period (31), and we cultured the sperm for an additional 9 days after infection, we can exclude the hypothesis that the sperm were acting as a carrier for unintegrated viruses.

Despite the use of high titer retrovirus (1.8 × 10⁸ cfu in the culture dish), the total efficiency of production of transgenic fish from 12-day-cultured sperm was ≈5% (5 transgenic fish per 89 total analyzed fish). This efficiency compares favorably to the traditional method used for generating transgenics in zebrafish (34). Generally 5–20% of the fish injected at the one-cell stage will transmit the transgene to the next generation. As the injected fish are mosaic for the DNA integrations, typically only 1–20% of the F₁ generation is transgenic. Thus it requires hundreds of fish and two generations to create a desired transgenic line. Whereas infection of male germ cells creates transgenics at the low end of the rate by traditional DNA injection, the transgenic fish are not mosaic, so screening for transgenics and fish husbandry is reduced and an entire generation is skipped. Furthermore, we observed that the efficiency between each experiment was not consistent, from 0% in experiments 2 and 4 to >10% in experiments 1 and 3, suggesting there is significant room for improvement in the present protocol.

The efficiency for this pseudotyped retrovirus to infect the zebrafish embryonic cell line PAC2 has been reported to be ≈2-fold less than that in mouse 3T3 cells (26). Estimating the number of dissociated testicular cells plus feeder cells at ≈ 1 × 10⁷ cells, we added the retrovirus to a multiplicity of infection (moi) of nearly 10, or nearly 10 infectious virions per cell. The rate of growth for the spermatogonia colonies revealed that the spermatogonia divided within 2 days in our culture conditions (data not shown), which is ≈½ the rate of cell division for PAC2 cells. Our observed efficiency for the production of transgenic fish was considerably lower than the estimated moi. We currently do not understand what factors influence infection rates.

Although spermatogonial stem cells and primary spermatogonia can be infected with retrovirus (21, 23) or lentivirus (22) in vitro, there is no report demonstrating germ-line transmission when lentiviruses are injected into the seminiferous tubule (19) and retroviruses into the intraperitoneum of newborn mice (13) or even embryos in utero (16). Because Sertoli cells (19) and Leydig cells (13) were efficiently infected in these experiments, it is possible that germ cells may have lower sensitivity for infection from such viruses. Pluripotent stem cells possess mechanisms to prevent viral transduction by expressing specific proteins that bind the murine leukemia viral (MLV) LTR that prevent retrovial gene transcription (35). This effect might result in lower sensitivity for infection in germ cells from such viruses. Recently, De Miguel and Donovan (36) reported better infection efficiency of the avian leucosis viruses to mouse spermatogonia than MLV. Future efforts should focus on determining the efficiency of infection with other types of viruses and transfection by nonviral techniques such as electroporation.

Other approaches to improve the efficiency of transgenics are the improvement of the fertilization method and culture system. We showed here that 12-day culture is enough to detect the virus-mediated lacZ gene expression in sperm. Because it requires several thousand sperm to fertilize a single egg with our in vitro fertilization method, the evidence that transgenic embryos were obtained from 12-day-cultured sperm reveals that sperm were produced in significant numbers after 12 days in culture. Selection of transfected sperm by using the GFP gene and intracytoplasmic injection of the sperm (10, 37) could lead to a more efficient system to produce transgenics from sperm-based vectors.

We enzymatically dissociated cells of adult mature testes and removed the sperm; the culture then contained germ cells at various stages of differentiation and development ranging from spermatogonia to spermatids. Because retroviral transduction requires cell division, isolation of spermatogonia and longer, actively dividing culture would improve the efficiency of infection. Enrichment of spermatogonia will be performed by using transgenic fish expressing GFP using the vas promoter (38), which express GFP in spermatogonia and spermatocytes before the first meiotic division. Furthermore, we recently established two additional testicular cell lines of zebrafish: one, ZtA6-2, that stimulates proliferation of spermatogonia; and the other, ZtA6–12 (used in the present study), which stimulates the differentiation into sperm (28). Longer culture of spermatogonia without differentiation could be possible on the ZtA6-2 feeder cells. Subsequently, the spermatogonia will be transferred to the ZtA6–12 feeders to differentiate into sperm.

The use of zebrafish as a model system is growing rapidly as the research community has generated literally thousands of developmental mutants (39) supported with a full genome sequencing effort. However, like most model systems other than mice, no gene targeting method is available because of the lack of an embryonic stem cell culture system to transmit targeted or altered genes into the germ line. Recently, it has been reported that zebrafish short-term embryonic cell cultures are able to produced germ-line chimeras (40), and fertile transgenic zebrafish can be obtained by nuclear transfer using embryonic fibroblast cells from long-term culture (41). Although spermatogonial stem cells do not proliferate in our culture system without differentiating, the present study shows that sperm can be transfected in this culture and that the sperm can produce transgenic zebrafish by insemination in vitro. The longer culture durations from spermatogonia and scaling up the culture to produce more sperm will allow for the possibility of selection for homologous recombination events through antibiotic resistance.

Potentially the most exciting application of generating transgenic sperm from primary cultures is the potential to apply a similar technology to mammals, especially humans. The ability to culture human spermatogonia would allow gene therapy to occur before the embryo ever starts to develop, thus allowing prophylactic treatment for certain genetic disorders. Using the zebrafish cultures as a starting point, many fundamental aspects of culturing spermatogonia and their transfection in vitro can be determined with the ultimate goal of using the technique for many animal systems.