Targeted genetics in Drosophila cell lines: Inserting single transgenes in vitro

ABSTRACT

A long-standing problem with analyzing transgene expression in tissue-culture cells is the variation caused by random integration of different copy numbers of transfected transgenes. In mammalian cells, single transgenes can be inserted by homologous recombination but this process is inefficient in Drosophila cells. To tackle this problem, our group, and the Cherbas group, used recombination-mediated cassette exchange (RMCE) to introduce single-copy transgenes into specific locations in the Drosophila genome. In both cases, φC31 was used to catalyze recombination between its target sequences attP in the genome, and attB flanking the donor sequence. We generated cell lines de novo with a single attP-flanked cassette for recombination, whereas, Cherbas et al. introduced a single attP-flanked cassette into existing cell lines. In both approaches, a 2-drug selection scheme was used to select for cells with a single copy of the donor sequence inserted by RMCE and against cells with random integration of multiple copies. Here we describe the general advantages of using RMCE to introduce genes into fly cells, the different attributes of the 2 methods, and how future work could make use of other recombinases and CRISPR/Cas9 genome editing to further enable genetic manipulation of Drosophila cells in vitro.

Introduction

Analysis of primary cultures and continuous cell lines has supported whole animal studies in Drosophila since the first primary cultures were developed in 1965 and the first cell lines were derived in 1969 (reviewed in¹). The uniform morphology and physiology of cells from cell lines have made them a useful system for biochemical and molecular analyzes, including, for example, the relationship between structure and the function of proteins and genome-wide RNAi screens (reviewed in²).

The most commonly used fly cells are the embryonic cell lines S2, Kc167, and their derivatives.^3,4 In addition to these, other cell lines have also been established from specific tissues such as the central nervous system,⁵ imaginal discs^6,7 the larval blood system⁸, and the ovary.⁹ Currently, there are 163 Drosophila melanogaster cell lines available in the collection maintained at the Drosophila Genome Resource Center (DGRC). These include some lines derived by our group using activated Ras (Ras^V12) to drive cell proliferation.^10-12 The method is efficient and also enables the generation of cell lines with specific genotypes.^11,13 Here we describe the use of our method to derive cells that harbor a single site for insertion of transgenes.¹⁴ Using a different approach, the Cherbas group engineered existing cell lines to permit insertion of a single transgene.¹⁵ Both methods use Recombination-Mediated Cassette Exchange (RMCE) for transgene insertion. Here we highlight this work and how it will enable more refined analysis in Drosophila cells.

Current methods are hampered by variation in transgene number and insertion site

Drosophila cultured cells take up exogenous DNA by transfection or directly, as in the case of S2 cells that are phagocytic.¹⁶ Typically the DNA is a plasmid carrying a transgene encoding the product of interest. After transfection, the DNA is expressed transiently for a few days and gene products can be analyzed during this time window. Alternatively, a stable line can be established in which the transgene is inserted into the genome. To generate stable cell lines, the transgene is transfected together with a selectable marker gene, which can be on a separate plasmid. Resistance genes developed for mammalian cell culture system also typically work in Drosophila cells.^17-21

In the stable lines, the transfected DNA is randomly integrated and there is a large variation in the number of transgenes that insert in a given cell. This variation in copy number stems from the formation of tandem arrays of the transfected DNA.¹⁶ These arrays can have abnormal chromatin structure that results in silencing of the transgene, as well as instability in the length of the array. Moreover, transgenes integrate randomly and are subject to different genomic contexts, which can also affect expression level.²²

Site-specific systems to introduce transgenes at defined sites

To control copy number and insertion site, homologous recombination has been used extensively in mammalian cells in culture and in whole animals.^23,24 In flies, homologous recombination is possible,^25,26 but occurs at low frequency and transgenes have been typically inserted using transposons, such as P elements, piggyBac transposons, and Minos elements.^27-29 These insert as a single copy with some bias to different regions in the genome.³⁰ Recently, site-specific integration using the viral recombinases, φC31 and Cre/loxP have become the methods of choice for transgene insertion in Drosophila (reviewed in ref.³¹). φC31 catalyzes the recombination between 2 heterotypic sites attP and attB, converting them into attL and attR sites, which are no longer substrates for the recombinase, thus making the event nonreversible. In a variation of this approach, φC31 can be used to promote recombination between a cassette in the genome flanked by attP sequences and a donor cassette flanked by attB sites. The donor cassette is inserted and the genomic cassette is eliminated resulting in so-called Recombination-Mediated Cassette Exchange (RMCE).³² Recombination mediated by φC31 is an efficient process that requires a small recognition site (< 40 bp for each attP or attB site) and allows the insertion of long donor sequences (at least 146 KB).^32-34 attP sites, including those amenable to RMCE, have been introduced into multiple sites in the genome using transposable elements.^35,36 φC31-mediated recombination, including RMCE, has been demonstrated in S2 cells,^33,37 but a robust selection system to purify only those cells that have undergone RMCE had not been developed until our work and that of the Cherbas group.^14,15

Generating cell lines with cassettes for RMCE in Drosophila cell culture

We developed 2 cell lines using the expression of UAS-Ras^V12 with Act5C-Gal4 that also harbor an attP-flanked cassette (Fig. 1A). The lines are called Ras-attP-L1 and L2. The cassette (attP. w⁺. attP), generated by Bateman et al., is in the taranis locus and has been used for transgene expression in flies (Fig. 1B).³² Cherbas et al. took an alternative approach and introduced an attP-flanked Act-5C>GFP cassette into Kc167 and Sg4 (an S2 derivative) cells using P-element transformation (Fig. 1D). The lines are called, for example, Kc167-PP-93E, which identifies the parental line and the insertion site (Fig. 1E). They generated multiple lines with a single cassette, which provide a variety of genomic sites for RMCE.

Figure 1.

Generation of cell lines with an attP-flanked cassette for insertion of transgenes by RMCE. A-C, Manivannan et al.; D-F, Cherbas et al. (A) Flies with an Act5C-Gal4 gene were crossed to flies with a UAS-Ras^V12 gene, which also have an attP-flanked cassette (attP.w⁺.attP). In the progeny, Act5C-Gal4 induces expression of UAS-Ras^V12. Primary cultures derived from these embryos gave rise to the Ras-attP-L1 and Ras-attP-L2 cell lines. (B) Schematic of the attP.w⁺.attP cassette at taranis (tara) locus. (C) In Ras-attP-L1 and -L2 cells, Act5C-Gal4 induces UAS-genes (DHFR, GFP, TK) and a transgene of interest (your favorite gene, UAS-yfg). Cells expressing DHFR are selected by MTX and cells expressing TK are killed by GCV. (D) P-element transformation was used to insert an attP-flanked cassette (attP-GFP-attP) into established cell lines, Sg4 (an S2 derivative) and Kc167. Clones of cells with a single insertion of the cassette were selected. Seventeen independent lines with different insertion sites were generated. (E) Schematic showing 2 different attP-flanked cassettes used. Top, cassette with eGFP flanked by attP sites (PP). Bottom, similar cassette with the addition of insulator sequences introduced as a part of the P-element (IPPI). (F) In the Sg4 and Kc cells, the DHFR and the TK genes are encoded by transgenes with an Act5C promoter. DHFR, dihydrofolate reductase; GFP, green fluorescent protein; MTX, Methotrexate; GCV, Ganciclovir; TK, Thymidine kinase.

There are advantages to both types of cell lines. The lines we developed are diploid for the major autosomes (second and third chromosomes) and heterozygous for the attP-cassette, which is on the third chromosome. There are variations in the number of the sex chromosomes and the small fourth chromosomes, but the near normal karyotype obviates uncertainty about the number of alleles for most genes. If an experiment requires information about the ratio of endogenous to transgene-encoded product, this could be an advantage over S2 and Kc cells, which have polyploid karyotypes and variable numbers of alleles.³⁸ Proliferation of the Ras-attP-L1 and -L2 cells depends upon expression of UAS-Ras^V12 driven by the Actin5C-Gal4 gene. The presence of the Actin5C-Gal4 gene in the cells allowed us to devise a selection scheme based on selectable markers driven by Gal4-induced transcription and also means that the cells can be used to express transgenes from the collections of UAS-regulated constructs available for many genes.^39,40 Cherbas et al. used Kc and Sg4 cells, which have the advantage of being well-characterized cell lines that have been widely used and for which the transcriptome has been determined.^38,41-46 The absence of the Act5C-Gal4 gene also means there is the flexibility to express genes with other Gal4 drivers.

Selection for RMCE events in Drosophila tissue culture cells

The key feature of RMCE is that only sequences within the attB-flanked cassette are inserted allowing counter selection for cells randomly inserting a whole vector, which has a toxicity factor in the backbone. Following precedents in mammalian systems, we and Cherbas et al., designed vectors that include a dihydrofolate reductase (DHFR) gene within the cassette. A single copy of DHFR provides resistance to methotrexate (MTX). We used UAS-DHFR, which is induced by Act5C-Gal4 already present in the cells (Fig. 1C), whereas, Cherbas et al. used DHFR directly fused to an Act5C promoter (Fig. 1F). Only cells that integrate the DHFR gene grow in the presence of MTX, however, cells with a random insertion of the donor plasmid are also resistant to MTX. To select against these cells, there is a thymidine kinase (TK) gene in the backbone of the donor plasmid. This gene is excluded only in cells with RMCE events (Fig. 2). Ganciclovir (GCV) is an antiviral drug that is converted into a toxic product by TK, correspondingly, any cell with an active TK gene (UAS-TK, in the Ras-attP-L1 and -L2 cells, and Act5C-TK in the Sg4 and Kc derivatives) will die (Fig. 2).

Figure 2.

Selection of cells with single transgenes inserted by RMCE. (A) Ras-attP-L1 and Ras-attP-L2 cells are transfected with an attB-donor plasmid encoding the selectable markers (DHFR and TK), GFP, and your favorite gene (yfg). The donor plasmid can insert by RMCE or non-specifically and cells with both types of inserts are GFP positive and resistant to MTX. Non-transfected cells (gray) die. In the presence of GCV only the cells that have undergone RMCE survive because the TK gene is excluded. A polyclonal line or colonies of cells with a single transgene are selected. One orientation is shown, but both are possible because the attP sites are indirect repeats. (B) Sg4 or Kc167 cells with the genomic cassette, which includes a GFP gene, are transfected with an attB-donor plasmid encoding the selectable markers (DHFR and TK), and yfg. The donor plasmid can insert by RMCE or non-specifically. Cells in which RMCE has occurred are GFP negative and resistant to MTX. Non-transfected cells and cells with non-specific insertions are GFP positive. The non-transfected cells die. In the presence of GCV cells that have undergone RMCE survive because the TK gene is excluded. However, some GFP positive cells, with non-specific insertions, survive and a subsequent FACS step is required to purify cells with a single transgene inserted by RMCE. Only a single orientation is possible because the attP sites are direct repeats.

Ras-attP-L1 and -L2 cells that have undergone RMCE can be selected simply by the 2-drug combination (Fig. 2A). In contrast, the Sg4-PP and Kc-PP cells do not express the TK gene at sufficiently high levels from the Act5C promoter and require fluorescence activated cell sorting (FACS) to derive a pure population (Fig. 2B). This difference in the level of TK expression is most likely because the Gal4-UAS system amplifies transcription of UAS-TK. The Sg4-PP and Kc-PP cells in which RMCE has occurred lose the GFP gene and can be sorted as GFP negative (Fig. 2B). After RMCE, each system provides comparable efficiency and requires about 6–8 weeks to purify a cell population with a single transgene insertion.

Using donor vectors for RMCE and expression of your favorite gene

The donor vector we developed has a multiple cloning site (pMCS-Donor) to allow insertion of transgenes either with a UAS-promoter or an independent promoter. We tested the vector with 3 genes (Tubulin, Jupiter, and Cas9) and were able to generate stable cells lines that expressed the proteins.¹⁴ Cherbas et al added a Gateway Cloning Cassette that will allow users to clone existing ORF collections into the vector.¹⁵

In the Ras-attP-L1 and L2 lines, the genomic cassette is flanked by indirect repeats and transgenes can be inserted in either orientation. We tested whether this influenced gene expression and found that levels of GFP were similar, as determined by western blot analysis, from a transgene inserted in either orientation.¹⁴ For reasons we do not understand, however, insertions occur predominantly in one orientation, so that the cell population is essentially comprised of cells with a single orientation.¹⁴ Moreover, the Ras-attP-L1 cells are strongly adherent and colonies resulting from single RMCE events can be cloned.¹⁴ In the Sg4 and Kc cells, the genomic cassette is flanked by direct repeats and transgenes can only insert in one orientation.¹⁵

RMCE and re-engineering Drosophila cells in culture

The major advantage of RMCE is that it allows the generation of stable lines expressing a single transgene from the same genomic location. This is ideal for structure-function relationship studies because it eliminates variation in expression levels between lines. With the 2 systems, researchers can choose either the more familiar Sg4/Kc cells or use the lines we developed that allow selection without FACS and have a more normal karyotype. All the lines and vectors are available at the DGRC.

The collection of lines that Cherbas et al., generated provide a range of possible insertion sites, however, these are not directed to particular sites within the genome. In subsequent work, using any cell line of choice, CRISPR/Cas9 could be used to insert an attP-flanked cassette at a particular site (and could simultaneously inactivate the target gene). RMCE using this entry point would allow reengineering of the site, as has been done in flies.^47-49 Currently, the selection schemes for insertion of the sequence by CRISPR/Cas9 are not very robust, but targeting multiple sites, followed by cloning and molecular characterization should allow for targeting of all alleles in a given cell line.^50-53

With the system we developed it is possible to derive cells of a particular genotype simply by constructing the line from flies with the desired alleles. We have developed a number of mutant cell lines demonstrating the feasibility of this approach.^{10,11,13,54,55} This may be useful for complex genotypes for which manipulation in the cells by CRISPR/Cas9 may be impractical. We also expect the system can be used to generate cells lines of a particular lineage by deriving lines from the cell-type specific expression of Ras^V12. Any such lines developed should include an attP-flanked cassette that will allow the subsequent introduction of transgenes by RMCE thus enabling, for example, the analysis of Huntington disease by comparison of lines expressing different mutant alleles in a single copy at the same site.

In closing, we expect that there will be increased interest in studying biological questions in Drosophila cell lines because RMCE and other recombination systems will enable more sophisticated and controlled somatic cell genetics.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.