SEARCH

SEARCH BY CITATION

Keywords:

  • conditional;
  • Cre recombinase;
  • hcRFP;
  • heat shock;
  • inducible;
  • lacZ;
  • transgenic;
  • zebrafish

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Cre-mediated site-specific recombination has become an invaluable tool for manipulation of the murine genome. The ability to conditionally activate gene expression or to generate chromosomal alterations with this same tool would greatly enhance zebrafish genetics. This study demonstrates that the HSP70 promoter can be used to inducibly control expression of an enhanced green fluorescent protein (EGFP) –Cre fusion protein. The EGFP–Cre fusion protein is capable of promoting recombination between lox sites in injected plasmids or in stably inherited transgenes as early as 2 hr post–heat shock induction. Finally, the levels of Cre expression achieved in a transgenic fish line carrying the HSP70-EGFPcre transgene are compatible with viability and both male and female transgenic fish are fertile subsequent to induction of EGFP–Cre expression. Hence, our data suggests that Cre-mediated recombination is a viable means of manipulating gene expression in zebrafish. Developmental Dynamics 233:1366–1377, 2005. © 2005 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Zebrafish represent a vertebrate developmental model with many forward genetic advantages to study embryogenesis. Forward genetic screens have generated thousands of mutants defective in many different biological or developmental processes (Driever et al., 1996; Haffter et al., 1996; Golling et al., 2002). These studies would be complemented by the ability to specifically analyze the function of a single gene of interest through reverse genetics. In addition, for studies of organogenesis or regeneration, the ability to alter gene activity only after the completion of embryonic development would be of great benefit. However, reverse genetic techniques in zebrafish have lagged behind other model systems, although advances are beginning to occur rapidly. For example, an allelic series of rag1 mutations was generated by combining N-ethyl-N-nitrosourea mutagenesis with high-throughput sequencing of the rag1 gene (Wienholds et al., 2002). This method is a viable alternative to gene targeting to allow reverse genetics in zebrafish (Wienholds et al., 2002, 2003). Furthermore, short-term embryonic stem cell cultures have been achieved in zebrafish, suggesting that the reverse genetic approach of targeted mutagenesis may be possible in this model system (Lin et al., 1992; Ma et al., 2001; Ciruna et al., 2002). In addition to these new reverse genetic approaches, GAL4-UAS bitransgenic zebrafish lines have been generated successfully. These lines have been used for tissue-specific and temporally controlled transgene expression to mark cell types or ectopically express proteins (Scheer and Campos-Ortega, 1999; Köster and Fraser, 2001; Ma et al., 2001; Scheer et al., 2002). Here, we describe experiments designed to determine whether or not Cre-mediated site-specific recombination is (1) another feasible tool for reverse genetics, and (2) an alternative means for temporal and spatial control of transgene expression in zebrafish.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

To generate an inducible system for expressing Cre recombinase, an EGFPcre fusion gene (Le et al., 1999b) was placed under the control of a zebrafish heat shock–inducible HSP70 promoter (Fig. 1A). Earlier studies have shown efficient transient induction of GFP using this HSP70 promoter (Halloran et al., 2000). In addition, laser induction allowed control of expression to be restricted to single cells, while subjecting an entire embryo or fish to heat shock resulted in global induction of expression. Hence, by analogy, the HSP70-EGFPcre transgene should allow Cre expression to be controlled in specific cells or, alternatively, globally throughout the developing or adult zebrafish.

thumbnail image

Figure 1. Cartoon of HSP70-EGFP-cre and recombination reporter plasmids. A: HSP70-EGFP-cre construct. B: CMV driven hcRFP reporter construct. Gray triangles denote lox511 sites, and the red stop sign denotes the synthetic polyadenylation site and C2 termination site. C: PGK promoter driven lacZ reporter construct. Gray triangles denote loxP sites, and the red stop sign denotes the insertion of the neomycin (or G418) resistance gene and SV40 polyadenylation site. D: CMV-driven EGFP reporter (Cre Stoplight) construct. Gray triangles denote loxH (immediately 3′ to the CMV promoter) and loxP (3′ to the stop sign) sites, and the red stop sign denotes the insertion of the dsRFP gene and bovine growth hormone gene polyadenylation site. B–D: Small bent arrows denote transcriptional start sites, and the EGFP–Cre fusion protein is denoted by the fused green and blue spheres.

Download figure to PowerPoint

The original HSP70-EGFP construct was injected into one-cell embryos to generate mosaic transgenic zebrafish. These mosaic transgenic embryos at 24 hr postfertilization (hpf) were heat shock–induced for 90 sec at 42°C. Confirming the results of an earlier study (Halloran et al., 2000), no EGFP expression was seen in the absence of heat shock (Fig. 2A), but heat shock induction caused strong EGFP expression in a subset of cells (Fig. 2B). To test initially the Cre system, mosaic transgenic embryos were generated by injection of the HSP70-EGFP-cre vector into one-cell embryos. As expected, no EGFP-Cre expression was detected in the absence of heat shock (Fig. 2C). In contrast, strong mosaic EGFP-Cre expression (at comparable levels to the EGFP expression generated using the HSP70-EGFP vector) was seen subsequent to 90 sec of heat shock (compare Fig. 2D with 2B). Notably, initiation of EGFP-Cre expression was evident as early as 2 hr, appeared to achieve peak levels of expression by 4–6 hr and was still evident 24 hr post–heat shock induction.

thumbnail image

Figure 2. Heat shock regulated expression of EGFP and EGFP–Cre. A: Mosaic transgenic zebrafish (24 hours postfertilization [hpf]) for the HSP70-EGFP construct in the absence of heat shock. B: Mosaic transgenic zebrafish (24 hpf) for the HSP70-EGFP construct heated for 90 sec at 42°C. C: Mosaic transgenic zebrafish (24 hpf) for the HSP70-EGFP-cre construct in the absence of heat shock. D: Mosaic transgenic zebrafish (24 hpf) for the HSP70-EGFP-cre construct heated for 90 sec at 42°C. E: Confocal Z-stack showing a representative HSP70-EGFP-cre Tg/Tg zebrafish at approximately 24 hpf. Brightfield images of the embryos in A–D are shown in the insets. Scale bars = 250 microns in B (applies to inset and A–D), 500 microns in E.

Download figure to PowerPoint

Some of the mosaic HSP70-EGFP-cre transgenic fish were raised to sexual maturity. The offspring of these fish were examined for germline transmission of the transgene by heat induction for 90 sec at 42°C followed by screening under fluorescence imaging. Approximately 15% (84 of 575) of the progeny of one of the mosaic founders carried the transgene and demonstrated uniform expression of EGFP-Cre (Fig. 2E), whereas only 0.5% (3/660) of the progeny from a second founder carried the transgene (data not shown). Importantly, the frequency of long-term survival of fish carrying the EGFP-cre transgene subsequent to heat shock induction of EGFP-Cre was no different than wild-type hatch mates (Table 1). Both male and female transgenic fish were fertile. Furthermore, an intercross of fish carrying one copy of the transgene (Tg/+) yielded Mendelian ratios of Tg/Tg, Tg/+, and +/+ offspring (N = 27, 52, and 21, respectively). In this cross, the Tg/Tg offspring were distinguished from the Tg/+ offspring by the increased intensity and earlier uniform detection of EGFP fluorescence after heat shock induction (data not shown). Again, no increased lethality was observed in the heat shock–induced Tg/+ and Tg/Tg offspring.

Table 1. Survival of Offspring Following Heat Shock Induction at 42°C for 90 sec at 24 hpfa
Age that fish were analyzedTg-positiveTg-negative
  • a

    hpf, hours postfertilization; Tg, transgene.

1 day100% (N = 90)100% (N = 77)
2 days99%100%
3 days99%100%
4 days99%100%
5 days98%100%
6 days98%100%
1 week98%100%
2 weeks97%79%
3 weeks46%35%
6 weeks17%16%
12 weeks17%16%

To more accurately compare heat shock induction schemes in the HSP70-EGFP-cre Tg/+ embryos, a HSP70-EGFP-cre Tg/Tg male was crossed to a wild-type female and the resulting Tg/+ embryos from the same clutch were either maintained at 28.5°C or heat shocked under varying conditions. No fluorescent signal above the background autofluorescence was seen at 24 hpf in the un-induced Tg/+ embryos (Fig. 3A). The edge effect on the embryo in Figure 3A was not true EGFP fluorescence, rather it was due to using exposure conditions that allow focusing on the un-induced embryo and reflection from the yolk sac. In contrast, induction for 15 min at 37°C at 20 hpf resulted in low but detectable levels of EGFP-Cre fluorescence at 24 hpf (Fig. 3B). Furthermore, when the population of embryos for the various induction schemes was compared, 30 min of induction at 37°C at 20 hpf resulted in noticeably higher levels of EGFP-Cre fluorescence than the 15-min induction (compare Fig. 3B with 3C). Slight increases in intensity of fluorescence were seen for 60 min (or 270-min induction) at 37°C starting at 20 hpf (data not shown). In addition, an increasing number of the embryos in the heat shocked population achieved the highest fluorescence level with longer induction. Finally, a 5-min induction at 42°C at 20 hpf also gave strong EGFP-Cre fluorescence in the anterior end of the embryo, but fluorescence in the body was weaker than the 30-min 37°C induction (compare Fig. 3D with 3C). As we had already carefully monitored embryo survival after a 90-sec heat shock at 42°C and found no discernible survival differences relative to wild-type embryos (Table 1), we monitored the survival of embryos that were heat shocked under these different parameters for 2 weeks post-heat shock (Table 2). Importantly, these different regimens did not negatively affect embryo survival through 2 weeks post-heat shock.

thumbnail image

Figure 3. Analysis of heat shock induction schemes for HSP70-EGFP-cre transgenic line. A: Un-induced HSP70-EGFP-cre Tg/+ embryo photographed at 24 hours postfertilization (hpf). B: HSP70-EGFP-cre Tg/+ embryo photographed at 24 hpf after being induced for 15 min at 37°C at 20 hpf. C: HSP70-EGFP-cre Tg/+ embryo photographed at 24 hpf after being induced for 30 min at 37°C at 20 hpf. (D) HSP70-EGFP-cre Tg/+ embryo photographed at 24 hpf after being induced for 5 min at 42°C at 20 hpf.

Download figure to PowerPoint

Table 2. Viability of HSP70-EGFP-cre Fish Following Different Heat-Shock Parameters
Method of heat shockPre-heat shock1 day post-heat shock2 days post-heat shock1 week post-heat shock2 weeks post-heat shock
  1. aAll embryos were heat-shocked at 24 hours postfertilization. The number of embryos surviving at each time point for each heat shock treatment is shown. The percentage of embryos surviving each heat shock treatment at each time point is shown in parentheses.

37°C, 15 min117116 (99%)115 (98%)113 (97%)104 (89%)
37°C, 30 min113112 (99%)112 (99%)108 (96%)105 (93%)
37°C, 60 min127126 (99%)125 (98%)125 (98%)117 (92%)
37°C, 270 min151151 (100%)151 (100%)151 (100%)132 (87%)
42°C, 5 min106102 (96%)102 (96%)102 (96%)93 (88%)

The first stable HSP70-EGFP-cre transgenic line was examined by several means. First, Southern transfer analysis was used to verify integration of the transgene. As shown in Figure 4A, restriction fragments of the expected sizes were detected in the transgenic line, but not in wild-type fish, demonstrating that the line carries one copy of the transgene. Specifically, both HindIII–AflII (Fig. 4A) and BglII–AflII (data not shown) double digestions yielded the expected specific single band of 3.6 kb. Therefore, the transgene was intact from the BglII or HindIII sites in the polylinker upstream of the heat shock promoter through the AflII site following the SV40 polyadenylation site. There were no other useful sites 5′ to the HSP70 promoter, but the presence of a 4.5-kb hybridizing fragment upon double digestion with HindIII and StuI (Fig. 4A) or BglII and StuI (data not shown) revealed that the transgene was intact for another 0.9 kb 3′ to the SV40 polyadenylation site. Second, reverse transcriptase-polymerase chain reaction (RT-PCR) was performed to verify expression of the transgene after heat shock induction using primers upstream of the EGFPcre fusion point and downstream of the cre stop codon. As shown in Figure 4B, the expected PCR product was only detected in the samples containing reverse transcriptase. Furthermore, the 1.2-kb PCR product from two independent RT-PCR reactions was subjected to DNA sequencing, and no point mutations were found within the cre gene. Hence, a wild-type fusion transcript is conditionally expressed in this transgenic line. Next, Western blot analysis was used to detect expression of the fusion protein. As shown in Figure 4C, antisera against Cre detected the expected 68-kDa fusion protein in transgenic fish subjected to heat shock induction (Fig. 4, lanes 4, 6, and 8), but not in un-induced transgenic fish (Fig. 4, lane 2), un-induced wild-type fish (Fig. 4, lane 1), or in wild-type fish subjected to any of the heat shock induction schemes (Fig. 4, lanes 3, 5, and 7). In contrast, different heat shock induction strategies resulted in the accumulation of varying amounts of EGFP–Cre fusion protein in transgenic fish. Induction for 1 hr at 39°C (lane 6) appeared to induce the most fusion protein and 1 hr at 37°C (lane 4) induced an intermediate amount. Of interest, a 5-min heat shock induction at 42°C (lane 8) gave the lowest level of fusion protein, yet (as will be shown below) an even shorter (90 sec) 42°C heat shock induction actively allowed Cre-mediated recombination in the transgenic fish. Only low levels of Cre protein are required for recombination of substrates and useful murine Cre-expressing lines have not generated detectable levels of Cre protein using this antisera (B.S., unpublished data). Hence, the 42°C induction appears to generate experimentally useful levels of EGFP-Cre protein. Thus, this transgenic line conditionally generates the EGFP–Cre fusion protein.

thumbnail image

Figure 4. Analysis of HSP70-EGFP-cre transgenic line. A: Southern transfer analysis of HSP70-EGFP-cre transgenic line. B: Reverse transcriptase-polymerase chain reaction analysis of HSP70-EGFP-cre transgenic line. “H20” denotes no DNA control reaction. “+” and “−” denote addition and absence of reverse transcriptase, respectively. C: Western blot detection of EGFP–Cre fusion protein. Lane 1: un-induced wild-type fish; lane 2: un-induced HSP70-EGFP-cre Tg/+ fish; lane 3: wild-type fish heat shock induced for 1 hr at 37°C; lane 4: HSP70-EGFP-cre Tg/+ fish heat shock induced for 1 hr at 37°C; lane 5: wild-type fish heat shock induced for 1 hr at 39°C; lane 6; HSP70-EGFP-cre Tg/+ fish heat shock induced for 1 hr at 39°C; lane 7: wild-type fish heat shock induced for 5 min at 42°C; lane 8: HSP70-EGFP-cre Tg/+ fish heat shock induced for 5 min at 42°C.

Download figure to PowerPoint

Modeled on the reporter systems routinely used to monitor Cre recombination in mice (Lobe et al., 1999; Mao et al., 1999, 2001; Srinivas et al., 2001), an hcRFP recombination reporter construct was generated. A CMV promoter-driven hcRFP, pHcRed1-C1 (Clontech), was modified to include two lox511 sites flanking an RNA polymerase II termination signal between the promoter and translational start site of hcRFP (see Fig. 1B). This reporter construct was transformed into Cre-expressing Escherichia coli (Sauer and Henderson, 1988) to excise the termination signal and generate the expected final recombination product containing a single lox511 site (see Fig. 1B). When this recombination plasmid was injected into single-cell embryos, 78% of the embryos yielded mosaic hcRFP expression (Fig. 5B; Table 3), demonstrating the reporter vector should successfully monitor Cre-mediated recombination in zebrafish.

thumbnail image

Figure 5. Detection of hcRFP after heat shock induction of EGFP-Cre. A: Lack of detectable expression from the hcRFP reporter construct (depicted at the top of the panel) in mosaic transgenic embryos at 24 hours postfertilization (hpf). B: Detection of expression from final recombination product (depicted at the top of the panel) in mosaic transgenic embryos at 24 hpf. C–F: Detection of hcRFP-positive cells after heat shock induction of EGFP-Cre. C: Brightfield image showing the trunk of an HSP70-EGFP-cre embryo injected with the hcRFP reporter construct. D: EGFP fluorescent image of embryo shown in C, showing remaining EGFP intensity 24 hr post heat shock. E: hcRFP fluorescent image of embryo shown in C. Arrow denotes a cluster of hcRFP-positive cells 24 hr post heat shock. F: Artificial merge of brightfield and fluorescent images. Arrow denotes the cluster of hcRFP-positive cells shown in E. Scale bars = 250 microns in A (applies to A,B and insets).

Download figure to PowerPoint

Table 3. Analysis of Recombination Events in hcRFP Transient Injection Studies
Plasmid injectedInjected lineHeat shockHeat shock atAnalyzed atRFP− totalRFP+ totalN% RFP+Distribution of RFP-positive embryos by no. of positive cells per embryo
1–56–910–1920–2930–4950–100100+
  • a

    Same injection as group listed directly above.

lox2 hcRFPHSP70 EGFP-cre Tg/+37°C, 30 min5 hpf29 hpf6142070%5432000
lox2 hcRFPHSP70 EGFP-cre Tg/+37°C, 60 min24 hpf48 hpf14233762%15260000
lox2 hcRFPHSP70 EGFP-cre Tg/+NONEN/A24 hpf15251573%4100000
lox2 hcRFPaHSP70 EGFP-cre Tg/+NONEN/A48 hpf14261484%5100000
lox hcRFPwild-typeNONEN/A29 hpf11395078%65126631
lox hcRFPawild-typeNONEN/A48 hpf11395078%12786510

Finally, to test the fully integrated system, the hcRFP reporter construct was injected into one-cell embryos from crosses of wild-type and HSP70-EGFP-cre (Tg/Tg) fish. In the absence of heat shock, 97% of the embryos (152 of 157) exhibited a complete absence of hcRFP-expressing cells at 24 hpf (Fig. 5A; Table 3). Twenty-four hours after heat shock induction of EGFP-Cre, however, 62% to 70% of the injected embryos contained mosaic groupings of hcRFP-expressing cells (Fig. 5C–F; Table 3). Because the distribution of any injected reporter construct is mosaic, hcRFP fluorescence was only observed in clusters of cells, ranging from only a few hcRFP-positive cells per embryo (Fig. 5C–F) to more than 20 cells per embryo (Table 3).

To further verify the ability of EGFP-Cre to catalyze recombination in the transgenic line, a second recombination reporter expressing β-galactosidase subsequent to Cre-mediated site-specific recombination was used (Fig. 1C). Injection of the recombination product, derived by transformation of Cre-expressing Escherichia coli (Sauer and Henderson, 1988), into wild-type one-cell embryos led to mosaic expression of β-galactosidase at 24 hpf (Fig. 6A). At 48 hpf, 41% of the injected embryos (N = 37) exhibited β-galactosidase–expressing cells (data not shown). Although this reporter construct is not as robust as the hcRFP reporter, it should still detect Cre activity in the EGFP-cre line. Hence, the recombination substrate was injected into one-cell embryos from a HSP70-EGFP-cre (Tg/+) × wild-type cross, and the embryos were heat shocked at 24 hpf. One day subsequent to heat shock induction, at 48 hpf, no β-galactosidase–expressing cells were visualized in the EGFP-Cre–negative embryos (N = 17 tested). However, β-galactosidase–expressing cells were visualized in 21% of the EGFP-Cre–positive embryos (N = 67) (Fig. 6B,C). Furthermore, injection of the recombination reporter into wild-type embryos (N = 57) and subsequent heat shock induction did not generate any β-galactosidase–positive cells, demonstrating the requirement of Cre-mediated recombination for the observed positive cells (data not shown). Hence, the EGFP–Cre fusion protein conditionally expressed in this transgenic line is capable of catalyzing site-specific recombination of two independent recombination reporter constructs.

thumbnail image

Figure 6. Detection of β-galactosidase after heat shock induction of EGFP-Cre. A: Detection of expression from final lacZ recombination product (depicted at the top of the panel) in mosaic transgenic embryos at 24 hours postfertlization (hpf). B: Detection of β-galactosidase-positive (blue) cells after heat induction of a lacZ reporter injected HSP70-EGFP-cre transgenic embryo at approximately 42 hpf. C: Enlargement of boxed region of B.

Download figure to PowerPoint

Whereas the previous two experiments demonstrated the ability of the EGFP-cre transgenic line to recombine injected plasmids, it is ultimately important to verify that the transgenic line is capable of recombining chromosomally integrated lox sites. Hence, the EGFP-cre transgenic line was crossed to a pCMV-loxH-STOP-loxP-EGFP transgenic reporter line (Fig. 1D). Unfortunately, expression of the reporter from the CMV promoter was extinguished in this generation of the transgenic line. PCR analysis, however, was capable of detecting the expected site-specific recombination product subsequent to heat shock-induction. As shown in Figure 7, for the EGFP-cre Tg/+ × pCMV-loxH-STOP-loxP-EGFP reporter cross, PCR amplification of the unrecombined transgene produces two products of 4.0 and 1.6 kb (Fig. 7A), while the recombined transgene yields only a 2.8-kb fragment with the same PCR primers (Fig. 7B). Importantly, the 2.8-kb recombinant PCR product is only detected after heat shock induction (Fig. 7C). The parental (unrecombined) fragments of 4.0 and 1.6 kb are detected at all time points due to the use of pools of embryos (20–30 embryos; half of which do not contain the HSP70-EGFP-cre transgene); hence, a subset of these embryos cannot recombine the transgene. Samples were taken at various time points subsequent to heat shock induction to determine how rapidly the production of EGFP-Cre could induce recombination. The recombined product was first apparent at 2 hr post–heat shock induction (Fig. 7C), and increasing amounts of recombined product were detected from 4, 8, and 12 hr post–heat shock (data not shown) through 24 hr post–heat shock (Fig. 7C). DNA sequencing of the recombined product at all three time points confirmed the accuracy of the Cre-mediated recombination between the loxH and loxP sites in the reporter. Furthermore, Cre-mediated recombination only occurred in pools of embryos from which PCR analysis could detect the presence of the HSP70-EGFP-cre transgene (Fig. 7C). Finally, PCR analysis and DNA sequencing determined that the recombined reporter was stably inherited in the subsequent generation (data not shown).

thumbnail image

Figure 7. Detection of Cre-mediated site-specific deletion by polymerase chain reaction (PCR) analysis. A: Cartoon of two expected unrecombined PCR products. B: Cartoon of expected recombined PCR product. C: Analysis of recombination in EGFP-cre Tg/+ × pCMV-loxH-STOP-loxP-EGFP reporter Tg/+ cross and confirmation of presence of EGFP-cre Tg. “H2O” refers to the water (no DNA) control PCR reaction, and “no hs” refers to genomic DNA isolated from the transgenic cross in the absence of heat shock induction.

Download figure to PowerPoint

Because smaller PCR products are preferentially amplified over large products using the same primers and pools of embryos were used for this experiment, the relative levels of recombined and unrecombined products cannot be accurately assessed in this experiment. Therefore, to determine the true efficiency of Cre-mediated recombination, we performed individual embryo PCR on offspring from a HSP70-EGFP-cre Tg/+ × pCMV-loxH-STOP-loxP-EGFP reporter cross. Embryos were heat shocked at 24 hpf and EGFP-Cre–positive and EGFP-Cre–negative embryos were harvested for PCR at 48 hpf. Of the embryos that contained both the EGFP-cre and the reporter transgenes, we found that 100% (N = 139) contained the recombined reporter product. In addition, a recombined reporter product was not observed in those EGFP-Cre–negative embryos that only contained the reporter transgene (N = 24). Therefore, the induction of the HSP70-EGFP-cre transgene is capable of relatively rapidly recombining chromosomally integrated lox sites after heat shock induction with incredibly high efficiency.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Cre is a highly efficient site-specific recombinase isolated from bacteriophage P1 (Hamilton and Abremski, 1984; Hoess et al., 1984). Cre mediates recombination between two 34-bp loxP sites (or variants thereof), such that recombination between directly repeated loxP sites leads to deletion of the intervening DNA and recombination between inverted loxP sites leads to inversion of the intervening DNA (Nagy, 2000). Cre recombinase can also be used to generate specific chromosomal translocations (Smith et al., 1995; Van Deursen et al., 1995), large scale chromosomal deletions (Ramirez-Solis et al., 1995; Li et al., 1996), or repeated high-frequency insertion of transgenes into the same genomic location (Fukushige and Sauer, 1992; Call et al., 2000). Hence, Cre has become a genomic engineering tool of vast utility in murine studies. Given the possible uses of these recombination events in zebrafish, we examined the feasibility of Cre expression and site-specific recombination.

A heat shock promoter-driven EGFPcre fusion transcript was efficiently induced in a single-copy transgenic line and the resulting fusion protein catalyzed site-specific recombination of two separate, injected loxP and lox511 reporter constructs. Furthermore, the fusion protein promoted recombination of chromosomally integrated loxP and loxH sites in a transgenic line beginning 2–4 hr post–heat shock induction. Thus, the Cre recombinase is capable of recombining at least three different flavors of lox sites in zebrafish, suggesting that simultaneous deletion or inversion of several different transgenes carrying different variant of lox sites is possible.

Two independent experiments demonstrated that the Cre-mediated recombination was highly efficient. First, transient injection assays using the hcRFP reporter revealed that 62 to 70% of the EGFP-Cre–positive embryos catalyzed the recombination and expressed the reporter, which is comparable to 78% of the embryos injected with the recombined reporter exhibiting hcRFP expression (Table 3). Thus, the Cre-mediated production of mosaic hcRFP-expressing embryos was 90% of the level produced by injection of the previously recombined hcRFP reporter. A more appropriate test of the efficiency of EFGP-Cre was our PCR screen on 139 individual embryos that contained both the HSP70 EGFP-cre transgene and the pCMV-loxH-STOP-loxP-EGFP reporter transgene and the result that all 139 (100%) contained a recombined product. Taken together, these data indicate that the high efficiency of Cre-mediated recombination reported in other model systems can be achieved in zebrafish with the HSP70 EGFP-cre transgenic line.

Our transient reporter assays using the hcRFP reporter revealed that 3–4% of the injected EGFP-Cre embryos showed hcRFP-expressing cells, even in the absence of heat shock induction. This low level of recombination seen in the un-induced group may represent “leakiness” in the HSP70 promoter, possibly due to stressors other than heat or to promiscuous expression of HSP70 in certain tissues. In this vein, the original HSP70-EGFP transgenic line was reported to be “leaky” in the developing zebrafish eye (Halloran et al., 2000). Although we have not observed “leaky” EGFP-Cre expression in the eye or lens with the HSP70 EGFP-cre line, we have, in a few clutches, observed low-levels of EGFP in the developing notochord in un-induced embryos. Again, we attribute this either to this promoter being sensitive to other stressors besides heat in certain tissues or to promiscuous expression of HSP70 in the developing zebrafish notochord. This occasional EGFP expression in the notochord of uninduced embryos may account for the low level of hcRFP expression in the uninduced embryos, rather than a Cre-independent recombination event of the hcRFP reporter.

Rarely, Cre experiments have detected cytotoxicity in mammalian systems (Schmidt et al., 2000; Loonstra et al., 2001). Presumably, these results involve the utilization of the cryptic, yet functional, loxP sites that have been detected in mammalian genomes (Thyagarajan et al., 2000). Careful examination in tissue culture cells has revealed a dose-dependence to the cytotoxicity, with significant effects only occurring at very high levels of Cre expression (Loonstra et al., 2001). Experiments showing toxicity in mice involved expression of Cre from the protamine promoter. This highly active promoter produced Cre, while the histones were being replaced by protamines for compaction into the sperm head. Hence, the high level of Cre expression as well as the potentially sensitive form of chromosomes at this developmental stage may both contribute to the observed toxicity. In general, however, mouse lines expressing Cre have been relatively easy to generate and maintain as demonstrated by the current availability of lines on the Web site maintained by the Nagy laboratory (http://www.mshri.on.ca/nagy/Cre-pub.html).

Our studies suggest that induction of Cre expression does not have a significant adverse effect on the viability of transgenic zebrafish. Two generations of heat shock–induced offspring from the first HSP70-EGFP-cre founder have survived to sexual maturity, demonstrating that low level Cre expression is compatible with survival and absence of developmental defects. Moreover, 53% (90 of 167) of the progeny from a Tg/+ outcross to a wild-type fish carry the transgene, with both male and female offspring carrying the transgene being fertile subsequent to heat shock induction of EGFP-Cre. Finally, we are currently raising third-generation larvae and have not found any problems with inducing the expression of the transgene.

We demonstrated the successful generation of an inducible, stable Cre transgenic line in zebrafish. Although the ability to use Cre in numerous systems (Nagy, 2000; Werdien et al., 2001) suggested Cre use should be feasible in zebrafish, it remained important to experimentally demonstrate such feasibility and to check for cytotoxicity. Our initial studies suggest cytotoxicity will not be a critical issue for use of Cre in zebrafish. We are able to vary the amount of EGFP–Cre fusion protein generated by varying the length of heat shock and temperature used for the induction allowing optimization of the amount of protein for various recombination schemes. In conclusion, this study demonstrates that, in principle, the myriad of reverse genetic techniques for which Cre-mediated site-specific recombination is currently used in murine studies (Nagy, 2000; Branda and Dymecki, 2004) should be readily applicable in zebrafish as well. In addition, Cre-mediated recombination may serve either as an alternative means to or in conjunction with the GAL4-UAS system to spatially and temporally control transgene expression in zebrafish.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Fish Maintenance

Wild-type zebrafish (Danio rerio) were maintained under standard laboratory conditions as previously described (Westerfield, 1995; Detrich et al., 1999) using a light schedule of 14 hr on, and 10 hr off, at 28.5°C. Embryos were collected on the morning of spawning and fed paramecia (Carolina Biological) from approximately day 5 until 20 days of age. From that point, fish were fed a combination of dry food and brine shrimp daily.

One-Cell Embryo Injection Protocol

After maxiprep purification (Qiagen), plasmids were dialyzed in 0.5× TE overnight on floating dialysis membranes (Millipore) and then resuspended at final concentrations of 50 ng/μl or 100 ng/μl in 0.1 M KCl as previously described (Nüsslein-Volhard and Dahm, 2002). Dilutions from this stock solution were injected into one-cell zebrafish embryos as previously described (Meng et al., 1997; Nasevicius and Ekker, 2000) using a PLI-90 injector system (Harvard Apparatus, Inc.).

Generation of HSP70-EGFP-cre Construct

The construct HSP70-EGFP(HA) (Halloran et al., 2000) was digested with SalI and NotI. Blunt ends were generated using the Klenow fragment of E. coli DNA polymerase I. The plasmid pBS598 (Le et al., 1999b), containing EGFP-cre, was digested with MluI and a blunt end was generated using the Klenow fragment of E. coli DNA polymerase I. Next, the MluI digested plasmid was digested with Eco47 III to isolate the EGFP-cre restriction fragment. This EGFP-cre restriction fragment was ligated into the HSP70-EGFP(HA) vector fragment to generate the final HSP70-EGFP-cre construct (Fig. 1A). The entire cre reading frame was confirmed to be wild-type sequence in the final construct by DNA sequencing.

Generation of hcRFP Recombination Reporter Construct

pHcRed1-C1 (BD Biosciences) was digested with BglII and BamHI and subsequently religated to remove the majority of the polylinker. To generate a new polylinker, two annealed oligonucleotides (5′ CTAGGTACCGAATTCGTCGACGGATCCAAGCTTC 3′ and 5′ CTAGGAAGCTTGGATCCGTCGACGAATTCGGTAC 3′) were ligated into the unique NheI site just upstream of the start codon for hcRFP. A lox511 site, with a C at the second position to reduce RNA secondary structure, that might affect translation efficiency (Le and Sauer, 2001), was cloned into the new unique HindIII site (generated in the previous step), using the following annealed oligonucleotides: 5′AGCTACAACTTCGTATAATGTATACTATACGAAGTTAG 3′ and 5′ AGCTCTAACTTCGTATAGTATACATTATACGAAGTTGT 3′. A BamHI–EcoRI restriction fragment containing SPAC2 (Eggermont and Proudfoot, 1993) was cloned into the unique BamHI and EcoRI sites in the new polylinker. SPAC2 consists of a synthetic polyA site followed by a C2 complement gene region generating a very strong RNA polymerase II termination signal. Finally, a second lox511 site with the same position 2 alteration was cloned into the plasmid, in the same orientation as the first lox511 site, at the unique KpnI site in the polylinker using the following annealed oligonucleotides: 5′ TACAACTTCGTATAATGTATACTATACGAAGTTAGTAC 3′ and 5′ TAACTTCGTATAGTATACATTATACGAAGTTGTAGTAC 3′.

The final construct contains the lox511 flanked SPAC2 sequence inserted between the CMV promoter and the hcRFP coding region (see Fig. 1B). At each step, the alterations were confirmed by DNA sequencing. In addition, the E. coli strain BS591 (Sauer and Henderson, 1988; Le and Sauer, 2001), which expresses Cre recombinase, was transformed. The resulting plasmids had the expected excision recombination product demonstrated by DNA sequencing, showing that the reporter construct was capable of undergoing Cre-mediated excision.

Generation of lacZ Recombination Reporter Construct

To construct the lox2-neoSTOP lacZ reporter plasmid pBS543, a 600-bp EcoRI fragment carrying the granulocyte colony-stimulating factor gene (G-CSF) polyadenylation signal from pEF-BOS (Mizushima and Nagata, 1990) was cloned into the EcoRI site of pBS428 (Bethke and Sauer, 1997). The resulting plasmid was digested with XbaI and NheI and re-ligated at their compatible sticky ends to remove a 346-bp fragment and thereby fuse the lacZ gene to the G-CSF polyadenylation signal. In the resulting pBS543 reporter plasmid, a 2.7-kb insertion into the 5′ end of the lacZ gene prevents lacZ expression from the upstream mouse PGK promoter. Because the insertion is a loxP-flanked cassette of the neo gene and the SV40 polyadenylation site, Cre-mediated excision at the loxP sites restores the proper reading frame to allow expression of the resulting recombinant loxPlacZ fusion gene (Bethke and Sauer, 1997) (Fig. 1C).

Generation of the pCMV-loxH-STOP-loxP-EGFP Recombination Reporter

To create the pCMV-loxH-STOP-loxP-EGFP recombination reporter, a multiple cloning site was introduced between the XhoI and NruI restriction sites of pT2KXIG, which contains the Tol2 transposable element (Kawakami et al., 1998). The pCMV-loxH-STOP-loxP-EGFP fragment was isolated from the pYY12 (Cre stoplight) plasmid (Yang and Hughes, 2001) as a SalI fragment and cloned into the SalI restriction site in the modified Tol2 vector. While this transgene (Fig. 1D) was transiently expressed in the injected zebrafish, the CMV promoter became transcriptionally silent in the F1 generation.

Generation of pCMV-loxH-STOP-loxP-EGFP Transgenic Zebrafish

After Maxiprep purification (Qiagen), plasmids were extracted with phenol:chloroform (1:1), precipitated, and resuspended in nuclease-free water. The plasmid pCSTZ2.8, which contains the Tol2 transposase cDNA, was linearized with NotI and used as the template to in vitro transcribe Tol2 transposase mRNA with the SP6 mMESSAGE mMACHINE kit (Ambion). The transcribed RNA was purified on a mini Quick Spin RNA column (Roche), extracted with phenol:chloroform (1:1), precipitated, and resuspended in nuclease-free water (K. Kawakami, personal communication). A solution containing 25 ng/μl of recombinant Tol2 target plasmid DNA and 25 ng/μl of transposase RNA was injected into embryos at the two- to four-cell stage as previously described (Kawakami et al., 2000) using an IM 300M microinjection system (Narishige).

After injection, embryos were raised to adulthood and crossed to the AB wild-type strain. Positive founders, containing the pCMV-loxH-STOP-loxP-EGFP recombination reporter were identified by PCR amplification of the transgene from genomic DNA isolated from embryos at 4–7 days postfertilization. The positive founders were remated with the AB strain, and the F1 progeny were raised to adulthood. Fin clips of the F1 progeny were used to isolate genomic DNA, which was PCR amplified to identify F1 individuals that contained the pCMV-loxH-STOP-loxP-EGFP transgene and served as the source for individual families.

Fluorescence Imaging and Screening for Germline Transmission of HSP70-EGFP-cre Transgene

After initial injection of the plasmids, embryos were incubated at 42°C for 90 sec (heat shock) to induce the HSP70 promoter. Similarly, mosaic transgenic founders were crossed with wild-type zebrafish (Danio rerio) and resulting progeny were heat shock induced for 90 sec at 42°C. Approximately 4 hr post–heat shock, fluorescence was visualized using a Leica MZFL III dissection microscope with a ×2 objective using a GFP2 filter set. For experiments using the hcRFP reporter construct, the same microscope was used with a Texas Red filter set and fluorescence was visualized 24 hr post–heat shock induction.

Examination of Heat Shock Induction Strengths and Subsequent Survivability

To assess accurately different heat shock induction protocols, the same clutch of offspring from a HSP70-EGFP-cre Tg/Tg X wild-type cross was used for the un-induced control and various induction schemes. Embryos were photographed using a manual setting of 4 sec using the MZFl III microscope and with a ×2 objective using a GFP2 filter set. This setting allows focusing on un-induced Tg/+ embryos and did not overexpose the induced Tg/+ embryos. Using this photography, an accurate representation of relative induction strengths can be obtained. This experiment was repeated to obtain a greater number of embryos to assess survivability under different heat shock induction schemes (Table 2). Dead embryos were recorded and removed daily as fish were monitored for 2 weeks post–heat shock.

Oligonucleotide Synthesis and DNA Sequencing

Oligonucleotides were synthesized by Integrated DNA Technologies, Inc., and Invitrogen. DNA sequencing needed for the experiments described in Figures 1–6 was performed by the University of Kansas Medical Center Biotechnology Support Facility staff using ABI Prism BigDye Terminators version 3 and cycle sequencing with Taq FS DNA polymerase. DNA sequence was collected and analyzed on an ABI Prism 377XL automated DNA sequencer (PE Applied Biosystems Division). For the experiments described in Figure 7, DNA sequencing was performed by Sequetech (Mountain View, CA) using a ABI Prism 3730 automated DNA sequencer.

Southern Blot Analysis of HSP70-EGFP-cre Transgenic Line

DNA was isolated from one hundred, 24 hpf embryos containing the transgene (as assayed by detection of GFP 4 hr after a 90-sec 42°C heat shock induction) as well as from wild-type control embryos. Isolated DNA (10 μg) was digested with the appropriate restriction enzymes and electrophoresed through a 0.8% GTG agarose gel (BioWhittaker Molecular Applications). DNA was transferred to nitrocellulose (Schleicher and Schuell) and probed with a 1.1-kb XhoI–MluI fragment encompassing the cre gene isolated from pBS185 (Gibco/BRL; Sauer and Henderson, 1990). Southern blot analysis was performed using standard conditions (Sambrook et al., 1989).

RT-PCR and Sequencing of cre Transcript

Total RNA was isolated using TRIZOL reagent (Gibco/BRL) from 100 transgene-positive 24 hpf embryos after 90 sec of heat shock induction at 42°C. DNA-free (Ambion) was used to remove contaminating genomic DNA. Next, oligo(dT) was used to prime cDNA synthesis using Superscript II (Invitrogen). Finally, the following primers were used to amplify the fragment using BioXAct (Bioline): 5′ GTGCTGCTGCCCGACAACC 3′ and 5′ CAAATGTGGTATGGCTGATTATGATC 3′. The following cycling conditions were used: 95°C for 5 min; 20 cycles of 95°C for 1 min, a 0.5°C/cycle decrease from 70°C to 60°C for 90 sec/cycle, 72°C for 45 sec; 10 cycles of 95°C for 1 min, 60°C for 90 sec, 72°C for 45 sec; a final 10-min 72°C extension; and then cooling to 4°C. The resulting fragment was purified using QiaQuick (Qiagen) and sequenced as described above using the following specific primers: 5′ CAGTCCGCCCTGAGCAAAG 3′, 5′ CAAGTGACAGCAATGCTG 3′, 5′ GTTCGAACGCTAGAGCCTG 3′, and 5′ GTTAATGGCTAATCGCCATC 3′.

Western Blot Analysis

Tail and body wall tissue from adult un-induced and heat shock induced wild-type and HSP70-EGFP-cre Tg/+ transgenic fish for the following induction schemes (1 hr induction at 37°C, 1 hr induction at 39°C, and 5 min induction at 42°C) were collected in RIPA (20 mM Tris pH 8.0, 137 mM NaCl, 10% glycerol, 1% Igepal CA-630, 0.1% sodium dodecyl sulfate, 0.5% Na deoxycholate, 2 mM ethylenediaminetetraacetic acid, 5 mM sodium orthovanadate, 5 mM benzamidine, and 1 mM phenylmethyl sulfonyl fluoride), dounced, and centrifuged for 5 min at 14,000 × g at 4°C. Supernatants were collected and quantitated using the Bio-Rad Protein Assay (Bio-Rad). Extracts (50-μg) were subjected to 0.8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electroblotted to activated Immobilon-P membranes (Millipore). Membranes were blocked with 5% dry milk in 1× phosphate buffered saline with 0.1% Tween 20 (PBSTw). Next, membranes were incubated with a 1:4,000 dilution of rabbit anti-Cre antibody (Sauer, 1987; Le et al., 1999a) in PBSTw with 1% bovine serum albumin (BSA) for 1 hr at room temperature followed by four separate 5-min washes in PBSTw. Then, membranes were incubated with a 1:40,000 dilution of anti-rabbit horseradish peroxidase–conjugated secondary antibodies (Jackson ImmunoResearch) in PBSTw with 1% BSA for 1 hr at room temperature and the same wash regimen was used as above. The Supersignal West Pico Chemiluminescent kit (Pierce) and a 30-sec exposure to X-OMAT 5 film (Kodak) were used for visualization.

Analysis of Cre-Mediated Recombination of the pCMV-loxH-STOP-loxP-EGFP Transgene in Pooled Embryos

Transgenic fish, that contained the pCMV-loxH-STOP-loxP-EGFP recombination reporter, were crossed with the HSP70-EGFP-cre transgenic zebrafish line. The progeny were heat shocked at 42°C for 5 min at 24 hpf. The embryos were raised until the desired age before genomic DNA was isolated for analysis.

Genomic DNA was isolated from pools of 20–30 F1 embryos using the standard high molecular weight isolation method from The Zebrafish Book (Westerfield, 1995). The genomic DNAs were used as templates for PCR amplification of the pCMV-loxH-STOP-loxP-EGFP recombination reporter using two forward primers (primer A, 5′ AGACAGACAATCTAATGCCAG 3′, complementary to the 5′ Tol2 sequences flanking the transgene; and primer B, 5′ TTGGAGTCCACGTAGTAGTAGCC 3′, complementary to dsRFP sequences) and a common reverse primer (primer C, 5′ CACTTCATTCTATTTTCCCTTGC 3′), complementary to the 3′ Tol2 sequences flanking the transgene (see Fig. 7A,B). PCR fragments were amplified using a Peltier 200 thermal cycler (MJ Research) under the following conditions: 2 min at 94°C, followed by 35 cycles of 94°C for 30 sec, 62°C for 1 min, and 68°C for 6 min. A final 10-min incubation at 68°C was used to ensure the complete extension of all the PCR products. To verify the presence of the HSP70-EGFP-cre transgene in the pools of embryos tested for Cre-mediated recombination, the same primers and PCR conditions were used as described for the RT-PCR and sequencing of the cre transcript. The PCR amplification products were separated on a 0.8% agarose gel.

Analysis of Cre-Mediated Recombination of the pCMV-loxH-STOP-loxP-EGFP Transgene in Individual Embryos

Offspring from a HSP70-EGFP-cre Tg/+ X pCMV-loxH-STOP-loxP-EGFP reporter cross were heat shocked at 42°C for 5 min at 24 hpf. Single embryos were harvested at 24 hr post–heat shock (48 hpf) and placed into 0.2 μl thin-wall PCR tubes. The tubes were centrifuged for 20 sec and placed in the −80°C freezer for at least 30 min. The PCR tubes were removed from the freezer and placed on ice, 10 μl of extraction buffer was added (10 mM Tris pH 8.0, 50 mM KCl, 2.5 mM MgCl2, 0.45% Tween-20, and 200 μg/ml proteinase K), and the tubes were spun briefly. The embryos were digested by incubation for 3 hr at 56°C, followed by a 15-min incubation at 95°C to inactivate the proteinase K. Before PCR, samples were briefly spun in a microfuge for 30 sec and placed on ice. To amplify genomic DNA, 40 μl of PCR mix (the same mix described in the preceding experiment with pooled embryos except that primer B was left out) was added. Cycling conditions were the same as described in the preceding experiment.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The authors thank M. Halloran for the HSP70 promoter construct, N. Proudfoot for the polII termination sequence, K. Kawakami for his generous gift of the Tol2 (pT2KXIG) and Tol2 transposase cDNA (pCSTZ2.8) plasmids, T.E. Hughes for the Cre stoplight plasmid, E. Roach and S. Fernald for assistance with preparation of the figures used in this article, the University of Kansas Medical Center Biotechnology Support Facility staff for DNA sequencing, and K.R. Peterson for critical reading of the manuscript. R.T. was supported, in part, by the KUMC Biomedical Training Grant.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES