Targeted gene expression by the Gal4-UAS system in zebrafish
Division of Molecular and Developmental Biology, National Institute of Genetics, and Department of Genetics, Graduate University for Advanced Studies (SOKENDAI), 1111 Yata, Mishima, Shizuoka 411-8540, Japan
Division of Molecular and Developmental Biology, National Institute of Genetics, and Department of Genetics, Graduate University for Advanced Studies (SOKENDAI), 1111 Yata, Mishima, Shizuoka 411-8540, Japan
*Author to whom all correspondence should be addressed. Email: firstname.lastname@example.org
Targeted gene expression by the Gal4-UAS system is a powerful methodology for analyzing function of genes and cells in vivo and has been extensively used in genetic studies in Drosophila. On the other hand, the Gal4-UAS system had not been applied effectively to vertebrate systems for a long time mainly due to the lack of an efficient transgenesis method. Recently, a highly efficient transgenesis method using the medaka fish Tol2 transposable element was developed in zebrafish. Taking advantage of the Tol2 transposon system, we and other groups developed the Gal4 gene trap and enhancer trap methods and established various transgenic fish expressing Gal4 in specific cells. By crossing such Gal4 lines with transgenic fish lines harboring various reporter genes and effector genes downstream of UAS (upstream activating sequence), specific cells can be visualized and manipulated in vivo by targeted gene expression. Thus, the Gal4 gene trap and enhancer trap approaches together with various UAS lines should be important tools for investigating roles of genes and cells in vertebrates.
Gal4 is a yeast transcriptional activator consisting of 881 amino acids (Fig. 1A). The DNA binding activity of Gal4 is located in the 74 amino acids in the N-terminus (Keegan et al. 1986). The transcriptional activation function of Gal4 is mapped in two regions (residues 148–196 and 768–881) (Ma & Ptashne 1987). Gal4 binds to its specific recognition sequence UAS (upstream activating sequence) and activates transcription of target genes. It has been demonstrated that the Gal4-UAS system can operate not only in yeast but also in various animal cells (Fischer et al. 1988; Ornitz et al. 1991; Scheer & Campos-Ortega 1999; Hartley et al. 2002).
The Gal4-UAS system was used as a two-component gene expression system in Drosophila (Brand & Perrimon 1993). A fly line expressing Gal4 and a fly line carrying the lacZ gene downstream of UAS were created. In the double transgenic progeny obtained from a genetic cross of these lines, LacZ was induced by Gal4. An advantage of this methodology is that not only the lacZ gene but also any gene downstream of UAS can be induced by the Gal4 activity. Since the P-element-mediated enhancer trapping can create a number of fly lines expressing Gal4 in specific cells, expression of a gene downstream of UAS can be targeted by using such Gal4 lines.
The Gal4-UAS system had not been applied effectively to vertebrate systems mainly due to the lack of an efficient transgenesis method. Recently, a highly efficient transgenesis method based on the medaka fish Tol2 element was developed in zebrafish (Kawakami et al. 2004). Taking advantage of the Tol2 transposon system, we and other groups developed the Gal4 gene trap and enhancer trap methods and established various transgenic Gal4 lines and UAS lines in zebrafish.
The Gal4-UAS system: before Tol2
The Gal4-UAS system was first applied to zebrafish by Scheer and Campos-Ortega (Scheer & Campos-Ortega 1999). To construct a Gal4-expressing fish line, plasmids containing the full-length yeast GAL4 gene downstream of ubiquitous promoters, such as the SV40/thymidine kinase enhancer/promoter or the carp β-actin promoter, were constructed. To construct UAS effector fish, a plasmid carrying a constitutively active form of the Notch receptor downstream of UAS (UAS:myc-notch1a-intra) was created. These plasmid DNAs were injected into fertilized eggs and the transgenic lines were established. In the double transgenic offspring from the cross of these fish, the myc-notch1a:intra protein was expressed in the regions where the GAL4 transcript was expressed. These results indicate that the full-length Gal4 can direct expression of a gene placed downstream of UAS in zebrafish.
To achieve stronger induction of a UAS transgene, Koster & Fraser (2001) used Gal4-VP16 (Fig. 1B), which has a stronger transcriptional activity than the full-length Gal4 (Sadowski et al. 1988; Koster & Fraser 2001). Gal4-VP16 consists of the DNA-binding domain from Gal4 and the transcriptional activation domain from the herpes simplex virus VP16 protein (Sadowski et al. 1988). Plasmid constructs, that harbor both the Gal4-VP16 gene downstream of a tissue-specific promoter and a reporter gene downstream of UAS tandemly, were injected into zebrafish embryos. In the injected embryos, high levels of reporter gene expression were observed in many cells in the notochord, neural tube, muscle and skin, regions where the tissue-specific promoter was thought to be active. Thus, Gal4-VP16 can induce expression of a gene placed downstream of UAS strongly in a tissue-specific manner. It was proposed that this Gal4-UAS transient expression method can be used to express effector genes ectopically and to label cells for time-lapse studies.
Sagasti et al. created a stable transgenic line harboring both the Gal4-VP16 gene under the control of an enhancer from the zebrafish islet-1 gene and the UAS:EGFP (enhanced green fluorescent protein) gene tandemly. This transgenic fish was created by integrating a plasmid in the genome with the meganuclease (I-SceI)-mediated transgenesis method (Thermes et al. 2002; Sagasti et al. 2005). In the transgenic embryos, the sensory neurons were labeled with green fluorescent protein (GFP) fluorescence, which allowed the investigation of the shapes and sizes of individual sensory arbors. However, in all of the transgenic lines that they established, the Gal4-VP16-driven GFP expression patterns varied and were observed only in a small population of cells in a region where the islet-1 enhancer should be active. This phenomenon is called variegated (or mosaic) expression. Scott et al. (2007) created a transgenic line by injecting the plasmid DNA carrying both the brn3c promoter-driven Gal4-VP16 and a membrane-bound GFP gene downstream of UAS. In the transgenic embryos, a small subset of retinal ganglion cells expressing Gal4-VP16 was labeled with GFP fluorescence, indicating that the Gal4-VP16-mediated GFP expression was also highly variegated.
Although these works are promising, only a handful of research that used transgenic Gal4 lines and UAS-transgene lines have been reported to date. This is mainly because of the following reasons. First, the transgenesis efficiencies with conventional transgenesis methods, either microinjection of plasmid DNA or the I-SceI-mediated method (Thermes et al. 2002), are not high enough to create hundreds of different Gal4 transgenic lines. Second, the availability of specific enhancers or promoters for targeting Gal4 expression in tissue or cell-type specific manners had been limited in zebrafish. Finally, the enhancers/promoters that are thought to target Gal4 expression did not always work as expected due to variegated expression.
The Tol2 transposon system
For the last 10 years, we have been developing transposon techniques in zebrafish by using the Tol2 transposable element (Kawakami 2007). We cloned a cDNA encoding the transposase protein from the Tol2 element and developed a two-component transposition system in zebrafish (Kawakami et al. 1998; Kawakami & Shima 1999). In this system, a transposon-donor plasmid harboring a Tol2 construct, that contains cis-sequences essential for transposition and a DNA insert between them, and the transposase mRNA, that is synthesized in vitro by using the cDNA as a template, are co-injected into fertilized eggs (Fig. 3A). The Tol2 construct is excised from the plasmid and integrated in the genome in the germ lineage (Kawakami et al. 2000). Since the transposon insertions are more efficiently transmitted to the next generation, in comparison to the DNA microinjection method or another vertebrate transposon system, Sleeping Beauty (Ivics et al. 1997; Davidson et al. 2003), the Tol2-mediated transgenesis has now become a standard approach to create transgenic zebrafish. Furthermore, since this method enables us to create hundreds and thousands of insertions in F1 fish, it has been applied to gene trapping and enhancer trapping. Random integrations of Tol2-based gene trap and enhancer trap constructs were created, and a number of transgenic fish lines expressing the GFP reporter gene in specific tissues, cells and organs have been generated (Kawakami et al. 2004; Parinov et al. 2004; Nagayoshi et al. 2008). These prompted us to pursue development of a Gal4-UAS system by using the Tol2 transposon system in zebrafish.
Tol2-mediated Gal4 enhancer trapping
Scott et al. (2007) created two types of enhancer trap constructs carrying the Gal4-VP16 gene downstream of long (1.5 kilo base pairs [bp]) and short (600 bp) versions of the zebrafish hsp70 promoter, generating HSP:GAL4 and HSP(600):GAL4, respectively (Fig. 2A,B). A plasmid DNA carrying either of these was injected into embryos with the Tol2 transposase mRNA (Fig. 3A). In the injected embryos, the trap construct is excised from the plasmid and integrated into the chromosome, thereby creating genomic insertions in germ cells. The insertions are transmitted to the next generation and express Gal4-VP16 when activated by a chromosomal enhancer(s). To visualize the Gal4-VP16 expression, the UAS:Kaede reporter fish, which carries a gene for the photoconvertible fluorescent protein Kaede (Ando et al. 2002) downstream of 14 × UAS (14-times repeat of a UAS module), was crossed with the injected fish. In the resulting embryos, various Kaede expression patterns were observed.
We generated two types of enhancer trap constructs by using another version of transcriptional activator Gal4FF (Fig. 1C,D) (Asakawa et al. 2008). Gal4FF contains the DNA-binding domain of Gal4 and two short transcriptional activation motifs from VP16, which have a weaker transcriptional activity than VP16 (Seipel 1992; Baron 1997). T2KhspGGFF contains a fusion gene between enhanced EGFP and Gal4FF (GGFF) placed downstream of a 650-bp fragment of the hsp70 promoter (Fig. 2C). T2KhspGFF contains the Gal4FF gene downstream of the same promoter (Fig. 2D). To visualize the GGFF or Gal4FF expression, the UAS:EGFP reporter fish harboring the EGFP gene downstream of 5 × UAS was crossed with the injected fish (Fig. 3A). Specific GFP expression patterns were observed in the resulting embryos (Fig. 3B). The enhancer trap events were detectable by fluorescence from the GGFF protein, but the fluorescence became much stronger when crossed with the UAS:EGFP fish. Thus, both GGFF and Gal4FF work as transcriptional activators.
These four Gal4 enhancer trap constructs worked essentially similarly. All of the constructs accomplished a high frequency of enhancer trapping and generated transgenic fish expressing Gal4 in various tissues from embryonic stages to adulthood. Importantly, these Gal4 enhancer trap lines showed a stable and consistent expression of the UAS-reporter and did not suffer from silencing of the transgene expression after passage of generations, which has been sometimes observed in transgenic fish created by the plasmid injection method. These revealed that the use of the Tol2-mediated transgenesis and the hsp70 promoter is effective for creating stable Gal4 transgenic lines.
Tol2-mediated Gal4 gene trapping
Two types of Tol2-based Gal4 gene trap constructs have been described. Davison et al. (2007) created a self-reporting SAGVG construct containing both the Gal4-VP16 gene placed downstream of a splice acceptor sequence and the UAS:EGFP reporter fragment (Fig. 1E). In the transgenic embryos, various GFP expression patterns were observed, indicating that Gal4-VP16 induces expression of the UAS:EGFP reporter on the same construct. As they expected, when the SAGVG transgenic fish were crossed with the UAS:Kaede reporter fish, a robust Kaede expression was observed in cells expressing GFP, indicating that Gal4-VP16 can also activate another transgene in trans under the control of UAS. Two cases where UAS:Kaede was not activated in the hindbrain or heart valve despite the activation of UAS:EGFP were mentioned. These may be due to Gal4-VP16-independent expression of UAS:EGFP.
We created T2KSAGFF containing the Gal4FF gene downstream of a splice acceptor sequence (Fig. 2F) (Kawakami et al. 2004; Kotani et al. 2006; Asakawa et al. 2008). The embryos injected with T2KSAGFF were raised and crossed with the UAS:EGFP reporter fish (Fig. 3A). In the resulting embryos, various GFP expression patterns were observed. We identified 129 GFP expression patterns in the F1 generation from 250 injected fish (Fig. 3B). This frequency is high enough for a small laboratory to collect hundreds of different Gal4FF expression patterns.
The splice acceptor of the rabbit β-globin gene was used in these gene trap constructs. A fusion of the 5′ upstream exon and the Gal4FF or Gal4-VP16 sequence was detected by reverse transcription–polymerase chain reaction (RT–PCR) and 5′ rapid amplification of cDNA ends (RACE), indicating that Gal4FF and Gal4-VP16 were indeed expressed dependently on endogenous transcription (KA and KK, unpubl. data., year) (Davison et al. 2007).
Visualization of Gal4-expressing cells: UAS-reporter lines
We constructed the UAS:EGFP and the UAS:mRFP1 reporter fish carrying the EGFP gene and the mRFP1 gene, respectively, downstream of 5 × UAS and a minimal TATA sequence (Table 1). Both of these reporter fish carry a single genomic insertion of the reporter construct and are able to visualize Gal4 expression in various cell types in embryos and adult fish (Asakawa et al. 2008) (KK unpubl. obs). Scott et al. used the 14 × UAS:Kaede fish in the Gal4 enhancer trapping to visualize Gal4 expression (Table 1) (Scott et al. 2007). An advantage of this approach is that the Gal4-expressing cells can be further dissected by photoconversion upon irradiation of UV or violet light (Ando et al. 2002; Sato et al. 2006). The photoconversion of Kaede and KikGR was applied to visualize the morphology of Gal4-expressing neurons (Table 1) (Tsutsui et al. 2005; Hatta et al. 2006; Scott et al. 2007). Also, migrating cells can be marked and traced by Kaede in vivo during organogenesis (Davison et al. 2007). For analysis of morphology of neurons, a UAS-reporter line carrying a gene for Dronpa, whose fluorescent intensity can be changed by light, is also available (Table 1) (Ando et al. 2004; Aramaki & Hatta 2006).
Table 1. Upstream activating sequence (UAS) reporter and effector lines
When these UAS reporters are used for visualization of Gal4 expression, the following two things need to be considered. One is a time lag between transcription of reporter gene and detection of reporter protein. The other is a turnover rate of reporter protein. Because of these two factors, reporter expression does not always recapitulate Gal4 expression (Phelps & Brand 1998). For example, because EGFP and mRFP1 are relatively stable, these can be detected even after the Gal4 expression is terminated (Asakawa et al. 2008) (K. Asakawa and K. Kawakami, unpubl. data, 2005).
Modifications of activities of Gal4-expressing cells: UAS-effector lines
An advantage of the Gal4-UAS approach is to modify functions of the Gal4-expressing cells by targeted expression of any gene of interest. We will introduce currently available transgenic lines carrying such effector genes downstream of UAS.
Analysis of gene function by targeted gene expression
Scheer and Campos-Ortegas established the UAS:myc-notch1a:intra line to induce a constitutively active form of the Notch receptor (Table 1) (Scheer & Campos-Ortega 1999). When transcription of UAS:myc-notch1a:intra was activated by deltaD:GAL4 or hsp70:GAL4, a large proportion of retinal cells entered glial development, revealing an instructive role of Notch in promoting gliogenesis in the zebrafish retina (Scheer et al. 2001). The hsp70:GAL4; UAS:myc-notch1a:intra double transgenic embryo was also used to examine a role of Notch in repressing the her3 gene activity (Hans et al. 2004), in development of arterial and venous blood vessels (Lawson et al. 2001), and in migration of angiogenic cells in artery formation (Siekmann & Lawson 2007). Jeong et al. (2007) constructed the UAS:fezl line to misexpress fezl, a zinc-finger-containing gene (Table 1) (Hashimoto et al. 2000; Jeong et al. 2007). The heat induction of fezl in the hsp70:GAL4; UAS:fezl double transgenic embryos revealed a crucial role of fezl in patterning diencephaon (Jeong et al. 2007). These works demonstrated the effectiveness of the GAL4-UAS method in elucidating gene function. In the future, establishing and sharing transgenic fish carrying various genes under the control of UAS will facilitate studies on gene function.
UAS:nfsB-mCherry: targeted cell ablation
To understand an in vivo role of a specific cell-type, ablation of such cells is an effective way. Davison et al. (2007) constructed a transgenic fish line carrying the Escherichia coli nfsB gene downstream of UAS (Table 1). The nfsB gene encodes nitroreductase B (NTR). The NTR protein converts a non-toxic prodrug (a nitroimidazole substrate such as Metronidazole) into a cytotoxic DNA cross-linking agent that kills a cell. The UAS:NTR-mCherry fish was crossed with a line expressing Gal4 in the floorplate or in the notochord. When the resulting double transgenic embryos were treated with the prodrug, the mCherry-positive cells were ablated by apoptosis. The NTR system has been shown to kill pancreatic β cells, cardiomyocytes and hepatocytes in zebrafish (Curado et al. 2007; Pisharath et al. 2007). The UAS:NTR-mCherry fish will allow us to ablate specific cell types by using appropriate Gal4 lines. It remains to be tested whether this system is applicable to all cell types.
UAS:iGluR6(L439C): remote control of neuronal activity by light
Szobota et al. (2007) developed the UAS:iGluR6(L439C) line to control the activity of neurons by light (Table 1). iGluR6(L439C) is a genetically engineered ionotropic glutamate receptor iGluR6, whose leucine at residue 439 in the ligand-binding domain is replaced with cysteine (Volgraf et al. 2006). This cysteine is conjugated with a chemical called MAG (malaimide-azobenzene-glutamate). When UV is irradiated to iGluR6(L439C)-expressing cells in the presence of MAG, photoisomerization of the conjugated MAG induces activation of iGluR6(L439C). This induces depolarization of iGluR6(L439C)-expressing neurons. When the UAS:iGluR6(L439C) fish was crossed with an enhancer trap line expressing Gal4-VP16 in subsets of neurons including sensory neurons, the resulting larvae show defects in the touch response upon illumination of UV. Moreover, the touch response was recovered by illumination with 488 nm light. Thus, the GAL4-UAS:iGluR6(L439C) system can be used for controlling activity of neurons in vivo.
UAS:TeTxLC: targeted inhibition of neuronal function by tetanus toxin light chain
We established the UAS:TeTxLC line to interfere with neuronal function (Table 1) (Asakawa et al. 2008). Tetanus toxin light chain (TeTxLC) inhibits synaptic transmission by cleaving a synaptic vesicle associated membrane protein VAMP-2 to block neurotransmitter release (Schiavo et al. 1992). Its ectopic expression in the nervous system has been shown to cause behavioral abnormality in Drosophila and mice (Sweeney et al. 1995; Yamamoto et al. 2003).
We applied the Gal4FF-UAS:TeTxLC system to dissect specific neural circuits regulating behaviors. We first carried out enhancer trap and gene trap screens to collect transgenic lines expressing Gal4FF in diverse patterns (Fig. 3A). Then, we crossed those Gal4FF lines with the TeTxLC line and carried out the touch response assay on the resulting embryos (Fig. 4A). Distinct abnormal behaviors in the touch response were observed in different Gal4FF lines (Fig. 4B).
One line failed to show the touch response and displayed the TeTxLC expression in large populations of interneuons in the brain and spinal cord (Fig. 4B,C). Another line failed to respond to the touch and the TeTxLC expression was detected in the Rohon-Beard sensory neurons (Fig. 3B,D,E). In addition, one line showed an abnormal escape response and expressed TeTxLC mainly in the spinal interneurons. These observations demonstrated our Gal4FF trap system and the UAS:TeTxLC line can be used for investigating neural circuits regulating behavior.
The choice of Gal4 variants
It has been known that expression of a strong transcription activator can cause toxicity due to titration of endogenous transcription machineries. This phenomenon is called ‘squelching’ (Gill & Ptashne 1988). Therefore, a high level of Gal4 expression may be harmful to cells. In fact, 4.3% of Gal4-VP16 expression patterns created by the enhancer trapping showed lethality (Scott et al. 2007). This is probably due to a strong transcriptional activity of VP16 (Argenton et al. 1996; Koster & Fraser 2001). Except these expression patterns, the majority of the Gal4-VP16 lines and all of the Gal4FF lines did not show lethal effects (Scott et al. 2007; Asakawa et al. 2008). Thus, both Gal4-VP16 and Gal4FF may be used for enhancer trapping and gene trapping.
Basal (background) expression associated with current constructs
In our gene trap screen, a weak basal level of Gal4FF expression was observed in non-neural cells in the spinal cord, which is presumably associated with the sequence around the splice acceptor (Asakawa et al. 2008) (K. Asakawa and K. Kawakami, unpubl. data, 2005). In the enhancer trapping using the hsp70 promoter, the basal expression of Gal4 was observed in the skeletal muscle and heart (Scott et al. 2007; Asakawa et al. 2008). The background Gal4FF expression did not affect the touch response behavior (Asakawa et al. 2008). Development of a splice acceptor and a minimal promoter with little or no background expression will be important in the future.
Variegated (Mosaic) expression
Variegated expression of UAS-reporters has been observed in transgenic lines expressing Gal4-VP16 under the control of specific enhancer/promoters that were created by the meganuclease (I-SceI)-mediated transgenesis method or by microinjection of the plasmid DNA (Sagasti et al. 2005; Scott et al. 2007). In contrast, in the Gal4FF and Gal4-VP16 transgenic lines created by Tol2-mediated transgenesis, specific expression patterns of a reporter gene were observed consistently among siblings and through generations, and gross variegated expression was not detected (Davison et al. 2007; Scott et al. 2007; Asakawa et al. 2008). This indicates that the transposon-mediated enhancer trap and gene trap methods can abolish or reduce such effects. It should be noted that the basal (background) expression pattern, i.e. UAS:EGFP expression in the skeletal muscle, was still variegated in some of our Gal4FF enhancer trap lines. An enhancer trap construct that can eliminate such effects remains to be developed.
Position effect of UAS-transgenes
In the course of creating UAS lines, we found that different UAS lines show different levels of transgene expression even when crossed with the same Gal4 line, suggesting that the UAS-transgenes are subjected to position effect (Phelps & Brand 1998). For example, three independent UAS:TeTxLC effector lines, each of which contains a single integration of the UAS:TeTxLC construct, had different effects on behavior when crossed with a line expressing Gal4FF in specific neurons (Asakawa et al. 2008) (K. Asakawa and K. Kawakami, unpubl. data, 2006). Therefore, when a UAS transgenic line is created, it is crucial to test several different lines and select the best one for experiments. Development of a method for integrating a UAS-transgene in a defined chromosomal locus will help to obtain a strong and consistent expression of UAS-transgene.
This work was supported by postdoctoral fellowships from the Japan Society for the Promotion of Science (to K.A.), a Sasakawa Scientific Research Grant from the Japan Science Society (to K.A.), NIH/NIGMS R01 G069382 (to K.K.) and grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to K.K.).