A toolkit to generate inducible and interconvertible Drosophila transgenes

The existence of three independent binary systems for conditional gene expression (Gal4/UAS; LexA/LexAop; QF/QUAST) has greatly expanded versatile genetic analyses in the fruit fly Drosophila melanogaster; however, the experimental application of these tools is limited by the need to generate multiple collections of non-interchangeable transgenic fly strains for each inducible gene expression system. To address this practical limitation, we developed a modular vector that contains the regulatory elements from all three binary systems, enabling Gal4-, LexA- or QF-dependent expression of transgenes. Our methods also incorporate DNA elements that facilitate independent site-specific recombination and elimination of regulatory UAS, LexAop or QUAST modules with spatial and temporal control, thus offering unprecedented possibilities and logistical advantages for in vivo genetic modulation and efficient interconversion of overexpression transgenic fly lines.


MAIN
The inducible Gal4/UAS gene expression system revolutionized genetic experimentation in fruit flies 1 . This binary genetic tool facilitates gene expression in larval and adult fly tissues with cellular specificity and temporal control 1 . The Gal4/UAS system relies on the production of the yeast Gal4 transcription factor from gene-specific regulatory enhancers and adjacent promoters 1 . The Gal4 protein can bind to cognate DNA upstream activating sequences (UAS), thereby inducing transcription of any DNA sequence of interest inserted downstream of the UAS 1 . The Gal4/UAS system has been instrumental in deciphering many biological processes, and has inspired the development of analogous conditional binary gene expression tools such as the LexA/LexAop and QF/QUAST systems 2,3 . The use of these bipartite gene expression tools, singly or in combinations, has greatly broadened conditional genetics in fruit flies 4 . Currently, each of these gene expression tools requires the generation of independent transgenic fly lines, limiting experimental interchangeability and combinatorial usage, as well as generating cost inefficiencies related to the production and maintenance of multiple transgene collections.
To circumvent these limitations, we have developed a versatile modular vector (MV, hereafter) that incorporates the gene-activating regulatory sequences for three binary gene expression systems (5x lexAop / 5x UAS / 5x QUAS) upstream of a multi-cloning site (Fig.   1a). This design was predicted to enable the in vivo expression of an inserted DNA sequence under the regulation of the specific transactivator chosen. To demonstrate the flexible function of the MV, we generated a transgenic fly line encoding for a synthetic protein containing the small peptide HA fused to a fragment of a split venus fluorescent protein (VC fragment) 5 . We subsequently intercrossed the new MV-HA-VC fly strain with fly lines individually expressing different transcriptional activators . Each of these intercrosses generated F1 larval progeny and restricted HA-VC expression within Gal4, LexA and QFtranscribing cells, respectively ( Fig.1b-d). Furthermore, the use in parallel of two transcriptional drivers (dpp-lexA/Hh-QF; dpp-lexA/sal-G4) induced the concomitant expression of HA-VC within their cellular domains (Fig. 1e,f). Confirming the expression of our full-length HA-VC protein, immunostaining for HA and GFP largely colocalised in all of the cells expressing the MV construct (Supplementary Figure 1b). Similar results were obtained expressing a form of the initiator caspase Dronc tagged with V5 and VC peptides at the C-terminus (Supplementary Fig. 1c-e). In this case immunoprecipitation experiments followed by Western blotting confirmed the full-length expression of the Dronc-V5-VC (the expected size of Dronc-V5-VC is ~62KDa; Fig. 1g). These results demonstrated that the MV allows the expression of full-length cDNAs under the control of either separate or combined bipartite gene expression systems. We speculate that enhanced spatial and temporal precision will be achievable by incorporating Gal80 or QS, the respective repressors of Gal4/LexA or of QF 4 . These MV features can streamline genetic configurations for investigating gene function in cells with different origins, and for studying inter-organ cell communication.
Next, we explored the ability of our plasmid to generate loss-of-function phenotypes expressing short-hairpin RNA interference constructs. To this end we cloned short inverted repeats for targeting the gene wingless (flybase reference number = HMS00844) in the multicloning site of MV. Importantly, a TRIP line expressing this construct (HMS00844) was previously shown to downregulate the expression of Wg in vivo under the regulation of patch-Gal4 9 . However, our construct failed to induce noticeable wg downregulation or wg mutant phenotypes upon overexpression with several drivers (e.g. en-Gal4, Hh-QF and dpp-lexA), Similar unexpected results were obtained using UAS-Dicer to enhance the efficiency of the knockdown in combination with en-G4 and nubbin-Gal4, respectively. Although these preliminary experiments suggest that further optimization of our plasmid is likely needed to support robust RNA interference (e.g. expansion of the number of activation sequences), MV is still an excellent template for a wide range of biological applications requiring gene overexpression.
Drosophila genetics has benefited from three independent mechanisms of sequence specific genomic DNA recombination, especially the Flippase/FRT (Flp/FRT) system 10 . This system capitalises on Flp DNA recombinase activity that targets Flp-recognition sites (FRT sequences) 10 . Recently, directed mutagenesis of the 'wild type' Flp enzyme (hereafter, Flp WT ) and the FRT-recognition sites generated the mFlp 5 /mFRT 7.1 system 11 . Another alternative, the Cre/LoxP system, relies on Cre-mediated recombination of LoxP-recognition motifs 12 . Because no cross-reactivity has been reported between these three recombination systems 11,13 , gene regulatory elements in the MV were flanked by specific pairs of these recombination sequences (Fig. 2a) to investigate if sequence-specific excision of generegulatory sequences could be achieved. Specifically, the UAS sites were flanked by FRT (Flp sensitive), the QUAST repeats by loxP (Cre sensitive) and the LexAop site by mFRT 7.1 (mFlp 5 sensitive: Fig.2a). Upon expression of the cognate recombinase (Flp WT , Cre or mFlp 5 ), the specific regulatory components were successfully excised from the genome in somatic cells (Fig. 2a-c; Methods). This facilitates the termination of the overexpression (GFP negative clonal cell populations indicated by red arrowheads in the right panels of Fig. 2a-c) induced by each transcriptional activator with temporal precision and the generation of genetic mosaics within somatic tissues.
Next, we assessed whether gene cis-regulatory elements could be excised in the germline of a founder MV fly strain; if so, this would provide a means to efficiently generate derivative strains with permanently altered gene induction capacities. We intercrossed the MV-Dronc-V5-VC founder line with flies expressing Flp WT from a Beta2-tubulin promoter (BTub2, hereafter) ( Fig. 3a) to direct highly efficient FRT-dependent recombination in the male germline 14,15 . F1 male progeny containing both transgenes (MV-Dronc-V5-VC and Btub2-Flp WT ) were subsequently intercrossed with females harbouring both the sal-Gal4 and dpp-lexA driver transgenes (Fig. 3b). Remarkably, 95% (40/42 wing discs from 26 larvae) of the F1 (5 independent males were used to perform the experiment) generated larvae that did not express the Dronc-V5-VC construct in the sal-Gal4 domain but retained the activation in the dpp-lexA territory (Fig. 3c). These results demonstrated that most of the males expressing Flp WT in the germline permanently excised the UAS regulatory repeats without compromising the LexA-binding sites. Comparable results were obtained when Cre was expressed under the regulation of the Btub2 promoter ( Fig. 3b, 3c).
Unlike the results obtained when using Flp WT or Cre, we found that intercrosses using flies expressing Btub2-mFlp 5 only affected expression of the Dronc-V5-VC in approximately 2.82% of analysed larvae (2/70 wing discs from 35 larvae). These results are likely due to a low recombination frequency mediated by mFlp 5 in the male germline 15 , which could be alleviated by creating in the future new fly strains with enhanced Flp5 expression within male gametes (e.g. nanos-Gal4 UAS-Flp5). Despite this potential limitation, our findings confirm the suitability of the MV to efficiently obtain new fly strains starting from a single founder MVline through a simple intercross strategy for at least two of the gene regulatory elements included in the MV plasmid ( Fig. 3b). Moreover, the efficient germline excision of specific regulatory elements allows use of novel MV fly strains in combination with previously created lines in public repositories for multiplexed.
In summary, we have generated a next-generation modular vector that offers unprecedented experimental opportunities, whilst providing enhanced versatility to conduct targeted genetics in Drosophila. This expansion of the fruit fly genetic toolkit allows development of interconvertible fly collections of overexpression transgenes with a substantial reduction of logistical costs, including experimenter time and effort. Resulting lines retain specific gene regulatory elements of interest and can be rapidly interconverted through a simple scheme of intercrosses to allow altered inducible transgene regulation.

Molecular cloning
All PCRs were performed with Q5 High-Fidelity polymerase from New England Biolabs (NEB, M0492L). Standard subcloning protocols were used to generate all the DNA plasmids (see details below). Genomic DNA was extracted via standard protocols was used as a template to amplify different DNA fragments. Transgenic flies harbouring the new transgenes were obtained by attP/attB PhiC31-mediated integration. Some of the transgenes were generated by Bestgene Inc, whilst others were generated in house by Sangbin Park. Fly strains generated will be deposited at the Bloomington Stock Centre.
Whilst resources are transferred to Bloomington, reagents will be provided upon request.

Generation of the Module Vector (MV)
The different modules included in the vector were first designed in silico using the SnapGene software. Each module contains 5 repeats of each upstream regulatory sequence type, followed by an hsp70 minimal promoter. Each module was also flanked by specific recombination sites (see diagram Figure 1). This fragment was synthesized as a large DNA fragment by GENEWIZZ and subcloned in the PUC57 vector. The modular construct was then extracted from PUC57 as a SpeI-KpnI fragment and subcloned into a UASt-attB-w+ vector previously digested with NheI-KpnI. PmeI, XhoI, KpnI are potential unique restriction sites available for conventional cloning downstream of the gene regulatory elements (see diagram Figure 1). Sequence of the plasmid will be provided upon request until the vector is deposited in a public repository.

Generation of Beta2-Tubulin Flippase 5 and Cre transgenic flies
We extracted genomic DNA and amplified a Beta2-tubulin promoter via PCR using the following primers: Forward primer Beta2-tubulin promoter:

5' aaatttaatctgcaggcggccgcgaattcaagcttcgcccctttttcacaccg 3'
Convenient restriction sites for cloning were placed at the 5' (AvrII) and 3' (NotI and EcoRI) end of the PCR product. We then digested a UAS-attB-w+ vector with NheI and EcoRI, thus replacing the UAS repeats with the Beta2-tubulin promoter but keeping the rest of the plasmid backbone. The PCR was digested with AvrII and EcoRI and ligated with the aforementioned backbone, creating an intermediate plasmid  suitable to subclone the cDNA of the Flippase-5 and Cre. The Flippase5 cDNA was amplified via PCR from a plasmid vector kindly provided by Iris Salecker using the following primers: Forward primer Flippase-5:

5' AAtATAaaggcctTctagattatatgcgtctatttatgtagg 3'
The NotI and StuI restriction sites were conveniently placed in the PCR product of the Flippase-5 to facilitate the cloning in the Beta2-tubulin-attB plasmid. Before ligation, the Beta2-tubulin-attB plasmid and the Flippase-5 PCR product were digested with NotI and StuI.
To generate the Beta2-tubulin-CreV5-attB-w+ vector, we replaced the UAS repeats included in a UAS-Crev5-attB-w+ plasmids previously generated in the laboratory by the Beta2-Tubulin minimal promoter. To that end and prior ligation, we digested the UAS-CreV5-attB-w+ plasmid and the Beta2-Tubulin minimal promoter PCR product with NheI-EcoRI. Notice that the Cre cDNA was tagged with the V5 epitope.

MV-HA-VC
The VC fragment corresponding to the split Venus GFP was amplified by PCR with the primers indicated below. The HA epitope was incorporated in the N-terminus within the forward primer. Primers also contained suitable restriction sites to facilitate the subcloning.
The PCR product was first subcloned as a PmeI-XhoI fragment in an Actin SV40-polyA attB plasmid of the laboratory. Finally, the HA-VC fragment with the SV40 polyA was subcloned in the MV vector using the PmeI-SpeI restriction sites. Sequence of the plasmid will be provided upon request until the vector is deposited in a public repository.

MV-Dronc-V5-VC
The Dronc-V5-VC fragment was synthetised by Twist Bioscience. A cDNA wildtype of Dronc was fused to the V5 and VC peptides at the C-terminus. The constructs was subcloned in the MV vector using PmeI-KpnI restriction sites. Sequence of the plasmid will be provided upon request until the vector is deposited in a public repository.

MV-wg-RNAi
The construct was built using the primers indicated below. The primers included the inverted repeats for targeting the gene wingless previously described in the Drosophila Trip collection (HMS00844) (https://fgr.hms.harvard.edu/fly-in-vivo-rnai). The primers were synthetised and subsequently annealed using standard protocols. Then the resulting double strand DNA was subcloned in the MV vector previously opened with PmeI-KpnI restriction sites. After incubation for 2h with the secondary antibodies, samples were washed in PBT several times and mounted in Vectashield.

Immunoprecipitation and Western Blot
Twenty larvae of each respective genotype were snap-frozen at −80°C, thawed, macerated on ice in 300 μ l of RIPA buffer (complemented with protease inhibitor). Then the sample was cleared at 20.000 g for 20 min and supernatant toped-up with PBS + 0.3% Triton X-100 to 500 μ l. Next the sample was incubated with 15 μ l of V5 magnetic beads (chromotek) pre-washed three times in PBS + 0.3% Triton X-100 head-over-head at 4°C on a spinning wheel for 90 min. Beads were washed three times in PBS + 0.3% Triton X-100. Bound material was eluted from the beads in 40 μ l Laemmli buffer and loaded on a 4-12% gradient gel. Dronc-V5-VC was detected after completing the Western blotting with goat anti GFP antibody (Abcam, 1:2500).

Imaging of wing discs
Confocal imaging of wing imaginal discs was performed using the Olympus Fluoview FV1200 and the associated software. 55 focal planes were taken per wing disc using the 40x lens. Acquired images were processed using automated Fiji/ImageJ. Generally, Zstacks were projected, and the channels split. The Reslice function was used to obtain transversal sections of the images. Figures were produced using Adobe photoshop CC2018.

DATA AVAILABILITY
All data are incorporated into the article and its online supplementary material. All of the experimental resources generated in this manuscript will be publicly available through different public repositories upon publication and until then will be freely distributed upon reasonable request to the corresponding author. The full sequence of the MV plasmid will be accessible in the Supplementary Text of the manuscript. The plasmids described in the manuscript will be sent to a public repository (e.g. Drosophila Genome Resource Repository; https://dgrc.bio.indiana.edu/Home) upon publication. The new Drosophila strains described in the manuscript will be submitted to the Bloomington Drosophila Stock Center (https://bdsc.indiana.edu/index.html) upon publication.

ACKNOWLEDGEMENTS
We are grateful to the following investigators for providing flies and reagents: Drs. I.
Salecker (for kindly providing all the reagents and flies used in the study related with mFlippase-5, and hs-Cre line on the second chromosome), K. Basler (for providing the ( 1 9 9 6 ) .