Introduction

The zebrafish (Danio rerio) has emerged as a powerful model organism and is increasingly used in various research fields and disciplines. Zebrafish share 70% of their genome and more than 80% of disease genes with humans (Howe et al. 2013). This has led to the development of many distinct zebrafish disease models (Santoriello and Zon 2012) including models of hematopoietic and neurodegenerative diseases, cardiovascular diseases (Bowley et al. 2022) such as atherosclerosis (Tang et al. 2021), cancer (Liu and Leach 2011; Russo et al. 2022), and tuberculosis (Bouz and Al Hasawi 2018; Varela and Meijer 2022), among others. Transgenic animals continue to play a key role for many experimental strategies, including study of gene function, genetic screens, and development of disease models. Transgenic zebrafish lines have been widely generated by applying the Tol2 transposon system with a transgenesis rate of 50% (Kawakami et al. 2004). In this system, a transposon-donor plasmid containing a DNA sequence of interest flanked by Tol2 sites and synthetic mRNA encoding the Tol2 transposase are introduced into the single cell of fertilized eggs by microinjection. Upon translation, the transposase protein integrates the Tol2 flanked gene fragment randomly into the genome through a cut-and-paste mechanism without causing rearrangement or modifications at the target site (Kawakami and Shima 1999; Kawakami, Shima, and Kawakami 2000; Kawakami et al. 2004; Kawakami 2007).

A number of approaches have previously been used to generate Tol2 transgenesis constructs. Traditionally, multi-step restriction enzyme and ligase based subcloning was used. However, this has given way to many advanced cloning approaches, including Gateway cloning (Hartley, Temple, and Brasch 2000; Kwan et al. 2007; Katzen 2007), Gibson Assembly (Gibson et al. 2009), In-Fusion cloning (Zhu et al. 2007; Sleight et al. 2010) and Golden Gate cloning, which have been used to speed up and simplify the procedure of transgene assembly. Each of these approaches have different advantages and disadvantages. Gateway cloning allows for the establishment and reuse of a library of fragments (e.g. promoters, genes and fluorescent proteins) but requires the use of lambda phage-derived att sequences to facilitate insertion of the desired sequences. These att sequences result in the addition of ∼20-25 bp of additional sequence in between each element and expansion of the number of fragments to be assembled can require recloning of the sequence with new att ends (Katzen 2007; Engler, Kandzia, and Marillonnet 2008; Sone et al. 2008). By contrast, Gibson and In-Fusion cloning allow for joining of short (15-20 bp) regions of homology, enabling the seamless cloning of many transgenes. However, the requirement for homology for seamless cloning generally means that individual components have to be adapted to upstream and downstream elements to generate new transgenes and have an upper limit of approximately 5-6 fragments with currently available kits. Golden Gate cloning utilizes type IIS restriction enzymes to cleave and join multiple fragments in a one pot reaction (Padgett and Sorge 1996; Engler, Kandzia, and Marillonnet 2008). It balances the tradeoffs of the previous cloning systems, requiring inclusion of only a short 4 base pair sequence, while enabling researchers to generate a library of vectors that can be combinatorially assembled to generate distinct transgenes. Large numbers of DNA fragments can also be assembled by Golden Gate with reports of assembly of up to 52 fragments in a single reaction (Pryor et al. 2022). Despite the advantages of Golden Gate cloning, this approach has not been widely applied in zebrafish transgene assembly. Only a few examples of this approach have been described for Tol2 transgenesis including the recently described GoldenFish toolkit which uses Golden Gate cloning to gain more flexibility in transgene design (Jiang et al. 2023) and the Golden Gateway tool (Kirchmaier, Lust, and Wittbrodt 2013) which combines the well-established cloning tools Gateway and Golden Gate cloning to simplify creation of new components for Gateway assembly of transgenes.

Here, we have sought to apply Golden Gate cloning approaches to create a toolkit for zebrafish transgenesis. We have developed ImPaqT, a new Immunological toolkit for PaqCI-based Golden Gate Assembly of Tol2 Transgenes, to facilitate the production of transgenic animals. Studies of immune function have been of growing interest in the zebrafish community due to the close genetic similarity of the immune systems of humans and zebrafish (Howe et al. 2013), but there are few compiled resources for these cell types. ImPaqT provides a collection of commonly used immunological promoters targeting distinct lineages of vertebrate immunity, genetic tools and fluorophores supporting flexible generation of transgenes for immunological research. This toolkit utilizes Golden Gate cloning due to the advantages of this technique and incorporates a newly described type IIS restriction enzyme, PaqCI, with a 7 base recognition sequence to reduce the frequency of off target cutting (from 1 in 4,096 bases (4⁶) for a 6-base cutter to 1 in 16,384 bases (4⁷), assuming random cutting). We designed the kit to have consistent overhang sequences for each position in the backbone (e.g. 5’, middle, and 3’ element) in order to make the toolkit modular and simplify the construction of additional transgenes. As proof-of-concept, we show that this toolkit can facilitate the production of expression constructs and we demonstrate the generation of transgenic animals using the Tol2 transgenesis with these constructs. ImPaqT is modular, enabling easy swapping of gene sequences, easily expandable, allowing the introduction of multiple new genetic elements if needed, and relies only on the use of short (4 bp) overlaps that can also be designed to be scarless if desired. To establish an immunological toolkit that will be broadly useful to zebrafish researchers, we have generated a variety of immunological and non-immunological promoters, including mfap4 (macrophages), irg1 (macrophages, induced by infection), lyz (neutrophils), lck (T cells), ubb (ubiquitious) and hsp70 (heat shock inducible), tools for genetic recombination (icre), cell killing (kid, nitroreductase) and cell isolation tools (LNGFR), and an array of fluorescence proteins including mTurquoise2, tdStayGold, mCitrine, mNeonGreen, dLanYFP and tdTomato that can be used to rapidly develop new transgenes and new transgenic lines.

Results

PaqCI-based Golden Gate cloning enables the construction of Tol2 expression vectors

We sought to create a toolkit for immune transgene construction for the zebrafish community. Ultimately, we decided to design this toolkit to use Golden Gate cloning, given the short regions of homology that this approach required, the ability to structure this toolkit in a modular fashion and the expandability of the Golden Gate cloning approach, enabling incorporation of additional genetic elements as needed. Golden Gate assembly uses type IIS restriction enzymes to generate a set of unique overhangs that are used to join fragments in a desired order. To create a system for modular cloning of transgenes, we chose consistent 4-base overhang sequences for each of the three elements such that DNA fragments amplified with these overhangs can be cloned in the same position in the final construct, allowing these elements to be used combinatorially. As zebrafish transgenes generally consist of a cell-type specific promoter, a gene of interest, and a fluorescent marker, we structured our toolkit to have three elements that would ultimately be joined within a backbone vector possessing Tol2 repeats for transgenesis. These elements were named based on their location within the final construct - 5’ element (5E) for the first position, a middle element (ME) and a 3’ element (3E) for the final position within the backbone.

To generate the overhangs required to join individual components for Golden Gate assembly, we chose the 7-base cutter PaqCI to limit off-target cutting of inserts used in assembly. Genes and regions of interest were generated by PCR amplification using primers containing the 7-base PaqCI recognition sequence along with a spacer consisting of any four nucleotides that will be excised from the final construct, and a 4-base overhang sequence specific to the location of the sequence of interest in the final construct (Fig. S1). These PCR products were cloned into vectors and sequence verified to generate a stable library of constructs for transgene assembly. Cleavage with PaqCI leads to the creation of specific 4-base overhangs for each position at their 5’ and 3’ end enabling proper assembly and single-step ligation into a backbone vector for transgene generation. These backbone vectors possess inverted Tol2 repeats for transgenesis surrounding a PaqCI-flanked LacZ gene for blue-white screening to identify backbone constructs containing transgene components. A simplified schematic for the assembly using one insert is illustrated in Fig. 1. As we demonstrated below, we have found that this approach allows for efficient, combinatorial transgene generation.

Figure 1.

ImPaqT facilitates the flexibility of transgenes design and the generation of transgenic zebrafish

Having established this Golden Gate cloning-based technique for the assembly of Tol2 transposon vectors, we sought to create an array of constructs that would be beneficial to the immunology community to generate transgenic animals. To these ends, we have generated constructs for expression in commonly studied zebrafish immune populations, including T cells (lck) (Langenau et al. 2004), neutrophils (lyz) (Hall et al. 2007) and macrophages (mfap4 and the infection-inducible irg1) (Walton et al. 2015; Hall et al. 2013). To enable lineage tracing approaches for labeling of hematopoietic derived cells, we also included the hematopoietic stem cell (HSC)-specific promoter runx1 +23 (Tamplin et al. 2015) as well as tools for broader expression including the ubiquitious promoter ubiquitin B (ubb) (Mosimann et al. 2011), and the temperature inducible heat shock protein 70-like promoter (hsp70) (Shoji, Yee, and Kuwada 1998; Halloran et al. 2000; Kwan et al. 2007). We have also integrated a number of tools to allow for manipulation of these immune populations, including tools for genetic manipulation (icre (Shimshek et al. 2002)), for rapid purification of cell populations (truncated human LNGFR (Wattrus and Zon 2020)) and cell-specific ablation (nitroreductase (Curado et al. 2007; Curado, Stainier, and Anderson 2008) and kid (Labbaf et al. 2022)). To visualize labeled cell populations and genes, we provide a diverse array of best-in-class fluorescent proteins in a range of colors to facilitate simultaneous imaging of distinct cell types. These include tdStayGold (Hirano et al. 2022), mTurquoise2 (Goedhart et al. 2012), tdTomato (Shaner et al. 2004), mCitrine (Zacharias et al. 2002), mNeonGreen (Shaner et al. 2013) and dLanYFP (Shaner et al. 2013). A summary of all available constructs is shown in Table 1. ImPaqT is also designed to be easily expandable through simple cloning approaches.

Table 1.

Generation of new transgenic zebrafish lines using ImPaqT

To test our PaqCI-based Golden Gate assembly toolkit as well as generate new immunological zebrafish lines for the community, we used ImPaqT to assemble a number of new transgenic lines. Using the neutrophil-specific promoter lyz, we generated a Tg(lyz:mTurquoise2)^ber2 construct through Golden Gate assembly of 5E lyz, ME mTurquoise2 and 3E ubb:polyA to create the lyz:mTurquoise2 transgene construct (Fig. 2A-B). The construct was coinjected with Tol2 transposase mRNA into single cell of fertilized eggs and at 2 dpi we observed strong blue fluorescence signal within likely neutrophil populations in the CHT (Fig. 2C). Positive animals were raised to adulthood and screening of F1 progeny of these animals revealed robust expression of mTurquoise2 within putative neutrophil populations throughout the animal (Fig. 2D). Crossing to the already established neutrophil reporter Tg(lyz:EGFP)^nz117 (Hall et al. 2007) zebrafish confirmed specific expression of our transgene within neutrophils (Fig. S2).

Figure 2.

Similarly, we used the ImPaqT toolkit to generate a new macrophage-specific transgene using the infection-inducible macrophage-specific promoter irg1 and the recently described bright green fluorescent protein tdStayGold (Hirano et al., 2022). To generate the Tg(irg1:tdStayGold)^ber3 construct, the 5E irg1, ME tdStayGold and 3E ubb:polyA elements were combined into the backbone vector (Fig. 3 A-B). We injected zebrafish with this construct and as irg1 is an inflammation responsive promoter that is only weakly expressed in the absence of immune stimulation, we injected larvae with LPS before screening. After LPS injection, we were able to identify a number of F0 founders with strong green fluorescence within subpopulations of their macrophages (Fig. 3C). These positive F0 larvae were raised to adulthood and screening of F1 progeny revealed expression of tdStayGold throughout macrophage populations upon stimulation with LPS (Fig. 3 D).

Figure 3.

We also aimed to test the utility of the tools that we have designed within our kit. To these ends, we created an infection-inducible, macrophage specific icre transgene construct, Tg(irg1:icre-p2A mTurquoise2). This construct was generated by combining the 5E irg1, ME icre and 3E p2A mTurquoise2 (Fig. 4A and 4B). Animals injected with this transgene showed macrophage-specific mTurquoise2 signal after injection with LPS (Fig. 4C). To further confirm that this construct was able to drive rearrangement of floxed reporters, we injected Tg(−3.5ubb:LOXP-EGFP-LOXP-mCherry)^cz1701 (Mosimann et al. 2011) zebrafish with this construct. These animals were injected with LPS and imaged for 48 h to visualize rearrangement of the floxed construct. LPS treatment drove the induction of icre and the co-translationally cleaved p2a mTurquoise2 transgene within macrophages starting ca. 4 h post LPS injection. While the induction of the irg1:icre-p2a mTurquoise2 transgene was transitory, this short-term induction of Cre could induce permanent rearrangement of the loxP-EGFP-loxP-mCherry construct as seen by induction of red fluorescent mCherry within macrophage populations (Fig. 4D, S3, Movie S1). Control animals without the icre-construct neither showed mTurquoise2 nor mCherry signal in macrophages (Fig. S3, Movie S 2). This experiment showed not only the functionality of the individual insertion constructs but also the range of assembly constructs that can be generated with them.

Figure 4.

ImPaqT is expandable through the introduction of new overhang sequences

A limitation of many existing systems for transgenesis is that these systems have a fixed number of elements that can be combined to generate transgenes. Golden Gate methods however, are expandable through the addition of new overhangs and have been used to join 10-24 or even more than 50 fragments of DNA in a single reaction (Kucera 2021; Pryor et al. 2022). The ability to expand this kit to include larger numbers of constructs would simplify the generation of complex transgenes including multifragment promoters, loxP sites, localization tags (e.g NLS or prenylation signal) or multigene constructs for example. To provide proof of concept for the expandability of ImPaqT, vectors were designed to enable the construction of a ubb promoter-driven construct from three similar sized ubb promoter fragments, ubb₁, ubb₂ and ubb₃. These fragments were constructed such that the first fragment (ubb₁) and the last fragment (ubb₃) use the same O1 and O2 overhangs used by 5E constructs so that the array of promoter fragments can be cloned into the existing position of 5E elements and used alongside already existing ME and 3E constructs (Fig. 5A-B). New internal overhangs were designed for the ubb₁, ubb₂ and ubb₃ fragments to combine these distinct fragments (Fig. 5A-D). Individual ubb fragments were used for Golden Gate assembly in combination with ME mTurquoise2 and 3E ubb:polyA (Fig. 5C) which led to the expected transgenesis construct Tg(ubb:mTurquoise2)^ber4 (Fig. 5D) that was injected into embryos at single cell stage. Screening of injected animals confirmed that transgenic embryos established with this construct had the expected expression pattern (Fig. 5E). We also found robust expression in the F1 progeny (Fig. 5F) after raising positive F0 animals, confirming that injected animals are viable and able to pass on the transgene to the next generation. Together, these findings demonstrated proof-of-concept for the expansion of this transgene assembly system enabling users to add fragments as needed through the addition of novel unique 4-base overhangs while maintaining compatibility with existing ImPaqT constructs.

Figure 5.

Discussion

Transgenic animals remain essential for many research questions. They enable the visualization of cells and tissues, molecular pathways, immune responses and the perturbation of these pathways to identify the molecular mechanisms driving pathology. Use of the zebrafish as a model organism has been steadily increasing in many research disciplines due to the unique advantages of this model. This is true in immunology and host-pathogen interactions, where the optical transparency afforded by the zebrafish model have led to many research discoveries, but where, as a new immunology model, there are limited transgenic tools. Here, we demonstrate a new cloning toolkit for the generation of immunological transgenes that is based on Golden Gate cloning. The Golden Gate approach results in shorter overlap regions (4 bp) than recombination-based techniques like Gateway cloning, while compared to seamless, homology based cloning methods such as Gibson cloning, it allows for straightforward modular cloning approaches. The Golden Gate approach also takes advantage of off-the-shelf restriction enzymes and ligase, compared to other cloning approaches which use proprietary enzymes for cloning.

We have termed this toolkit ImPaqT - Immunological toolkit for PaqCI-based Golden Gate Assembly of Tol2 Transgenes. It includes a range of immunologically focused promoters along with genetic tools and fluorescent proteins that can be assembled using PaqCI-based Golden Gate cloning. These tools will enable the flexible assembly of transgenes to drive expression and to genetically manipulate, purify and ablate immune cell populations of interest. While zebrafish possess the same immune lineages of other vertebrates, as a relatively new model for immunology studies, the zebrafish has a more limited array of cell type-specific promoters than mouse models. For instance, despite the identification of potential NK- and ILC-specific (Hernandez et al. 2018) as well as mast cell-specific (Dobson et al. 2008) genes, promoters specific for these cell populations have not been described. As new immune-specific promoters are identified, these can be readily added to this toolkit. While we have targeted the immune system as a particular area of need, this Golden Gate approach can also be adapted to any other tissue or cell lineages within the zebrafish.

While Golden Gate cloning enables the use of shorter homology stretches and the combinatorial use of constructs for the assembly of diverse transgenes, as a restriction enzyme-based system, a limitation of Golden Gate systems, compared to other assembly approaches, has been the risk of off-target cutting within the cloned sequences. This off-target cutting requires mutation of insert sequences to eliminate these sites (termed domestication), (Marillonnet and Grutzner 2020; Bird, Marles-Wright, and Giachino 2022; Grützner and Marillonnet 2020), which can potentially alter function of domesticated promoters and genes. While domestication can result in changes in the sequence that may alter behavior of the genetic element, these in principle can be managed for genes of interest using silent mutations in critical bases. While promoters cannot be managed through the same approach, promoter fragments used for zebrafish are frequently large (commonly 2-6 kb) and it is unlikely that the mutation of a single base pair will inadvertently ablate the promoter activity. Consistent with this, we have not seen any changes in promoter activity or localization in the two promoters that we domesticated for the ImPaqT toolkit, although further cloning of additional promoters will more completely test this idea. Our choice of PaqCI itself seeks to address the limitations of domestication, as PaqCI uses a 7 base recognition site. In contrast to the 6 base-cutters that have been described in previous zebrafish Golden Gate approaches (Kirchmaier, Lust, and Wittbrodt 2013; Jiang et al. 2023), the rare cutting of PaqCI (if random, 1 in 16,384 bases) makes this enzyme better suited for the large DNA fragments that are frequently used for zebrafish promoters.

A limitation of existing systems has been the use of a fixed number of constructs for assembly of the transgene. One benefit of the Golden Gate directed system is that it can be expanded for cloning of complex transgenes, with some Golden Gate assemblies encompassing 20 or even more than 50 fragments (Kucera 2021; Pryor et al. 2022). While we have designed the system to incorporate 4 distinct DNA segments (3 inserts and a backbone vector) the flexibility of Golden Gate cloning means that our system can be readily expanded and contracted while maintaining compatibility with the other components of the kit. As proof-of-concept, we were able to break a single 5E encoding the ubb promoter into 3 distinct plasmids that we could subsequently assemble together with existing ME, 3E and backbone elements to generate a functional ubb driven transgene. The flexibility of this system will simplify the construction of large and complex transgenes. Potential applications would include breaking up larger promoter fragments into smaller and easier to clone elements, insertion of recombination sites for enzymes like Cre or Flp for rearrangement, assembly of multipromoter constructs encoding genes such as Cas9 and CRISPR gRNAs for cell type-specific knockout, or the inclusion of epitope or localization tags.

Lastly, while we have applied this Golden Gate method for the construction of Tol2 transgenes, the advantages of the PaqCI-Golden Gate method will likely be useful for the editing of zebrafish by other approaches as well. For instance, CRISPR insertion has been increasingly used to generate knock-in zebrafish lines, via either homologous recombination, non-homologous methods or by using short homology stretches. Knock-in approaches require the construction of donor vectors which incorporate varied sequences such as homology regions, fluorescent proteins and stretches of genomic DNA (Albadri, Del Bene, and Revenu 2017). Similarly, many groups also use I-SceI insertion approaches to generate transgenic zebrafish (Hoshijima, Jurynec, and Grunwald 2016). The PaqCI Golden Gate method could be adapted to simplify the construction of transgenes and editing constructs for these other approaches as well.

Materials and Methods

Ethics statement

Zebrafish husbandry and experiments were approved by the “Landesamt für Gesundheit und Soziales” (LaGeSo) Berlin (Germany).

Zebrafish husbandry and handling

Zebrafish (Danio rerio) embryos were raised in petri dishes containing E3 medium at 28°C for up to five days. Animals older than 5 days were maintained in a circulating system with constant water conditions of 28°C, pH 7.0 - 7.3, and a conductivity of 600-700 µS on a 12-hour light/ 12-hour dark cycle. They were fed twice a day with artemia and once a day with dry food.

Generation of transgenic zebrafish lines

Zebrafish strains

All zebrafish strains, including Tg(lyz:mTurquoise2)^ber2, Tg(irg1:tdStayGold)^ber3,Tg(irg1:icre-p2A mTurquoise2) and Tg(ubb:mTurquoise2)^ber4, were in the ^*AB wildtype background and generated in this study by applying the Tol2 transgenesis system as previously described (Kawakami et al., 2004). Tg(−3.5ubb:LOXP-EGFP-LOXP-mCherry)^cz1701 zebrafish were obtained from the European Zebrafish Resource Center (EZRC). Tg(−3.5ubb:LOXP-EGFP-LOXP-mCherry)^cz1701 (Mosimann et al. 2011) and Tg(lyz:EGFP)^nz117 (Hall et al. 2007) have been previously described.

Preparation of the backbone vector and inserts

The backbone vector containing the LacZ gene flanked by PaqCI sites and Tol2 arms in a vector backbone with a pMB1 origin and an AmpR gene for bacterial selection (pTWIST Amp High Copy) was purchased from Twist Bioscience. The complete sequence is shown in Figure S4. Primer sequences are shown in Table S1. Gene fragments for runx1+23 (with the endogenous PaqCI site mutated), tdStayGold and truncated human LNGFR (zebrafish codon optimized and with endogenous PaqCI sites mutated) were purchased from Twist Bioscience and TOPO cloned into pCRII TOPO (Thermofisher Zero Blunt TOPO). The complete sequences are shown in Figure S5.

PaqCI recognition sites with 4-base spacer sequences and 4-base overhangs were added to DNA sequences via PCR amplification. Primer were designed to add PaqCI sites such as cleavage resulted in a gene fragment with a 4-base overhang that is compatible to the overhang of another gene fragment that will be adjacent to it in the final assembly. Appropriate 4-base overhang sequences were confirmed using the online tool NEBridge Ligase Fidelity Viewer^™. PCR amplification was performed using Q5 polymerase and analyzed by gel electrophoresis. The PCR product was cleaned up or, in case of multiple bands, the band of correct size was excised under 70% UV light using a clean scalpel and DNA was extracted using Nucleo Spin Gel and PCR Clean-up Kit (Macherey-Nagel). DNA sequences were confirmed by sequencing. Gel purified gene fragments were subcloned into a Zero Blunt TOPO vector (Invitrogen) and transformed into competent NEB 5α E. coli cells (C2987H). Cells were cultivated in LB medium containing 50 µg/ml Kanamycin and plasmids were isolated using the Nucleo Spin Plasmid Quick Pure protocol (Macherey-Nagel).

Both the ubb and lck promoters had endogenous PaqCI cleavage sites in the promoter sequence. To domesticate these promoters, we mutated these PaqCI sites with the indicated PCR primers (Table S1) and assembled them with In-Fusion cloning (Takara Bio).

Golden Gate based assembly of expression constructs

Golden Gate assembly was performed according to the NEB protocol. In detail, 75 ng of the backbone vector and each insertion construct was mixed with 10 U PaqCI, 0.5 µl PaqCI activator (20 µM) and 400 U T4 DNA ligase in T4 DNA ligase buffer to a final volume of 20 µl. Final assembly product was obtained after 60 Cycles of incubation at 37°C for 1 min and incubation at 16°C for 1 min, followed by 5 min incubation at 37°C and 5 min incubation at 60°C using a thermocycler. The final product was transformed into competent NEB 5α E. coli cells (C2987H) and cells were cultivated in LB medium containing 20 µg/ml Ticarcillin. Plasmids were isolated using the Nucleo Spin Plasmid Quick Pure protocol (Macherey-Nagel) and PCR purified using the Nucleo Spin Gel and PCR Clean-up Kit (Macherey-Nagel) before microinjection.

Microinjections

The injection mixture contained 25 ng/µL polyadenylated mRNA encoding the Tol2 transposase, 50 ng/µL transposon donor plasmid and 10% phenol red in a final volume of 10 µl 1 X Tango Buffer (Thermo Fisher). The mMessage mMachine kit (Life Technologies) was used to transcribe the mRNA from T3TS-Tol2 (Balciunas et al. 2006). A volume of ∼1 nL was injected into the single cell of fertilized eggs by using a microinjector.

Imaging

Injected larvae were screened for insertion of the transgene 2 days post injection (dpi) using a Nikon ECLIPSE Ti2 microscope (Nikon) with a Dragonfly 200 confocal system (Oxford Instruments). Animals that were positive for the transgene were raised to adulthood and outcrossed to the ^*AB wildtype strain for selection of transgenic founder zebrafish. Before screening Tg(irg1:icre-p2A mTurquoise2) and Tg(irg1:tdStayGold) embryos, the activity of the irg1 promoter was induced by injection of LPS. For that purpose, animals were anesthetized with tricaine (MS-222) at a final concentration of 0.016% and injected into the caudal vein with LPS (Sigma-Aldrich, F3665) (final: 0.8 mg/ml). Screening was performed at least 4 h after injection. To reduce pigmentation, Tg(−3.5ubb:LOXP-EGFP-LOXP-mCherry) embryos were treated with 45 µg/ml 1-phenyl-2-thiourea (PTU) starting ca. 20 h post fertilization. Image analysis was performed using the Fiji ImageJ software (National Institute of Health, USA).

Line and reagent availability

All new lines and plasmids are available from the corresponding author upon contact and will also be deposited to AddGene.

Data availability

All relevant data can be found within the article and its supplementary information.

Acknowledgements

We would like to thank Miriam Herbert for sharing the ME kid construct. Additionally, we would like to thank our animal welfare officer Antje Dreyer and our animal caretakers Silke Lehmann, Marten Walk, Carsten Weiland, Ines Neumann, Jens-Phillip Otto und Janine Bleske for their great work and support.

Additional information

Funding

This work was supported by the Max Planck Society.

Additional files

Supplementary figures and tables

Movie 1

Movie 2