Functional editing of endogenous genes through rapid selection of cell pools (Rapid generation of endogenously tagged genes in Drosophila ovarian somatic sheath cells)

Abstract

The combination of genome-editing and epitope tagging provides a powerful strategy to study proteins with high affinity and specificity while preserving their physiological expression patterns. However, stably modifying endogenous genes in cells that do not allow for clonal selection has been challenging. Here, we present a simple and fast strategy to generate stable, endogenously tagged alleles in a non-transformed cell culture model. At the example of piwi in Drosophila ovarian somatic sheath cells, we show that this strategy enables the generation of an N-terminally tagged protein that emulates the expression level and subcellular localization of the wild type protein and forms functional Piwi-piRNA complexes. We further present a concise workflow to establish endogenously N-terminally and C-terminally tagged proteins, and knockout alleles through rapid selection of cell pools in fly and human models.

Figure 1.

A universal strategy for simple and rapid genomic editing of cell populations. (A) Drosophila ovarian somatic sheath cells (OSC) represent a unique but delicate model to study Piwi–piRNA mechanism ex vivo. OSC represent somatic follicle cells (FC) of the Drosophila ovary, and express one of the three Drosophila PIWI proteins, Piwi. PiRNAs are generated from long piRNA cluster transcripts by the endonuclease Zucchini (Zuc). Mature Piwi–piRNA silencing complexes transition into the nucleus, recognize nascent transposon transcripts by base-pairing complementarity and induce epigenetic silencing. (B) Endogenous tagging of piwi in OSC. An sgRNA was designed to target the endogenous piwi gene in the vicinity of the start codon (ATG). The donor construct for homologous repair contained a FLAG-HA (FH)-tag and a puromycin resistance gene. The FH-tag was fused in frame with piwi's open reading frame (ORF) to generate an endogenously N-terminally tagged protein (eFH-). The puromycin resistance was placed into a synthetic intron and transcribed from the opposite genomic strand. The edited allele is designed to express two independent transcripts: The piwi transcript remains under the control of the endogenous promoter and contains an additional intron and a tag. The mature modified mRNA differs from the wt piwi mRNA only by an additional exon-exon junction and the Flag-HA tag. The second transcript is independently generated from the opposite genomic strand and produces an mRNA encoding a puromycin resistance under the control of an Actin promoter. (C) Rapid and simple generation of stably edited OSC:eFH-piwi. Ovarian somatic sheath cells (OSC) were transfected with an expression plasmid for the sgRNA, the Cas9 endonuclease, and the donor plasmid. Cells were treated with SCR7, an inhibitor of non-homologous end joining (NHEJ) to increase the probability for homologous repair. Antibiotic selection with Puromycin (Puro) was started 48 h after transfection. After 2–3 weeks, a puromycin resistant cell population has repopulated the dish

Figure 2.

Characterization of the genomic edit and the resulting eFH-piwi transcript. (A) Schematic representation of the wild type and the edited piwi allele. The priming sites for universal and gene-specific primers that were used for genotyping and cDNA characterization are indicated. (B) Genotyping of OSC:eFH-piwi reveals a heterozygous editing event. PCR with the indicated primers (A) was performed on genomic DNA (gDNA). gDNA from WT OSC was used as control. (C) Characterization of eFH-piwi transcripts indicate accurate splicing of the synthetic intron. PCR was performed on complementary DNA (cDNA) using the indicated primers. Primers were designed to detect the unspliced and spliced transcript (A). The donor plasmid and gDNA served as control for the unspliced transcript.

Figure 3.

eFH-Piwi protein emulates the expression and subcellular localization of wt Piwi in OSC. (A) Heterozygous OSC:eFH-piwi expresses WT Piwi and eFH-Piwi protein to similar levels. Wt Piwi and eFH-Piwi were detected with an anti-Piwi antibody. Different amounts of cell extracts were analyzed as indicated. Protein quantification was performed by western blotting using fluorescent antibodies and the LI-COR Odyssey technology for accurate automated quantification. (B) Confocal microscopy images of Piwi and eFH-Piwi in OSC. Piwi was detected in wild type (WT) and edited (OSC:eFH-piwi) using a polyclonal anti-Piwi antibody (green). The edited eFH-Piwi was specifically detected in OSC:eFH-piwi using a monoclonal anti-HA antibody. DAPI was used to visualize nuclei. (C) Detection of Piwi in OSC:eFH-Piwi cells. Counterstain with an anti-Tubulin monoclonal antibody and DAPI.

Figure 4.

eFH-Piwi associates with piRNAs to form mature piRNA silencing complex. (A) eFH-Piwi was specifically immunopurified (IP) from OSC:eFH-piwi using an anti-Flag antibody. (B) Small RNAs were extracted from the purified Piwi–piRNA complexes and prepared for Illumina sequencing. 3′ and 5′ adaptors were sequentially ligated to the small RNAs before reverse transcription and PCR amplification. A total of 10 unique molecular identifiers (UMI: 10N) was accommodated in the ligated adaptors and allowed for removal of PSC duplicates during analyses. (see also Supplemental Figure S5). (C) PiRNAs associated with eFH-Piwi exhibit the same phased 1U-signatures as WT Piwi–piRNAs, indicating biogenesis by the Zuc-processor complex. Metagene analysis of uniquely mapping piRNAs aligned at their 5′ end across an extended genomic interval. The observed piRNA population is indicated as colored box. Nucleotide frequencies are shown across a 100 nt interval. Both piRNA populations show the characteristic patterns of phased processing by the Zucchini processor complex indicated by a preference for Uridine in the first position (1U), and both one piRNA length upstream (-26) and one piRNA length downstream (26) of the observed piRNAs. (D) Length profiles of piRNAs associated with eFH-Piwi and WT Piwi in nucleotides (nt). Both piRNA populations show a length distribution characteristic for Piwi–piRNAs. (E) eFH-Piwi–piRNAs, like WT Piwi–piRNAs are enriched for sequences that are antisense to annotated repeats (rmsk). More than half of either piRNA population represents sequences with antisense complementarity to transposons and other repeat elements (repeat masker, rmsk). The orientation with respect to the matching feature is indicated (sense, s; antisense, as)

Figure 5.

N-terminal tagging of endogenous protein coding genes using an intronic antibiotic resistance for rapid selection of stable cell pools. (A) sgRNAs were designed to target the 5′end of Shutdown (Shu) and Gle1 for N-terminal tagging. (B) After antibiotic selection with Puromycin, cell pools were probed for the expression of the endogenously tagged proteins by western blotting using an anti-HA antibody. An anti-Lamin antibody was used as loading control. Differences in the expression of the endogenously tagged proteins agree with differences in their mRNA expression, Shu (72 rpkm) and Gle1 (11 rpkm) (modENCODE PRJNA75285, rpkm … reads per kilobase per million).

Figure 6.

C-terminal tagging of endogenous proteins using a promoter-less selection marker. (A) sgRNAs were designed to cleave the target gene close to the STOP codon. The donor template for homologous repair (HR) removed the STOP codon to generate an in-frame fusion with a 3xFLAG-3XHA tag followed by P2A self-cleaving peptide and a Puromycin resistance (PuroR). Upon successful homologous repair, the endogenously tagged allele is designed to express and extended messenger RNA (mRNA) that produces two proteins, the C-terminally tagged target protein and a Puromycin Resistance. (B) We targeted Virilizer (Vir) and Nibbler (Nbr) as examples for C-terminal tagging. The sgRNAs used to target the endogenous genes are shown and the position to the coding sequence (CDS) and the STOP codon (TGA). (C and D) After antibiotic selection for 2–3 weeks, Drosophila ovarian somatic sheath cells (OSC) were probed for the expression of the tagged target protein. (C) Endogenously FH-tagged Vir (Vir-eFH) was detected in edited cells with the expected molecular weight of ∼220 kDa using an anti-Flag antibody for western blotting (WB). (D) To probe the reproducibility and flexibility of our method, C-terminally tagged Nibbler cells (Nbr-eFH) were established simultaneously with either one of two different sgRNAs (sgRNA 1 and 2) or a combination of both. After 2–3 weeks of selection with Puromycin, all three combinations resulted in stable cells that expressed endogenously tagged Nibbler (Nbr-eFH) with an expected molecular weight of ∼80 kDa.

Figure 7.

Insertion of multiple different selection alleles enables effective knock-out. (A) To disrupt the transcriptional unit of ANKRD1 in human embryonic kidney cells (HEK293), we designed an sgRNA upstream of the translational START codon. We first edited one allele by insertion of a Puromycin Resistance (PuroR) driven by the eukaryotic elongation factor-1 alpha promoter (Prom(EF1a)). After two weeks of selection with Puromycin, we performed a second round of editing to insert a Hygromycin resistance (HygroR) into the remaining wild type (WT) allele. (B) After double selection with Puromycin and Hygromycin, we probed the presence of the wild type and edited alleles by genomic PCR (position of the primers indicated in A). We observed both edited alleles with the expected size difference for the Puromycin and Hygromycin cassettes in the edited cells. The wild type (WT) allele was only observed in the WT cells suggesting a complete knock-out of ANKRD1. (C) In an alternative approach, we omitted the EF1a promoter in the donor cassette and instead inserted the coding sequence (CDS) of the selection marker in frame with ATG of the endogenous gene. Upon successful edit, the endogenous promoter drives the antibiotic resistance instead of the endogenous gene. Homology arms were designed to eliminate parts of the target-gene. We targeted YTH N6-Methyladenosine RNA Binding Protein 2 (YTHDF2) as an example. We co-transfected two different donor plasmids and selected for biallelic editing event with Puromycin and Hygromycin for three weeks. (D) We probed genomic DNA for the presence of the edited alleles. We observed a remaining wild type (WT) allele in addition to the two edited alleles suggesting that HEK293 cells are triploid (or partly triploid) for YTHDF2. To eliminate the third allele, we inserted an additional Neomycin Resistance (NeoR) using the same strategy as for PuroR and HygroR. After triple selection with Puromycin, Hygromycin and Neomycin, our genomic PCR showed the three integrated selection alleles and loss of the wild type allele suggesting a functional knock-out.