1. Genetics and Genomics
Download icon

An efficient CRISPR-based strategy to insert small and large fragments of DNA using short homology arms

  1. Oguz Kanca
  2. Jonathan Zirin
  3. Jorge Garcia-Marques
  4. Shannon Marie Knight
  5. Donghui Yang-Zhou
  6. Gabriel Amador
  7. Hyunglok Chung
  8. Zhongyuan Zuo
  9. Liwen Ma
  10. Yuchun He
  11. Wen-Wen Lin
  12. Ying Fang
  13. Ming Ge
  14. Shinya Yamamoto
  15. Karen L Schulze
  16. Yanhui Hu
  17. Allan C Spradling
  18. Stephanie E Mohr
  19. Norbert Perrimon
  20. Hugo J Bellen  Is a corresponding author
  1. Baylor College of Medicine, United States
  2. Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital, United States
  3. Howard Hughes Medical Institute, Harvard Medical School, United States
  4. Harvard Medical School, United States
  5. Howard Hughes Medical Institute, United States
  6. Howard Hughes Medical Institute, Baylor College of Medicine, United States
  7. Howard Hughes Medical Institute, Carnegie Institution for Science, United States
Research Advance
  • Cited 10
  • Views 8,857
  • Annotations
Cite this article as: eLife 2019;8:e51539 doi: 10.7554/eLife.51539

Abstract

We previously reported a CRISPR-mediated knock-in strategy into introns of Drosophila genes, generating an attP-FRT-SA-T2A-GAL4-polyA-3XP3-EGFP-FRT-attP transgenic library for multiple uses (Lee et al., 2018a). The method relied on double stranded DNA (dsDNA) homology donors with ~1 kb homology arms. Here, we describe three new simpler ways to edit genes in flies. We create single stranded DNA (ssDNA) donors using PCR and add 100 nt of homology on each side of an integration cassette, followed by enzymatic removal of one strand. Using this method, we generated GFP-tagged proteins that mark organelles in S2 cells. We then describe two dsDNA methods using cheap synthesized donors flanked by 100 nt homology arms and gRNA target sites cloned into a plasmid. Upon injection, donor DNA (1 to 5 kb) is released from the plasmid by Cas9. The cassette integrates efficiently and precisely in vivo. The approach is fast, cheap, and scalable.

https://doi.org/10.7554/eLife.51539.001

Introduction

A main goal of the Drosophila Gene Disruption Project (GDP) is to create genetic tools that facilitate an integrated approach to analyze the function of each gene in detail. This involves assessment of the loss of function phenotype, identification of the cells that express the gene, determination of the subcellular protein localization, selective removal of the transcript or protein in any tissue, the ability to perform immunoprecipitation of the protein and its interacting proteins or DNA, rescue of the induced fly mutant phenotypes with fly or human cDNAs and assessment of the consequences of amino acid variants in vivo.

These elegant and precise manipulations are made possible by the integration of a Swappable Integration Cassette (SIC) in the gene of interest (GOI) using transposon mediated integration (Minos-mediated Integration Cassette [MiMIC]; Venken et al., 2011; Nagarkar-Jaiswal et al., 2015a) or homologous recombination mediated by CRISPR, a technique we named CRIMIC [CRISPR-mediated Integration Cassette] [Zhang et al., 2014; Diao et al., 2015; Lee et al., 2018a]). A SIC is typically flanked with attP sites and can be replaced using Recombinase Mediated Cassette Exchange (RMCE) (Bateman et al., 2006; Venken et al., 2011). The CRIMIC variety of SIC currently used by the GDP is an artificial exon consisting of attP-FRT-SA-T2A-GAL4-polyA-3XP3-EGFP-FRT-attP inserted in a coding intron (intron flanked by two coding exons) of the GOI (Lee et al., 2018a). This insert typically creates a severe loss of function allele and generates a GAL4 protein that is expressed in the target gene’s spatial and temporal expression pattern (Diao et al., 2015; Gnerer et al., 2015; Lee et al., 2018a). The resulting GAL4 can then be used to drive a UAS-nuclear localization signal (NLS)::mCherry to determine which cells express the gene or a UAS-membrane (CD8)::mCherry to outline the cell projections (Brand and Perrimon, 1993; Shaner et al., 2004). Alternatively, a UAS-GOI cDNA can be used to test for rescue of the loss of function phenotype induced by the insertion cassette. This provides a means for rigorous quality assessment of the genetic reagent and, when combined with mutant and/or truncated forms of the UAS-GOI cDNA, facilitates structure-function analysis. In addition, a UAS-human-homologue cDNA of the GOI permits humanization of the flies and assessment of human variants (Bellen and Yamamoto, 2015; Kanca et al., 2017; Şentürk and Bellen, 2018; Chao et al., 2017; Yoon et al., 2017).

The SIC can also be replaced by an artificial exon that consists of SA-Linker-EGFP-FlAsH-StrepII-TEV-3xFlag-Linker-SD, abbreviated SA-GFP-SD, which adds an integral Green Fluorescent Protein (GFP) and other tags to the gene product (Venken et al., 2011; Nagarkar-Jaiswal et al., 2015a). This tag does not disrupt protein function in 75% of cases examined and permits the determination of the subcellular protein localization (Venken et al., 2011; Nagarkar-Jaiswal et al., 2015a; Yoon et al., 2017; Lee et al., 2018a) as well as removal of the protein in any tissue using specific GAL4 drivers (Jenett et al., 2012) to drive a DeGradFP protein that leads to polyubiquitination and degradation of the protein of interest (Caussinus et al., 2012; Nagarkar-Jaiswal et al., 2015a; Lee et al., 2018b). The GFP tag can also be used as an epitope for immunoprecipitation to determine interaction partners of the tagged protein (Neumüller et al., 2012; Zhang et al., 2013; David-Morrison et al., 2016; Yoon et al., 2017).

Additionally, SICs can be replaced by other RMCE vectors that enable integration of additional binary or tertiary systems (e.g. SA-T2A-LexA, SA-T2A-split GAL4; Gnerer et al., 2015; Diao et al., 2015) to obtain finer tools to express transgenes in specific cell populations. SICs can also be used to generate conditional alleles of targeted genes (Flip-flop and FLPstop; Fisher et al., 2017; Nagarkar-Jaiswal et al., 2017). Finally, strategies have been developed to convert SICs through genetic crosses rather than by injection (Trojan Exons and Double Header; Nagarkar-Jaiswal et al., 2015b; Diao et al., 2015; Li-Kroeger et al., 2018).

Although the above reagents form a powerful toolset, the generation of libraries of many thousands of genes based on these methods is labor-intensive and costly. The cost for reagents and labor for the generation of a single CRIMIC fly line is $1,000–2,000. Indeed, to create each CRIMIC construct we need to amplify two 1 kb homology arms, clone these arms on either site of a SIC in a plasmid, sequence verify the constructs, amplify and inject the DNA with a target-specific gRNA into 600 embryos, screen to obtain several independent transgenic flies, establish several fly stocks for each construct, PCR-verify the insertions, and cross each line with UAS-mCherry to determine expression patterns. Since the production of each transgenic line involves multiple steps, low failure rates at each step accumulate and decrease the overall success rate to ~50%. Given that we are in the process of tagging ~5000 genes that contain suitable introns, it is highly desirable to develop a more efficient, less labor-intensive, and cheaper alternative. One of the main bottlenecks is the production of large (5 kb) SIC homology donor plasmids containing a visible dominant marker and flanked by two ~ 1 kb homology arms to promote homologous recombination (Beumer et al., 2008; Beumer et al., 2013; Bier et al., 2018; Lee et al., 2018a). We therefore explored a series of alternative strategies to reduce the construct size and facilitate cloning.

Here, we report the development of methods, using either a PCR-generated, single stranded DNA donor (ssDNA drop-in) or a synthesized double stranded homology donor (dsDNA drop-in) that greatly simplify the generation of homology donor constructs and improve the transgenesis rate. We tested both methods in vivo in Drosophila and targeted the same 10 genes with five different constructs to assess transformation efficiency and accuracy of integration. We show that the ssDNA drop-in method works efficiently in Drosophila S2R+ cells for constructs that are less than 2 kb and we used this method to mark several cellular organelles with GFP tagged proteins. The dsDNA drop-in strategy is based on short homology arms flanking SICs of up to 5 kb. The success rate for tagging the tested genes was 70–80%. The dsDNA drop-in donor vector is easy and cheap to produce, transformation efficiency is high, and the insertions are precise. Hence, these changes significantly decrease the costs of generating a transgenic CRIMIC library and make the CRIMIC technique more accessible to others. We anticipate that the drop-in approaches will also be useful in other species.

Results and discussion

ssDNA homology donors

dsDNA homology donors for insertion of large cassettes in Drosophila typically require stretches of 500 nt to 1 kb of homology to the target site on either side of the SIC (Rong and Golic, 2000; Beumer et al., 2013; Zhang et al., 2014; Diao et al., 2015; Bier et al., 2018; Lee et al., 2018a). The large size of the homology regions affects cloning efficiency of the donor constructs (Lee et al., 2018a). Single stranded homology donors typically rely on much shorter homology arms (50–100 nt) to successfully integrate short DNA segments in Drosophila (~200 nt) based on homology-directed repair (Beumer et al., 2013; Gratz et al., 2013; Wissel et al., 2016; Bier et al., 2018). To facilitate and speed up the preparation of homology donor constructs, we decided to test ssDNA donors for CRISPR-mediated homologous recombination of donors that are 1 to 2 kb.

To produce ssDNA homology donors, we established a cloning-free method based on PCR (Higuchi and Ochman, 1989). We generated SIC containing PCR templates that can be amplified with an M13 universal primer-derived primer pair (26 nt forward primer and 24 nt reverse primer). Gene-specific 100 nt homology arms are incorporated into these primers as 5’ overhangs (left homology arm to the forward primer and reverse complement of right homology arm to the reverse primer; Figure 1A). One of the primers is phosphorylated at its 5’ end. The resulting PCR product contains a 5’ phosphorylated strand and a non-phosphorylated strand. When this PCR product is treated with Lambda Exonuclease, a 5’-to-3’ nuclease with a preference for 5’ phosphorylated DNA (Little, 1967; Mitsis and Kwagh, 1999), the 5’- phosphorylated strand is degraded, leaving the non-phosphorylated strand as a ssDNA homology donor. This PCR-based approach has several advantages. First, the homology arms are included in the ~125 nt primers as 5’ overhangs such that a single PCR generates the complete donor without any cloning steps. Second, the same primers can be used to amplify many different SIC templates for the same GOI. Third, the priming sequences and overall construct length do not change between genes. Hence, the same protocol can be applied to create homology donors for different genes. Fourth, the PCR-based method does not require bacterial transformation, eliminating possible rearrangements associated with propagation of DNA in bacteria (Figure 1B). Finally, the risk of inserting PCR-induced mutations in the SIC is negligible when a proofreading polymerase is used.

ssDNA pipeline is faster than cloning CRIMIC constructs.

(A) Schematic of PCR-based generation of drop-in ssDNA constructs. Gray boxes, UTRs; orange boxes, coding exons; yellow line, coding introns; black line, outside coding introns and exons. (B) Comparison of donor generation pipelines for CRIMIC and PCR-based drop-in ssDNA homology donors. Making ssDNA donors is about 5X faster than making CRIMIC donors.

https://doi.org/10.7554/eLife.51539.002

We first tested the efficacy of ssDNA drop-in constructs as a substrate for homology-directed repair in Drosophila S2R+ cells stably transfected with Cas9 (S2R+-MT::Cas9; Viswanatha et al., 2018). Specifically, we were interested in generating a collection of Drosophila S2R+ cells in which different organelles are marked by a protein tagged with a superfolder GFP (sfGFP) in each cell line (Pédelacq et al., 2006). We therefore amplified a SIC consisting of attP-SA-Linker-sfGFP-Linker-SD-attP (1392 nt including homology arms) using 20 different gene-specific primer pairs (Figure 2A, Supplementary file 3). Each gene-specific ssDNA homology donor was electroporated into Cas9-positive cells along with a corresponding gene-specific gRNA and then subjected to fluorescence-activated cell sorting (FACS) (Figure 2B).

Figure 2 with 3 supplements see all
ssDNA homology donors are effective in S2 cells to tag organelles.

(A) Schematic of drop-in cassette encoding for sfGFP artificial exon. Size of the construct including the homology arms is indicated on the right. sfGFP: superfolderGFP; SA: Splice Acceptor of mhc; SD: Splice Donor of mhc; L: flexible linker that consists of four copies of Gly-Gly-Ser. (B) Diagram of steps to isolate cell clones resulting from successful homologous recombination events. (C) Examples of S2R+ cells with organelles marked with GFP. Left panel, antibody staining; middle panel, GFP signal; right panel, the merge.

https://doi.org/10.7554/eLife.51539.003

The frequency of GFP+ cells was determined by FACS (Figure 2—figure supplement 1) to range between ~1–4%. GFP-positive cells were clonally isolated, cultured, and analyzed. We observed that for any given clone, all cells had the same subcellular GFP localization, indicating that they were derived from a single cell and that the insertion was stably integrated. For 19 out of 20 genes targeted we observed GFP signal by FACS (Table 1). For 12 of 19 genes, we could establish GFP+ clones, verify correct integration by PCR, and determine subcellular localization by immunostaining (Table 1, Figure 2C, Figure 2—figure supplement 2). With the exception of Ref2P, all correct insertions of GFP resulted in fusion proteins with the expected subcellular distribution. The genes for which we observed GFP signal by FACS but could not successfully isolate GFP+ clones tended to be expressed at low levels in S2R+ cells based on modEncode expression profiling (Cherbas et al., 2011; Table 1). For these genes it is possible that the signal-to-noise ratio for GFP was insufficient to robustly select GFP+ clones, leading to the loss of positive cells in the population.

Table 1
Summary of ssDNA drop-in mediated homologous recombination in S2R+ cells
https://doi.org/10.7554/eLife.51539.007
OrganelleFly proteinclones obtainedpopulation imagedclones imagedinsertion sequence verifiedImmunostainedCorrect GFP localizationDGRC stock#S2R+ expression (modENCODE RPKM)
AutophagosomesAtg8a14NNNN153
Autophagosomes/aggregatesRef(2)P9Ya-Ref2P, a-FK2NN138
Endoplasmic reticulum (ER)Calnexin99A16Ya-Cnx99aY273235
Endoplasmic reticulum (ER), transitionalSec2330YN*294101
Endosomes, earlyRab59NNNN98
Endosomes, recyclingRab1123YN*274302
G-Bodies (cytoplasmic puncta)Pfk14NNNN23
Golgi (cis-Golgi)Gmap10Ya-GMAPY276, 27715
Golgi (trans-Golgi)Sec7110NNNN8
Golgi (trans-Golgi)Golgin2451Ya-Golgin245Y28027
KinetochorePolo2YNY27550
Lipid dropletsSeipin12NNNN9
Lysosomesspin2Ya-Arl8Y293112
LysosomesArl89Ya-Arl8Y29178
MitochondriaTim17b3Na-ATP5ANN266
MitochondriaTom2017Ya-ATP5AY302117
Nuclear membrane, innerdLBR0NNNNN42
Nuclear membrane, innerLamin53Ya-LaminY292249
NucleolusFibrillarin14Ya-FibY278, 27953
PeroxisomesPmp706NNNN26
  1. * = distribution was as expected, but no antibody available to test by co-stain

Sequencing of the SIC insertion sites (Table 1) for the 12 cell lines showed that the insertions are precise. Western blotting of cell lysates confirmed that the inserted tags lead to GFP fusion proteins of the expected molecular weights (Figure 2—figure supplement 3). Given the dynamic localization of Polo protein during mitosis (Llamazares et al., 1991) we recorded the Polo subcellular localization pattern in live cells through mitosis. Time-lapse confocal imaging of Polo-GFP showed that the protein is localized to centrosomes, spindle, and midbody during cell division, in agreement with the data obtained by immunofluorescence (Llamazares et al., 1991) or using polo-GFP transgenes (Moutinho-Santos et al., 1999; (Video 1).

Video 1
Polo-sfGFP knock-in in S2R+ cells through ssDNA drop-in faithfully reports dynamic localization of Polo throughout the cell cycle.
https://doi.org/10.7554/eLife.51539.008

Previously, PCR generated double-stranded constructs containing GFP and an antibiotic resistance gene have been used for homologous recombination in S2 cells (Böttcher et al., 2014; Kunzelmann et al., 2016). However, selection with a drug resistance gene was used to enrich the population and GFP integration frequency was about 2% as judged by FACS. We were able to generate GFP protein traps in S2 cells using long ssDNA in up to 4% transfected cells and obtain clones without using drug-based selection (Figure 2, Figure 2—figure supplement 2).

Previous studies by Richardson et al. (2016) analyzed the binding dynamics of Cas9 to target sites and showed an increase in homologous recombination efficiency by using homology donors with asymmetric homology arms. Whether the use of asymmetric homology arms will increase the knock-in efficiency remains to be tested for drop-in.

In summary, ssDNA drop-in constructs are simple to produce and provided an efficient homology substrate in S2R+ cells for about 60% of the tested genes. The generated cell lines with marked organelles—endoplasmic reticulum (2), recycling endosomes, cis and trans Golgi, kinetochores, lysosomes (2), mitochondria, nucleoli and nuclear envelopes (Table 1)—are listed at the Drosophila RNAi Screening Center (DRSC) website (https://fgr.hms.harvard.edu/crispr-modified-cell-lines) and available through the Drosophila Genomics Resource Center (DGRC; https://dgrc.bio.indiana.edu). These cell lines will be a useful resource for small- and large-scale studies of organelle biogenesis, organelle function, and/or subcellular distribution of organelles or specific fusion proteins, and genetic and pharmacological screens to identify regulators of organelle homeostasis.

Integration of ssDNA drop-in donors in the germline

We next generated a drop-in SIC for in vivo Drosophila transformation that can be amplified using PCR primers. Our initial experiments showed that the empirical size limit for ssDNA production was ~2 kb, significantly smaller than the ~5 kb SIC present in the CRIMIC cassettes currently used by the GDP (Lee et al., 2018a). A minimal SIC should contain attP sites to enable downstream RMCE applications and a dominant marker that allows detection of positive insertions in the targeted locus. The smallest self-sufficient visible dominant marker in Drosophila is 3XP3-EGFP (Horn et al., 2000). Hence, in order to remain under the size limit, we generated an attP-3XP3-EGFP-attP PCR template that is 1242 nt long and can be used for ssDNA donor (1442 nt including two 100 nt homology arms) generation (Figure 3A). The attP-3XP3-EGFP-attP cassette is not mutagenic unless inserted in exons and only generates a landing site for RMCE in the targeted region.

ssDNA constructs are not as efficient as double stranded CRIMIC constructs for fly transformation.

(A) Schematic of drop-in cassette used for fly transformation. Size of the construct including homology arms is indicated on the right. (B) Injection results for the 10 genes selected for comparison of transformation efficacy. Numbers indicate positive events/fertile G0 single fly crosses. Red, no positive events; light green, positive non-confirmed events; dark green, genes with PCR-confirmed events.

https://doi.org/10.7554/eLife.51539.009

To test the efficiency of this new ssDNA drop-in construct in vivo, we compared the efficiency of the ssDNA drop-in donor (with 100 nt homology arms) and the current CRIMIC dsDNA donor (attP-FRT-SA-T2A-GAL4-polyA-3XP3-EGFP-FRT-attP with ~1 kb homology arms) for the same 10 genes using the same gRNAs by injecting ~400–600 embryos for each gene with each construct. Surviving adults were crossed with y w flies as single fly crosses and the numbers of independent targeting events were quantified. We targeted a set of 10 genes with each construct to have a large enough set to reduce the possibility of locus-specific properties skewing the results. We were successful in targeting four genes with both approaches and obtained insertions in CG5009, CG9527, Khc, and NLaz with the CRIMIC approach and CG5009, endoB, Cp1, and Lst with the ssDNA drop-in construct. We found that the use of CRIMIC constructs was more efficient, as judged by the number of independent fly lines produced per successful gene (Figure 3B).

Short ssDNA homology donors were previously used in Drosophila to integrate small epitope tags or site-specific recombination sites (Gratz et al., 2013; Wissel et al., 2016). Due to size constraints, these constructs lacked a dominant marker and detection of successful events relied on labor-intensive PCR strategies. More recently, novel methods were developed to generate and use longer ssDNAs (~1000 nt) as homology donors in vivo in mice (Miura et al., 2015; Quadros et al., 2017; Lanza et al., 2018). Our results show that longer ssDNA constructs with visible dominant markers integrate in the fly genome in vivo albeit with lower efficacy than double stranded homology donors with large homology arms (i.e. CRIMIC donors).

In summary, the ssDNA constructs are efficient donors for S2R+ cells for protein tagging by knock-in but for fly transgenesis they are not more efficient than the standard CRIMIC method. Nevertheless, the ease and low cost of producing these constructs and the transformation efficiency in S2 cells may justify their use to integrate GFP tags or landing sites in cultured cells or the germline.

dsDNA drop-in donors of <2 kb are efficient homology donors for transgenesis

Given that the ssDNA drop-in constructs did not increase the success rate of fly transformation, we next attempted to optimize dsDNA homology donors for production ease and transformation efficiency. One means of increasing the targeting rate is to shorten the SIC, since homologous recombination is dependent on the size of the inserted cassette (Beumer et al., 2013). We used three strategies to shorten SICs (Figure 4A). First, we used a shortened GAL4 construct, referred to as miniGAL4 (1200 nt), which is about half the size of full-length GAL4 (2646 nt) and has about 50% of the transcriptional activity of full-length GAL4 in yeast (Ding and Johnston, 1997). Second, we used a very short dominant marker. Because 3XP3-EGFP is ~1 kb in length, we opted to use an alternative marker that is only ~200 nt long and contains a ubiquitous U6 promoter expressing a gRNA sequence (gRNA1) that does not have a target in the fly genome. This strategy is based on the single-strand annealing (SSA) pathway of DNA repair to reconstitute expression of a transgenic ubiquitous mCherry that is interrupted by the gRNA1 target sequence (Figure 4—figure supplement 1; Garcia-Marques et al., 2019). Upon Cas9-mediated double strand break, the modified non-functional mCherry gene is repaired and becomes functional, providing a visible marker. This SSA dependent repair reaction occurs in the F1 generation upon stable integration of U6:gRNA1 in the genome (Figure 4—figure supplement 1). Third, we shortened the polyA sequence from 135 nt to 35 nt (minipolyA) (McFarland et al., 2006). This resulted in attP-SA-T2A-miniGAL4-minipolyA-U6gRNA1-attP, which functions as a gene trap and is 1968 nt in length, including homology arms. We also generated a smaller minimal mutagenic construct, attP-SA-3XSTOP-minipolyA-U6gRNA1-attP, that is 715 nt in length including homology arms (Figure 4A). Both dsDNA drop-in constructs are small enough to be commercially synthesized at a low cost (less than $250).

Figure 4 with 1 supplement see all
Double stranded DNA synthetic constructs are efficient for fly transformation.

(A) Schematic of synthesized plasmid drop-in donors. mPA indicates minipolyA. (B) Schematic of the donor plasmid followed by linearization by Cas9 in germ cells and integration of donors in vivo. gRNAtarget is gene specific gRNA. (C) Injection results for the 10 genes selected for comparison of transformation efficacy. Numbers indicate positive events/fertile G0 single fly crosses. Red, no positive events; light green, positive non-confirmed events; and dark green, genes with PCR-confirmed events.

https://doi.org/10.7554/eLife.51539.010

To further improve the insertion efficiency, we decided to induce in vivo linearization of the plasmid constructs. Linearization has previously been shown to boost homologous recombination rates in cell culture, zebrafish and mouse transgenesis (Cristea et al., 2013; Hisano et al., 2015; Suzuki et al., 2016; Yao et al., 2017). Hence, use of short 100 nt left and right homology arms flanked by the gRNA1 target sites to linearize the construct in vivo upon injection may significantly increase the frequency of homologous recombination (Figure 4B).

It is worth noting that higher transgenesis rates have been reported in mice and worms using donor constructs linearized in vitro (Paix et al., 2014; Paix et al., 2015; Paix et al., 2017; Dokshin et al., 2018; Yao et al., 2018). However, in vitro linearized DNA or direct PCR products are very poor substrates for homologous recombination for germ line transformation in Drosophila (Beumer et al., 2008; Böttcher et al., 2014). We therefore opted to use dsDNA donors that are linearized in vivo by the Cas9 expressed in germ cells.

Upon injection of gene-specific gRNA coding plasmids and donor plasmids we isolated multiple transgenic lines for each of the 10 targeted genes. We obtained 65 independent transgenic lines for the attP-SA-3XSTOP-minipolyA-U6gRNA1-attP construct, and 119 independent transgenic lines for the attP-SA-T2A-miniGAL4-minipolyA-U6gRNA1-attP construct. We verified the insertion sites by genomic PCR in 7/10 genes for both methods. Analyses of the genomic DNA sequences of the three genes for which we could not verify the integration site revealed that these genes contain an unanticipated variation in the gRNA target site in the isogenized Cas9 injection stocks when compared to the FlyBase reference genome sequence (dos Santos et al., 2015; Goodman et al., 2018). This suggests that the precise homologous recombination rate is dependent on efficient cutting of the target site and is likely higher than 70%. Our current gRNA design platform considers these strain variations which can be substantial in introns.

Because we shortened the polyA tail from 135 nt (Okada et al., 1999) to 35 nt (McFarland et al., 2006), we assessed the mutagenic potential of new constructs that contain minipolyA. We used Western blotting with an antibody that recognizes the gene product of CG5009 (Drosophila ortholog of ACOX1) to compare protein levels in animals homozygous for the CRIMIC allele (with 135 nt polyA tail) with protein levels of animals homozygous for either of the two constructs that carry the minipolyA. As shown in Figure 5A, the artificial exon attP-SA-3XSTOP-minipolyA-U6gRNA1-attP leads to ~40% decrease in protein levels and creates a much less severe allele than the CRIMIC allele. However, the attP-SA-T2A-miniGAL4-minipolyA-U6gRNA1-attP cassette leads to ~80–90% decrease in protein levels, similar to a CRIMIC insertion in the same locus (Figure 5A). The lower mutagenic efficacy of attP-SA-3XSTOP-minipolyA-U6gRNA1-attP may be the result of read through of stop codons, inefficient transcriptional stop at the minipolyA sequence or smaller size of the inserted artificial exon. The observation that attP-SA-T2A-miniGAL4-minipolyA-U6gRNA1-attP is potently mutagenic indicates that the combination of SA-T2A with minipolyA is stronger than SA-3XSTOP-minipolyA. Alternatively, the increased size of artificial exon in attP-SA-T2A-miniGAL4-minipolyA-U6gRNA1-attP overcomes the limitation of attP-SA-3XSTOP-minipolyA-U6gRNA1-attP construct in mutagenesis efficacy.

T2A-miniGAL4 gene trap cassette is mutagenic and expresses an active GAL4.

(A) Western blot and quantification of the level of ACOX1 protein in flies homozygous for SA-T2A-miniGAL4-minipolyA, SA-3XSTOP-minipolyA or CRIMIC construct for CG5009. ****p<0.0001, *p<0.01. mPA1 and mPA2 indicate two independent lines of CG5009-SA-3XSTOP-minipolyA. (B) Western blot of β-galactosidase and quantification of heterozygous flies carrying a copy of the miniGAL4 construct compared to heterozygous flies carrying a CRIMIC.

https://doi.org/10.7554/eLife.51539.012

The miniGAL4 construct had never been tested in flies. We therefore compared full length versus miniGAL4 induction of a UAS-LacZ reporter by Western blot. As shown in Figure 5B, we observed a ~ 60% reduction of reporter levels with T2A-miniGAL4 as compared to CRIMIC T2A-GAL4 driven UAS-LacZ expression for the CG5009 locus (Figure 5B). Hence, the attP-SA-T2A-miniGAL4-minipolyA-U6gRNA1-attP drives lower levels of reporter expression when compared to CRIMIC, which may limit the use of miniGAL4 for genes expressed at low levels.

In summary, the decrease of SIC size and linearization increased the transgenesis rate compared to CRIMIC templates. We used novel selection markers, smaller polyA tails, and new GAL4 variants to decrease the size of the integrated construct. However, this decrease in size came with trade-offs. The U6-gRNA1 dominant marker is very easy to detect and is smaller than previously established dominant markers. However, this marker requires the presence of other transgenes for detection (Figure 4—figure supplement 1) and the reconstituted dominant marker transgene segregates independently from the targeted gene. These limitations make establishing and maintaining stocks more challenging. Moreover, T2A-miniGAL4 may not be strong enough to drive robust reporter expression in genes with low expression levels.

Large double stranded drop-in dsDNA donors (~5 kb) are efficient homology donors to integrate CRIMIC-like SICs

To avoid the issues raised in the previous section, we designed a strategy to integrate full length CRIMIC cassettes with short homology arms. To achieve this, we tested whether large DNA fragments with short homology arms (100 nt) could be integrated in target genes upon linearization using the gRNA1 in vivo. This would allow the use of dominant markers like 3XP3-EGFP as well as integration of the full length GAL4 gene with an extended polyA tail, that is the CRIMIC cassette SIC. Unfortunately, synthesis of a full length 5 kb CRIMIC cassette would be cost prohibitive (>$1,000) especially for thousands of genes. Hence, we developed a modified cloning strategy in which we first synthesize a gRNA1 target-100nt homology arm-Restriction cassette-100nt homology arm-gRNA1 target inserted into a pUC57 vector for each target gene (cost is $80). The SIC containing attP-FRT-T2A-GAL4-polyA-3XP3-EGFP-FRT-attP is then subcloned directionally into this plasmid in a single straightforward cloning step, replacing the restriction cassette with the SIC of interest (Figure 6A). We refer to these constructs drop-in int100-CRIMIC constructs. The SIC can be replaced by any other effector (e.g. SA-GFP-SD) to generate other homology donor constructs.

Figure 6 with 1 supplement see all
Single step cloning method allows efficient insertion of the CRIMIC cassette in coding introns.

(A) Schematic of a single step cloning vector pUC57. LHA Left Homology Arm, RE1 Restriction Enzyme 1, RE2 restriction Enzyme 2, RHA Right Homology Arm. (B) Injection results for the seven genes selected to estimate transformation efficacy. Numbers indicate positive events/G0 single fly crosses. (C) Third instar larval brain expression domain of CG5009 as determined by crossing conventional CRIMIC or drop-in int100-CRIMIC flies to UAS-NLS-mCherry reporter lines. Scale bar is 100 µm.

https://doi.org/10.7554/eLife.51539.013

We injected vectors containing the full length CRIMIC cassette for seven of the genes in which we previously successfully inserted dsDNA drop-in cassettes. Using the drop-in int100-CRIMIC cassette we obtained multiple knock-in alleles in five genes, as verified by PCR (Figure 6B). Hence, 100 nt homology arms are sufficient to integrate large SICs into target sites. In addition, increasing homology arm length to 200 nt should not increase synthesis costs, as the total length of the construct remains less than 500 nt. Whether use of 200 nt may improve efficiency remains to be tested.

To ensure functionality of knock-in alleles, we compared expression patterns of drop-in int100 CRIMIC with the corresponding CRIMIC for CG5009 (Figure 6C) as well as drop-in int100 CRIMIC with the T2A-miniGAL4 for Khc (Figure 6—figure supplement 1) by crossing the flies to UAS-NLS::mCherry reporter lines. In both cases the constructs lead to mCherry expression in very similar patterns.

In vivo linearization was previously shown to lead to knock-ins in zebrafish and cell culture even when there is no homology arm, provided that the homology donor and target site are cut concomitantly (Cristea et al., 2013; Auer et al., 2014; Schmid-Burgk et al., 2016; Suzuki et al., 2016). This process is more prone to generating small deletions in the target site and the insertions are non-directional. Recently, a homology independent knock-in method was established for Drosophila cell culture and germline transformation (Bosch et al., 2019). By simultaneously cutting the donor construct and target region in the absence of homology arms, Bosch et al. (2019) integrated a CRIMIC cassette in 4 of 11 genes targeted. However, these insertion events were error-prone and non-directional. Note that for drop-in constructs, the presence of short homology arms allows the donor vectors to be synthesized cheaply and introducing larger SICs is accomplished by a single straightforward cloning step, providing a good balance between ease of construct generation and efficient in vivo use.

In summary, we have developed efficient pipelines for CRISPR knock-in using ssDNA in Drosophila cells or a dsDNA approach in the germline. With respect to the dsDNA donors, we significantly improved the overall efficiency compared to the method described in Lee et al. (2018a). The previous method that required cloning large homology arm flanked cassettes has several pitfalls: homology arm PCRs often must be troubleshooted repeatedly; assembly products are often incorrect; and sequencing of the final product is often challenging and needs to be repeated to confirm the construct. Moreover, when the efficacy of fly transformation and the events that can be PCR verified on either side of knock-in region are included, the efficacy of CRIMIC constructs hover around 50% in optimal conditions. With the drop-in cassette strategy, not only is the cloning success rate nearly 100%, but the protocol requires little troubleshooting, dramatically reducing the bench time. Finally, factoring in the transformation and verification rate drop-in results in a 70–80% success rate. Key features that improved efficiency are: 1) shortening the homology arms to 100 nt, which allows synthesis followed by a simple cloning step, thus eliminating cloning failure and reducing the cost from ~$300 to ~$100 per construct; and 2) integration of target gRNA1 sites on either side of the SIC to linearize the donor with Cas9 in the germ cells. In conclusion, the methods we describe are efficient, simple, and precise.

Materials and methods

Generation of templates for ssDNA production

Request a detailed protocol

The sequences of scaffold vectors can be found in Supplementary file 1. Briefly, to generate the attP-3XP3-EGFP-attP, the scaffold vector 3XP3-EGFP cassette was amplified by PCR using the long primers 3XP3-EGFP-RI-for and 3XP3-EGFPrev-NotI and cloned between the EcoRI and NotI sites in pCasper3 (Thummel and Pirrotta, 1992). To generate attP-SA-sfGFP-SD-attP, a scaffold vector that we named pScaffold was produced by integrating annealed oligonucleotides with sequences M13For-attPfor-SbfI-AvrII-attPrev-M13rev in pCasper3 backbone in EcoRI-NotI sites. SA-sfGFP-SD was cloned as a three-fragment ligation in pScaffold with linker-SA (amplified from pDoubleHeader [Li-Kroeger et al., 2018]) with primers SA-for-Sbf and Linker-SA-rev_BamHI), sfGFP (amplified from pUAST-NLS-sfGFP-3XMyc-PEST with primers sfGFP-for_BamHI and sfGFP-rev_KpnI), and Linker-SD (amplified from pDoubleHeader with primers Linker-SD-for_KpnI and Linker-SD-rev_NotI).

Production of ssDNA drop-in constructs

Request a detailed protocol

Gene-specific homology donors were produced by incorporating the homologous sites (regions spanning 100 nt upstream of 3 nt prior to the PAM (protospacer adjacent motif) for left homology arm or 97 nt downstream of that region for the right homology arm (Supplementary file 1) as 5’ overhangs to primers with complementarity to the template (26 bases for the forward primer and 24 bases for the reverse primer). The long primers were ordered from IDT (Coralville, Iowa) as Ultramers. See Supplementary file 3 for sequences of primers. The reverse primers were 5’ phosphorylated. Four 50 μl PCR reactions were set up with Q5 Hot Start High-Fidelity 2X Master Mix (NEB #M0494L). PCR conditions were optimized using gradient PCR (Bio-Rad C1000 Touch). The optimal annealing temperature was 70°C. The elongation time was 1 min for attP-3XP3-EGFP-attP and attP-SA-sfGFP-SD-attP. Resulting PCR amplicons were pooled in two samples and isolated using Qiaquick spin columns (Qiagen #28106) following the manufacturer’s protocol. Samples were eluted in 2 × 50 μl elution buffer from the kit. Two lambda exonuclease (NEB #M0262L) digestion reactions of 100 μl were set up using 4–6 μg DNA and 8 μl enzyme each. Digestion conditions were 37°C for 1 hr, followed by 10 min at 75°C for heat inactivation. Reaction products were pooled in two samples and ssDNA isolated using NEB Monarch DNA purification kit following the manufacturer’s instructions (NEB #T1030L). Samples were isolated in 2 × 10 μl prewarmed (55°C) elution buffer from the kit and the DNA concentration was measured using NanoDrop One (ThermoFisher Scientific).

Generation of GFP Knock-In cell lines

Cell culture and regular media

Request a detailed protocol

Drosophila cells stably expressing Cas9 (S2R+-MT::Cas9; Drosophila Genomics Resource Center cell stock #268; Viswanatha et al., 2018) were cultured in Schneider’s Drosophila Medium 1X (ThermoFisher Scientific #21720024) with 10% FBS and 1% penicillin/streptomycin (referred to as regular media). We note that this cell line is a derivative of S2R+ NPT005 (DGRC #229) and thus contains an mCherry tag in the Clic locus (Neumüller et al., 2012).

Conditioned media

Request a detailed protocol

Conditioned media were prepared as previously described by Housden et al. (2017).

Cellular transfection

Request a detailed protocol

Electroporation of S2R+-MT::Cas9 cells was performed using a Lonza 4D Nucleofector (Lonza #AAF-1002B) following the manufacturer’s protocol. For each transfection, 1 µg of sgRNA (100 ng) and 2 µL of the sfGFP donor (100 ng) were used with 4 × 105 sub-confluent S2R+-MT::Cas9 cells. See Supplementary file 3 for all sgRNA, long primers and insert sequences. After electroporation, cells were immediately placed in regular media and health was monitored. Cell cultures were then maintained and expanded for FACS.

Isolation of Single-Cell clones

Request a detailed protocol

Prior to FACS, 106 cells were filtered through a 40 µM Falcon Cell Strainer (Corning #21008–949) into 15 mL conical tubes (ThermoFisher Scientific #14-959-70C). Single cells were then FACS-isolated on a BD Aria IIu, based on the presence of GFP with an expression level greater than 2 × 102 (see Figure 2—figure supplement 1). Single cells were sorted into wells of a 96 well plate (VWR #29444–010) filled with 100 µL of conditioned medium. Cells were observed ~14 days later and any viable clones were expanded.

Analysis of GFP Knock-In cell lines

Image analysis

Request a detailed protocol

Multiple clonal cell lines were generated per gene and image analysis was used to measure the expression of the GFP knock-in marker. Several images of each live clone were taken using an InCell Analyzer 6000 automated confocal fluorescence microscope (GE Healthcare Lifesciences) using the dsRed channel to detect mCherry fluorescence present in all cells and the FITC channel to detect GFP fluorescence. Images were analyzed using CellProfiler (version 2.1.1). As mCherry is present throughout each cell, this image was used by CellProfiler to determine the outline of individual cells using the command ‘Identify Primary Objects.’ The outlines generated from this step were then applied to the GFP image through the process ‘Identify Secondary Objects.’ GFP fluorescence intensity was then measured using the command ‘Measure Object Intensity,’ which averages the GFP fluorescence within all of the cellular outlines (see Figure 2—figure supplement 2). Mean GFP intensities for each clone were noted, and the three clones per gene with the highest means were chosen for molecular analysis. All other clones were pooled and stored for use in a subsequent FACS if needed. Image data were managed using OMERO, as supported by the Harvard Medical School Image Data Management Core.

Genomic DNA extraction

Request a detailed protocol

Genomic DNA was extracted from individual clones using a Quick-DNA MiniPrep Kit (Zymo Research #D3024) according to the manufacturer’s protocol. Concentrations were measured using a NanoDrop 8000 Spectrophotometer (ThermoFisher Scientific #ND-8000-GL).

Molecular analysis

Request a detailed protocol

Due to the large size of the insertion, the 5’ and 3’ integration sites of each insertion were analyzed by PCR amplifying and sequencing a fragment spanning the junction of each end of the insertion with flanking DNA (see Supplementary file 3 for a list the two primer sets per gene). These flanking sites were PCR amplified with High Fidelity Phusion Polymerase (NEB #M0530) using the following program: 1) 98°C for 30 s, 2) 98°C for 10 s, 3) 56°C for 30 s, 4) 72°C for 30 s (35 cycles), 5) 72°C for 10 min, 6) 4°C hold. PCR products were excised from a 2% TAE agarose gel and purified using a QIAquick Gel Extraction Kit (Qiagen #28704) and Sanger sequenced at the Dana-Farber/Harvard Cancer Center DNA Resource Core.

Immunostaining

Request a detailed protocol

Cells were fixed in 4% paraformaldehyde in phosphate-buffered saline with 0.1% Triton X-100 for 30 min. A standard staining protocol was used. Primary antibodies were used as indicated: Chicken anti-GFP (1:1000, Abcam ab13970), Mouse anti-Lamin (1:500, DSHB #ADL84.12), Mouse anti-Calnexin99A (1:5, DSHB #Cnx99A 6-2-1), Mouse anti-ATP5A (1:100, Abcam ab14748), Rabbit anti-Arl8 (1:500, DSHB Arl8), Goat anti-GMAP (1:2000, DSHB, #GMAP), Goat anti-Golgin245 (1:2000, DSHB, #Golgin245), Rabbit anti-Ref2P (1:500, Abcam #ab178440), Rabbit anti-Myc-tag (1:1000, Cell Signaling Technology #2278S). Secondary antibodies with Alexa Fluor conjugates and DAPI (Molecular Probes #D-1306) were used at 1:1000. Images were obtained using a GE IN Cell 6000 automated confocal microscope with a 60x objective. Time-lapse videos were generated by imaging every 30 s over a 2 hr period.

Western blotting

Request a detailed protocol

Clones that are generated using phase one template contained a 5 bp deletion in the splice donor portion of the insert. To determine if full-length fusion proteins were generated, we analyzed the following: S2R+-MT::Cas9 (unmodified control), Actin::GFP (a GFP-positive control), Rab11 GFP knock-in (no deletion in the insert control), spin GFP knock-in (5 bp deletion), two Lam GFP knock-in clones (5 bp deletion in the first, 206 bp deletion in the second), Ref2p GFP knock-in (5 bp deletion), and Cnx99a GFP knock-in (5 bp deletion). Cells were spun down and lysed by resuspension in Pierce RIPA Buffer (ThermoFisher Scientific #89901) and Halt Protease Inhibitor Cocktail, EDTA-Free (ThermoFisher Scientific #87785). Cells were agitated for 30 min at 4°C and then centrifuged at 15,000 rpm for 20 min at 4°C. The lysates (supernatant) were removed and held at −20°C. Protein concentrations were determined using a Pierce BCA Protein Assay Kit (ThermoFisher Scientific #23227) according to the manufacturer’s protocol. Appropriate volumes of each lysate were added to 4X Laemmli Sample Buffer (BioRad #1610747), vortexed, held at 100°C for 10 min, and then spun down at 13,000 rpm for 3 min before loading into a Mini-PROTEAN TGX Precast Gel (BioRad #4561095) and running at 100 V for 1 hr. The gel was then transferred onto a PVDF membrane (BioRad #1620177) using the Trans-Blot Turbo Transfer System (BioRad #1704150). After blocking with 5% blocking solution and washing in TBST, the membrane was incubated in the primary antibody rabbit anti-GFP (1:5000; Molecular Probes #A6455) shaking at 4°C overnight. The membrane was then washed four times with TBST and incubated with Donkey anti-rabbit HRP (1:3000; GE Healthcare #NA934) for 1 hr at room temperature, washed with TBST, and prepared for imaging using the SuperSignal West Pico PLUS Chemiluminescent Substrate (ThermoFisher Scientific #34580) according to manufacturer’s protocol. The blot was imaged using a ChemiDoc MP Imaging System (see Supplementary file 3 top panel; BioRad #17001402). The blot was then stripped with Restore PLUS Western Blot Stripping Buffer (ThermoFisher Scientific #46430), reprobed with mouse anti-tubulin (1:2000; SigmaT #5168) and Sheep anti-mouse HRP (1:3000; GE Healthcare #NXA931), and re-imaged (see Supplementary Figure # middle panel). The blot was then stripped again and reprobed with mouse anti-lamin (1:500; DSHB #ADL84.12) and Sheep anti-mouse Horse Radish Peroxidase (HRP) (1:3000; GE Healthcare #NXA931), and re-imaged (see Supplementary Figure # bottom panel).

For western analysis in adult flies, flies were dissected and lysed in 0.1% CHAPS buffer [50 mM NaCl, 200 mM HEPES, 1 mM EDTA and protease inhibitor cocktail (Roche)]. Tissue or cell debris were removed by centrifugation. Isolated lysates were subjected to electrophoresis using a 4–12% gradient SDS-PAGE gel and transferred to Immobilon-FL polyvinylidene difluoride membranes. Loading input was adjusted for protein concentration. Primary antibodies used were as follows: Rabbit anti-ACOX1 (1:1000; HPA021195, Sigma), Rabbit anti-β-galactosidase (1:1000; MP Biomedicals #55976), and Mouse anti-Actin-c4 (1:5000; Millipore Sigma #MAB1501). Secondary antibodies include Jackson ImmunoResearch HRP conjugated (1:5000). Blots were imaged on a Bio-Rad ChemiDocMP. The intensity of each band was measured and normalized to a loading control using Imagelab software (Bio-RAD). Three biological replicates were performed and ordinary one way ANOVA was used to compare expression levels of ACOX1 in different conditions. Two technical replicates were performed for β-galactosidase measurements.

Fly injections

Request a detailed protocol

ssDNA constructs were injected at 50–100 ng/µl concentration with 25 ng/µl gene specific gRNA encoding pCFD3 vector (Port et al., 2014). attP-SA-3XStop-minipolyA-U6gRNA1-attP and attP-SA-T2A-miniGAL4-minipolyA-U6gRNA1-att dsDNA drop-in constructs were injected at ~150 ng/µl concentration together with 25 ng/µl gene specific gRNA. dsDNA drop-in int100-CRIMIC constructs were injected at 300–400 ng/µl along with 25 ng/µl gene specific gRNA and 25 ng/µl pCFD3-gRNA1. Injections were performed as described in Lee et al. (2018a). 400–600 y w; iso; attP2(y+){nos-Cas9(v+)} embryos per genotype were injected. Resulting G0 males and females were crossed to y w flies as single fly crosses for 3XP3-EGFP detection and with actin5C-Cas9; actin5C-GF-gRNA2-FP; actin5C-mCherr-#1-ry flies for gRNA1 detection (Figure 4—figure supplement 1; Garcia-Marques et al., 2019). Up to five independent lines were generated per construct per gene. actin5C-GF-#2-FP is an internal control that would detect non-specific activation of dominant markers. We have not detected GFP in any of the screened flies, showing the specificity of dominant marker detection through gRNA1.

PCR validation

Request a detailed protocol

PCR primers that flank the integration site were designed for each targeted gene (Supplementary file 3 for primer sequences). These primers were used in combination with insert-specific primers that bind 5’ of the inserted cassette in reverse orientation and 3’ of the insert in forward orientation (pointing outwards from the insert cassette). 200–800 nt amplicons were amplified from genomic DNA from individual insertion lines through single fly PCR (Gloor et al., 1993) using OneTaq PCR master mix (NEB #M0271L). PCR conditions were denaturation at 95°C for 30 s, 95°C 30 s, 58°C 30 s, 68°C 1 min for 34 cycles and 68°C 5 min.

dsDNA drop-in constructs production

Request a detailed protocol

Templates for ordering the dsDNA drop-in constructs can be found in Supplementary file 2. dsDNA drop-in constructs were ordered for production from Genewiz (‘ValueGene’ option) in pUC57 Kan vector backbone at 4 μg production scale. When lyophilized samples arrived from production, samples were resuspended in 25 μl of ddH2O at 55°C for 30 min. 19 μl was mixed with 1 μl gene-specific gRNA plasmid (25 ng/ul final concentration of gRNA plasmid). The rest was stored at −20°C for back-up purposes.

Confocal imaging of transgenic larval brains

Request a detailed protocol

Dissection and imaging were performed following the protocols in Lee et al. (2018a). In brief, fluorescence-positive 3rd instar larvae were collected in 1x PBS solution and then cut in half and inverted to expose the brain. Brains were transferred into 1.5 mL eppendorf tubes and fixed in 4% PFA in 1xPBS buffer for 20 min. Brains were then washed for 10 min three times in 0.2% PBST. Finally, samples were mounted on glass slides with 8 µL of VectaShield (VectorLabs #H-1000) and imaged at 20x zoom with a Nikon W1 dual laser spinning-disc confocal microscope.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
    Targeted gene expression as a means of altering cell fates and generating dominant phenotypes
    1. AH Brand
    2. N Perrimon
    (1993)
    Development 118:401–415.
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
    Type I repressors of P element mobility
    1. GB Gloor
    2. CR Preston
    3. DM Johnson-Schlitz
    4. NA Nassif
    5. RW Phillis
    6. WK Benz
    7. HM Robertson
    8. WR Engels
    (1993)
    Genetics 135:81–95.
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
    An exonuclease induced by bacteriophage lambda. II. nature of the enzymatic reaction
    1. JW Little
    (1967)
    The Journal of Biological Chemistry 242:679–686.
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
    Nature Biotechnology
    1. CD Richardson
    2. GJ Ray
    3. MA DeWitt
    4. GL Curie
    5. JE Corn
    (2016)
    Nature Publishing Group.
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
    New pCaSpeR P element vectors
    1. C Thummel
    2. V Pirrotta
    (1992)
    Drosophila Information Service 71:150.
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
  65. 65
  66. 66
  67. 67

Decision letter

  1. K VijayRaghavan
    Senior and Reviewing Editor; National Centre for Biological Sciences, Tata Institute of Fundamental Research, India
  2. K VijayRaghavan
    Reviewer; National Centre for Biological Sciences, Tata Institute of Fundamental Research, India

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Acceptance summary:

This is an excellent study and the revisions address the points raised by the reviewers satisfactorily. This work is part of a larger effort to create a gene disruption line for every gene in Drosophila that has a human homolog. The strategy has been to generate interruptions in the introns of genes. These interruptions consist of a 'swappable cassette' due to the presence of attp sites. Nested within the two attp sites are various payloads such as a GFP (with splice acceptor and donor), resulting a 'protein trap', or a T2A-Gal4 with a PolyA tail similarly that interrupts translation. You have generated about 1500 such MiMIC and CRIMIC lines out of the 10,000 that they intend to make. Given the scale of the endeavour, both cost and efficiency of transgenesis are important considerations. The current methods are expensive (1500 USD per line), and because they involve multiple cloning steps and injections of large constructs into embryos, the effective efficiency is low (50%).

Decision letter after peer review:

Thank you for submitting your article "An efficient CRISPR-based strategy to insert small and large fragments of DNA using short homology arms" for consideration by eLife. Your article has been reviewed by three peer reviewers, including K VijayRaghavan as the Reviewing Editor, Senior Editor, and Reviewer #1.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

This is an excellent study and should be published after revision. This work is part of a larger effort to create a gene disruption line for every gene in Drosophila that has a human homologue. Their strategy has been to generate interruptions in the introns of genes. These interruptions consist of a 'swappable cassette' due to the presence of attp sites. Nested within the two attp sites are various payloads such as a GFP (with splice acceptor and donor), resulting a 'protein trap', or a T2A-Gal4 with a PolyA tail similarly that interrupts translation. They have generated about 1500 such MiMIC and CRIMIC lines out of the 10,000 that they intend to make. Given the scale of their endeavour, both cost and efficiency of transgenesis are important considerations. The current methods they use are expensive (1500 USD per line), and because they involve multiple cloning steps and injections of large constructs into embryos, the effective efficiency is low (50%). They have now tried multiple things to reduce cost and retain/enhance the efficiency of transgenesis and in this manuscript, they lay all this tinkering out for us.

They first attempted a cloning free, PCR based strategy that utilised universal primers combined with a 100 nucleotide homology arm overhang to generate single-stranded DNA donors for their 'drop-in' cassettes. They tested this strategy with 20 genes in S2 cells, and 10 in germline transformations. They were able to generate stable S2 lines for 12 (of 20) of the genes and these they verified to be accurate in terms of their precision of insertion, and subcellular localisation of the resultant GFP fusion protein. Unfortunately, this simple and cheap strategy had much lower efficiency in germline transgenesis that the previous methods.

So they turned to tinker with a few aspects of the previous strategy, namely: reducing the size of the selectable marker; reducing the size of the polyA tail that interrupts translation; reducing the Gal4 to a 'miniGal4'; linearising the injected plasmid in vivo; shortening the size of the homology arm (100nt as opposed to 1000nt); alternate cloning strategy.

Although these were cheaper to synthesise and gave excellent efficiencies in transformation, the miniGal4 was far less active, the miniPolyA far less effective in blocking translation and the mini-selectable marker segregated away. Finally the authors determine that the old CRIMIC cassettes, but with shorter homology arms generated via an alternate cloning strategy and an in vivo linearisation of the plasmid provided the right balance between cost (1/3 the previous cost) and efficiency (though I can't tell what their efficiency is!)

This is a paper about tinkering with various strategies and it's precisely because of this that we would recommend it: It lets you look under the hood. So much work goes into such an endeavour, and we are pleased that they are going to put this out. It's not just nice, but will inform many other systems where transgenesis using similar strategies are being attempted. There's a wealth of rigorously documented what-works-what-doesn't knowledge here that typically people don't talk about.

Essential revisions:

1) For the more laborious, double-stranded DNA (dsDNA) synthetic constructs a significantly improved frequency of HDR is nicely demonstrated. However, the results for the single-stranded DNA homology donors (ssDNA) are as persuasive. The authors may want to consider repeating these experiments (in cell culture and in flies) using the approach published by Richardson and colleagues, 2016. These authors presented an asymmetric ssDNA donor design that increased the rate of HDR in human cells up to 60%. The design is based on their observation that after DNA cleavage, Cas9 first releases the PAM-distal nontarget strand before dissociation. This preferential release of the cleaved DNA seems to offer a privileged entry point for the ssDNA donor which is accounted for in an asymmetric design of the ssDNA donor (36 and 91 nucleotides homology arms, respectively). Exploring this technology in the context of Drosophila has the potential to enhance the method described by Kanca and colleagues (as already pointed out in the review by Bier et al., 2018, as well as more profoundly benchmarking their technology beyond Lee et al., 2018. We understand that this could take time and, unless the authors have already embarked on these directions, the authors should discuss the strengths and limitations of their results for ssDNA homology donors.

2) The shortened double-stranded Swappable Integration Cassette (SIC) design comes with the cost of losing the marker 3xP3-EGFP. The 3xP3-EGFP cassette is a direct measure of HDR efficiency. The replacement is the expression of a sgRNA used for the repair of a non-functional m-Cherry reporter. In this case, reporter activation is used as a co-CRISPR strategy and does not directly reflect HDR efficiency. Co-CRISPR strategies using recessive markers have been described by: Ge et al., 2016 and Kane et al., 2017.

First, the replacement of the 3xP3-EGFP marker should be described as a co-CRISPRing strategy in the text, citing the papers that developed and used this strategy previously. Second, it is unclear to the reader if the dominant marker (CRISPR-based mCherry repair) indeed has been used to establish fly lines described in the paper. Were the flies listed in Figure 4C pre-selected based on mCherry expression? How many of the non-mCherry lines were nevertheless successfully modified in the GOI? These points should be mentioned more clearly. In case it has been used, it would be interesting to know how many mCherry-positive animals resulted in successfully targeted GOIs. If data is available please add it to the supplementary information.

3) The authors make a point that their drop-in int100 CRIMIC approach to germline transformation significantly improves the methods previously used by the GDP. However, the improvement appears to be primarily due to the reduction of "cloning failures." What are "cloning failures?" Are they what most labs deal with on a routine basis by adjustments in the protocol? One can imagine that for a large-scale project avoiding such failures is highly desirable. For an individual lab, it may also be a convenience-but not necessarily a big deal. In the absence of any discussion of what a "cloning failure" is and how it might otherwise be addressed, it is hard to know whether the improvement obtained with the drop-in int100 CRIMICs is anything more than incremental.

4) The authors show that there is a reduced translation product from an interrupted gene when the miniGal4 construct is used vs. a simple STOP cassette. They discuss this difference in the context of the mPA, but the mPA seems unlikely to be the cause of the diminished expression since both constructs have it. Do the authors suppose that the sequence of the inserted construct (STOP vs. minGal) affects the mPA efficacy? A more likely alternative is that the two constructs differentially affect SA function. Do the authors have reason to rule out this possibility?

5) For the generation of ssDNA donors the authors describe using 100 nt extensions of the PCR primers. This is a very nice idea, but it would be useful to know how much this reduces the efficiency of PCR amplification and/or gives rise to non-specific products?

6) The manuscript refers to "gRNA1" but it seems unlikely to be a single sequence, otherwise the U6gRNA1 sequence would be a target of the guide RNA it produces. It looks from the information provided that the gRNA1 is actually two (slightly different) sequences designed to avoid cross-reaction when the "U6gRNA1" is expressed. This ambiguity is confusing and if the interpretation given here is correct it would be best to assign different names to the two sequences. Also, it would be useful if the authors showed a picture of an animal expressing the U6gRNA1-activated mCherry selection marker. It seems that this marker will be ubiquitously expressed and the authors should probably comment on the potential complication of using a marker that will interfere (until removal) with any other fluorescent label in the red channel.

7) Several figure and table legends are short on detail, in some cases making interpretation difficult. For example in Table 1, what does "N/A" mean? Particularly for "Insertion sequence-verified?" It seems like it either it was (Y) or wasn't (N). Also, in Figure 3B (and elsewhere) what does "Failed at Cloning" mean? And what does it mean that an event is "non-confirmed." In Figure 5A, what is the difference between mPA-1 and mPA-2? Were two versions of the mPA made?

https://doi.org/10.7554/eLife.51539.023

Author response

Essential revisions:

1) For the more laborious, double-stranded DNA (dsDNA) synthetic constructs a significantly improved frequency of HDR is nicely demonstrated. However, the results for the single-stranded DNA homology donors (ssDNA) are as persuasive. The authors may want to consider repeating these experiments (in cell culture and in flies) using the approach published by Richardson and colleagues, 2016. These authors presented an asymmetric ssDNA donor design that increased the rate of HDR in human cells up to 60%. The design is based on their observation that after DNA cleavage, Cas9 first releases the PAM-distal nontarget strand before dissociation. This preferential release of the cleaved DNA seems to offer a privileged entry point for the ssDNA donor which is accounted for in an asymmetric design of the ssDNA donor (36 and 91 nucleotides homology arms, respectively). Exploring this technology in the context of Drosophila has the potential to enhance the method described by Kanca and colleagues (as already pointed out in the review by Bier et al., 2018, as well as more profoundly benchmarking their technology beyond Lee et al., 2018. We understand that this could take time and, unless the authors have already embarked on these directions, the authors should discuss the strengths and limitations of their results for ssDNA homology donors.

We thank to reviewers for this suggestion. When Richardson and colleagues published their work we tried the asymmetric homology arm constructs for one of the targeted genes in S2 cells as well as in fly transformation. In both assays we did not notice a difference in efficiency for knock-in in Rab11 locus. However, since we did not perform a systematic study for multiple genes we did not discuss these preliminary results in our manuscript. We can mention this in the Discussion if the reviewer wishes, but we prefer not to do this as we can only speculate as to what the data mean. We therefore suggest that asymmetry may increase the efficiency of single stranded transformation, referring to Richardson et al. paper. We added the paragraph Previous studies by Richardson et al. (Richardson et al., 2016) analyzed the binding dynamics of Cas9 to target sites and showed increase in homologous recombination efficiency by using homology donors with asymmetric homology arms. Whether the use of asymmetric homology arms will increase the knock-in efficiency remains to be tested for Drop-in.” Note that given that a cassette with 3XP3-GFP combined with the full length GAL4 is more desirable for fly transformation, these constructs would be much too large to incorporate in ssDNA donors with any of the currently established methods.

2) The shortened double-stranded Swappable Integration Cassette (SIC) design comes with the cost of losing the marker 3xP3-EGFP. The 3xP3-EGFP cassette is a direct measure of HDR efficiency. The replacement is the expression of a sgRNA used for the repair of a non-functional m-Cherry reporter. In this case, reporter activation is used as a co-CRISPR strategy and does not directly reflect HDR efficiency. Co-CRISPR strategies using recessive markers have been described by: Ge et al., 2016 and Kane et al., 2017.

First, the replacement of the 3xP3-EGFP marker should be described as a co-CRISPRing strategy in the text, citing the papers that developed and used this strategy previously. Second, it is unclear to the reader if the dominant marker (CRISPR-based mCherry repair) indeed has been used to establish fly lines described in the paper. Were the flies listed in Figure 4C pre-selected based on mCherry expression? How many of the non-mCherry lines were nevertheless successfully modified in the GOI? These points should be mentioned more clearly. In case it has been used, it would be interesting to know how many mCherry-positive animals resulted in successfully targeted GOIs. If data is available please add it to the supplementary information.

We appreciate the suggestion from the reviewer. However, the gRNA1 used in our constructs as a dominant marker is not used in a Co-CRISPR context. The injected flies do not contain the Actin-mcherr-#1-ry marker that gets repaired by single strand annealing. This repair happens in the next generation in the presence of U6:gRNA1 that is integrated in the genome together with the SIC. When the U6:gRNA1 is not integrated in the genome of F1 flies, Actin-Cas9 does not result in the repair of Actin-mcherr-#1-ry. Hence, the U6:gRNA1 behaves like a dominant marker, and not like a co-CRISPR gRNA. We have clarified this in the text by adding “This SSA dependent repair reaction occurs in the F1 generation upon stable integration of U6:gRNA1 in the genome (Figure 4—figure supplement 1).”

3) The authors make a point that their drop-in int100 CRIMIC approach to germline transformation significantly improves the methods previously used by the GDP. However, the improvement appears to be primarily due to the reduction of "cloning failures." What are "cloning failures?" Are they what most labs deal with on a routine basis by adjustments in the protocol? One can imagine that for a large-scale project avoiding such failures is highly desirable. For an individual lab, it may also be a convenience-but not necessarily a big deal. In the absence of any discussion of what a "cloning failure" is and how it might otherwise be addressed, it is hard to know whether the improvement obtained with the drop-in int100 CRIMICs is anything more than incremental.

Based on years of troubleshooting the cloning protocol, for a given batch of 96 designs, we are able to amplify ~90% homology arms by PCR. We then ligate the homology arms, insertion cassette, and backbone by golden gate assembly. The best-case scenario for a given batch is ~90% successful assembly, for an overall cloning efficiency of 80%. We must emphasize that this is our highest possible success rate, and depending on the batch, we sometimes can only achieve ~50% cloning efficiency. There is also a tremendous amount of bench time that goes into successfully cloning large homology arm flanked cassettes and many areas where the cloning can fail: homology arm PCRs must be troubleshooted repeatedly; many of the assembly products are incorrect; and sequencing of the final product is often challenging and needs to be repeated several times to confirm the construct. For a lab that does not specialize in making such constructs, the failure rate and cost in time is likely even greater. With the drop-in cassette strategy, not only is the cloning success rate nearly 100%, but the protocol requires little troubleshooting, dramatically reducing the bench time.

We explain the sources of cloning failures by adding the section “The previous method required cloning large homology arm flanked cassettes has several pitfalls: homology arm PCRs often must be troubleshooted repeatedly; assembly products are often incorrect; and sequencing of the final product is often challenging and needs to be repeated to confirm the construct. […] With the drop-in cassette strategy, not only is the cloning success rate nearly 100%, but the protocol requires little troubleshooting, dramatically reducing the bench time. The transformation and verification rate result in a 70-80% success rate.”

4) The authors show that there is a reduced translation product from an interrupted gene when the miniGal4 construct is used vs. a simple STOP cassette. They discuss this difference in the context of the mPA, but the mPA seems unlikely to be the cause of the diminished expression since both constructs have it. Do the authors suppose that the sequence of the inserted construct (STOP vs. minGal) affects the mPA efficacy? A more likely alternative is that the two constructs differentially affect SA function. Do the authors have reason to rule out this possibility?

The difference in the efficiency of STOP vs. miniGAL4 is indeed puzzling. We believe that T2A in the miniGAL4 acts as a stronger translational abort signal than the stop codons. In the presence of a strong transcriptional stop signal such as the full-length SV40 polyA transcriptional stop compensates for the stop codon read through. Although the splice acceptor sequence is identical between SA-STOP and SA-T2AminiGAL4 we cannot rule out that they affect splicing at different levels. One possible explanation for possible decreased SA efficacy is the size of SIC that we integrate. We now add to the Discussion “The lower mutagenic efficacy of attP-SA-3XSTOP-minipolyA-U6gRNA1-attP may be the result of read through of stop codons, inefficient transcriptional stop at minipolyA sequence or smaller size of inserted artificial exon. […] Alternatively, the increased size of artificial exon overcomes the limitation of attP-SA-3XSTOP-minipolyA-U6gRNA1-attP construct in mutagenesis efficacy”.

5) For the generation of ssDNA donors the authors describe using 100 nt extensions of the PCR primers. This is a very nice idea, but it would be useful to know how much this reduces the efficiency of PCR amplification and/or gives rise to non-specific products?

After the optimization of annealing temperature through gradient PCR, our protocol does not suffer from noticeable decrease in PCR yield or from non-specific bands.

6) The manuscript refers to "gRNA1" but it seems unlikely to be a single sequence, otherwise the U6gRNA1 sequence would be a target of the guide RNA it produces. It looks from the information provided that the gRNA1 is actually two (slightly different) sequences designed to avoid cross-reaction when the "U6gRNA1" is expressed. This ambiguity is confusing and if the interpretation given here is correct it would be best to assign different names to the two sequences.

gRNA1 sequence is the same between the target and U6:gRNA1. The difference is the presence of PAM sequence in gRNA1 target sites and lack of it in U6:gRNA1 construct. Since CRISPR/Cas9 cut is dependent on the PAM sequence U6:gRNA1 is not cut by the gRNA is produces.

Also, it would be useful if the authors showed a picture of an animal expressing the U6gRNA1-activated mCherry selection marker. It seems that this marker will be ubiquitously expressed and the authors should probably comment on the potential complication of using a marker that will interfere (until removal) with any other fluorescent label in the red channel.

U6gRNA1-activated mCherry selection marker is indeed ubiquitously expressed but as indicated in the last paragraph of the subsection “dsDNA drop-in donors of ˂ 2kb are efficient homology donors for transgenesis”, this reconstituted actin-mCherry segregates independently from the insert. Selecting flies with knock-in allele on a balancer that do not inherit the actin-mCherry transgene on the different chromosome will ensure that it will not interfere with downstream applications. The Figure 4—figure supplement 1 contains the crossing scheme to ensure proper balancing while removing the actin-mCherry transgene from the background. Moreover, as discussed in the following section, use of int100-CRIMIC constructs remove these complications.

7) Several figure and table legends are short on detail, in some cases making interpretation difficult. For example in Table 1, what does "N/A" mean? Particularly for "Insertion sequence-verified?" It seems like it either it was (Y) or wasn't (N). Also, in Figure 3B (and elsewhere) what does "Failed at Cloning" mean? And what does it mean that an event is "non-confirmed." In Figure 5A, what is the difference between mPA-1 and mPA-2? Were two versions of the mPA made?

We appreciate the reviewers pointing this out to us. In Table 1, we used N/A to indicate where it was not relevant to evaluate the data (for example, if there were no clones, we could not sequence-verify them). However, this is unnecessarily complicated, so we have changed the table to a binary Y/N in each category, unless otherwise noted. In Figure 5A mPA1 and mPA2 are two independent fly lines that result from the injection of same mPA construct. We now indicate this in the Figure 5 legend.

https://doi.org/10.7554/eLife.51539.024

Article and author information

Author details

  1. Oguz Kanca

    1. Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    2. Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital, Houston, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Validation, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5438-0879
  2. Jonathan Zirin

    1. Howard Hughes Medical Institute, Harvard Medical School, Boston, United States
    2. Department of Genetics, Harvard Medical School, Boston, United States
    Contribution
    Conceptualization, Data curation, Methodology, Project administration, Writing—review and editing
    Competing interests
    No competing interests declared
  3. Jorge Garcia-Marques

    Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, United States
    Contribution
    Conceptualization, Resources, Writing—review and editing
    Competing interests
    No competing interests declared
  4. Shannon Marie Knight

    1. Howard Hughes Medical Institute, Harvard Medical School, Boston, United States
    2. Department of Genetics, Harvard Medical School, Boston, United States
    Contribution
    Data curation, Validation, Methodology
    Competing interests
    No competing interests declared
  5. Donghui Yang-Zhou

    1. Howard Hughes Medical Institute, Harvard Medical School, Boston, United States
    2. Department of Genetics, Harvard Medical School, Boston, United States
    Contribution
    Data curation, Visualization, Methodology
    Competing interests
    No competing interests declared
  6. Gabriel Amador

    1. Howard Hughes Medical Institute, Harvard Medical School, Boston, United States
    2. Department of Genetics, Harvard Medical School, Boston, United States
    Contribution
    Data curation, Methodology
    Competing interests
    No competing interests declared
  7. Hyunglok Chung

    1. Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    2. Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital, Houston, United States
    Contribution
    Data curation, Methodology
    Competing interests
    No competing interests declared
  8. Zhongyuan Zuo

    1. Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    2. Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital, Houston, United States
    Contribution
    Data curation, Visualization, Methodology
    Competing interests
    No competing interests declared
  9. Liwen Ma

    Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    Contribution
    Data curation, Visualization, Methodology
    Competing interests
    No competing interests declared
  10. Yuchun He

    1. Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    2. Howard Hughes Medical Institute, Baylor College of Medicine, Houston, United States
    Contribution
    Data curation, Supervision, Methodology
    Competing interests
    No competing interests declared
  11. Wen-Wen Lin

    Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    Contribution
    Data curation, Supervision, Methodology
    Competing interests
    No competing interests declared
  12. Ying Fang

    Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    Contribution
    Data curation, Methodology
    Competing interests
    No competing interests declared
  13. Ming Ge

    Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    Contribution
    Data curation, Validation, Methodology
    Competing interests
    No competing interests declared
  14. Shinya Yamamoto

    1. Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    2. Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital, Houston, United States
    3. Program in Developmental Biology, Baylor College of Medicine, Houston, United States
    4. Department of Neuroscience, Baylor College of Medicine, Houston, United States
    Contribution
    Conceptualization, Writing—review and editing
    Competing interests
    No competing interests declared
  15. Karen L Schulze

    1. Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    2. Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital, Houston, United States
    3. Howard Hughes Medical Institute, Baylor College of Medicine, Houston, United States
    Contribution
    Resources, Project administration, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-1368-729X
  16. Yanhui Hu

    1. Howard Hughes Medical Institute, Harvard Medical School, Boston, United States
    2. Department of Genetics, Harvard Medical School, Boston, United States
    Contribution
    Resources, Software, Methodology
    Competing interests
    No competing interests declared
  17. Allan C Spradling

    Department of Embryology, Howard Hughes Medical Institute, Carnegie Institution for Science, Baltimore, United States
    Contribution
    Funding acquisition, Project administration, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-5251-1801
  18. Stephanie E Mohr

    1. Howard Hughes Medical Institute, Harvard Medical School, Boston, United States
    2. Department of Genetics, Harvard Medical School, Boston, United States
    Contribution
    Conceptualization, Data curation, Investigation, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9639-7708
  19. Norbert Perrimon

    1. Howard Hughes Medical Institute, Harvard Medical School, Boston, United States
    2. Department of Genetics, Harvard Medical School, Boston, United States
    Contribution
    Supervision, Funding acquisition, Project administration, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-7542-472X
  20. Hugo J Bellen

    1. Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, United States
    2. Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital, Houston, United States
    3. Howard Hughes Medical Institute, Baylor College of Medicine, Houston, United States
    4. Program in Developmental Biology, Baylor College of Medicine, Houston, United States
    5. Department of Neuroscience, Baylor College of Medicine, Houston, United States
    Contribution
    Supervision, Funding acquisition, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    hbellen@bcm.edu
    Competing interests
    Reviewing editor, eLife
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5992-5989

Funding

NIGMS (R01GM067858)

  • Hugo J Bellen

Howard Hughes Medical Institute

  • Norbert Perrimon
  • Allan C Spradling
  • Hugo J Bellen

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

The Drosophila Gene Disruption Project is supported by NIH NIGMS R01GM067858 to HJB. Confocal microscopy at the NRI is supported by the Neurovisualization Core of the IDDRC, funded by NIH U54HD083092. We thank the Harvard Medical School Image Data Management Core for support with the OMERO platform and the Harvard Medical School Division of Immunology’s Flow Cytometry Facility for the use of their FACS equipment. We also thank Raghuvir Viswanatha, Justin Bosch, Denise Lanza and Jason Heaney for helpful discussions, Robert Levis for critical reading and editing of the manuscript and Tzumin Lee for fly stocks. Development of the tagged cell line resource at the Drosophila RNAi Screening Center (DRSC) was supported by NIH ORIP R24 OD019847 (PI: NP, Co-PI: A Simcox, Co-I: SEM). Additional relevant support for the DRSC includes NIH NIGMS R01 GM084947 and P41 GM132087 (PI: NP, Co-I: SEM). NP, ACS, and HB are investigators of Howard Hughes Medical Institute.

Senior and Reviewing Editor

  1. K VijayRaghavan, National Centre for Biological Sciences, Tata Institute of Fundamental Research, India

Reviewer

  1. K VijayRaghavan, National Centre for Biological Sciences, Tata Institute of Fundamental Research, India

Publication history

  1. Received: September 9, 2019
  2. Accepted: October 31, 2019
  3. Accepted Manuscript published: November 1, 2019 (version 1)
  4. Version of Record published: November 14, 2019 (version 2)

Copyright

© 2019, Kanca et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 8,857
    Page views
  • 1,705
    Downloads
  • 10
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

  1. Further reading

Further reading

    1. Evolutionary Biology
    2. Genetics and Genomics
    J Dylan Shropshire et al.
    Review Article

    Cytoplasmic incompatibility (CI) is the most common symbiont-induced reproductive manipulation. Specifically, symbiont-induced sperm modifications cause catastrophic mitotic defects in the fertilized embryo and ensuing lethality in crosses between symbiotic males and either aposymbiotic females or females harboring a different symbiont strain. However, if the female carries the same symbiont strain, then embryos develop properly, thereby imparting a relative fitness benefit to symbiont-transmitting mothers. Thus, CI drives maternally-transmitted bacteria to high frequencies in arthropods worldwide. In the past two decades, CI experienced a boom in interest due to its (i) deployment in worldwide efforts to curb mosquito-borne diseases, (ii) causation by bacteriophage genes, cifA and cifB, that modify sexual reproduction, and (iii) important impacts on arthropod speciation. This review serves as a gateway to experimental, conceptual, and quantitative themes of CI and outlines significant gaps in understanding CI’s mechanism that are ripe for investigation from diverse subdisciplines in the life sciences.

    1. Developmental Biology
    2. Genetics and Genomics
    Meltem Weger et al.
    Research Article

    The glucose-sensing Mondo pathway regulates expression of metabolic genes in mammals. Here, we characterized its function in the zebrafish and revealed an unexpected role of this pathway in vertebrate embryonic development. We showed that knockdown of mondoa impaired the early morphogenetic movement of epiboly in zebrafish embryos and caused microtubule defects. Expression of genes in the terpenoid backbone and sterol biosynthesis pathways upstream of pregnenolone synthesis was coordinately downregulated in these embryos, including the most downregulated gene nsdhl. Loss of Nsdhl function likewise impaired epiboly, similar to MondoA loss of function. Both epiboly and microtubule defects were partially restored by pregnenolone treatment. Maternal-zygotic mutants of mondoa showed perturbed epiboly with low penetrance and compensatory changes in the expression of terpenoid/sterol/steroid metabolism genes. Collectively, our results show a novel role for MondoA in the regulation of early vertebrate development, connecting glucose, cholesterol and steroid hormone metabolism with early embryonic cell movements.