A Combination of CRISPR/Cas9 and Standardized RNAi as a Versatile Platform for the Characterization of Gene Function

Abstract

Traditional loss-of-function studies in Drosophila suffer from a number of shortcomings, including off-target effects in the case of RNA interference (RNAi) or the stochastic nature of mosaic clonal analysis. Here, we describe minimal in vivo GFP interference (miGFPi) as a versatile strategy to characterize gene function and to conduct highly stringent, cell type-specific loss-of-function experiments in Drosophila miGFPi combines CRISPR/Cas9-mediated tagging of genes at their endogenous locus with an immunotag and an exogenous 21 nucleotide RNAi effector sequence with the use of a single reagent, highly validated RNAi line targeting this sequence. We demonstrate the utility and time effectiveness of this method by characterizing the function of the Polymerase I (Pol I)-associated transcription factor Tif-1a, and the previously uncharacterized gene MESR4, in the Drosophila female germline stem cell lineage. In addition, we show that miGFPi serves as a powerful technique to functionally characterize individual isoforms of a gene. We exemplify this aspect of miGFPi by studying isoform-specific loss-of-function phenotypes of the longitudinals lacking (lola) gene in neural stem cells. Altogether, the miGFPi strategy constitutes a generalized loss-of-function approach that is amenable to the study of the function of all genes in the genome in a stringent and highly time effective manner.

Keywords: CRISPR; Drosophila; loss-of-function; stem cell.

Figure 1

miGFPi as a versatile strategy to characterize gene function. (A) Schematic drawing of GSC and NSC lineages. GSCs and NSCs divide asymmetrically to generate another stem cell and a differentiating daughter cell. Differentiating daughter cells have limited proliferative potential. GSC differentiating daughter cells divide four times with incomplete cytokinesis to generate a 16-cell cyst. NSCs generate ganglion mother cells (GMCs) or intermediate neural progenitor cells (INPs) that generate postmitotic neurons after a limited number of divisions. (B) Outline of the miGFPi strategy. CRISPR/Cas9-mediated insertion of an oligo into the coding sequence of a gene. The oligo contains one component encoding for an immunotag (e.g.: V5 or HA) and one component constituting an RNAi target sequence derived from eGFP. When homozygous, this chromosome allows for RNAi-mediated loss-of-function of the gene. When heterozygous, the nontagged allele of the gene is unaffected by RNAi and preserves gene function, serving as a control. (C) miGFPi allows for simultaneous loss of function of numerous genes into which the oligo has been inserted. (D) miGFPi allows for loss-of-function studies of individual or combinations of isoforms from a gene. (E) Potential applications of the two components of the oligo-based tag. CRISPR, clustered regularly interspaced short palindromic repeats; eGFP, enhanced GFP; GFP, green fluorescent protein; GSC, germline stem cell; iGFPi, in vivo GFP interference; LOF, loss-of-function; miGFPi, minimal in vivo GFP interference; NSC, neural stem cell; RNAi, RNA interference; shRNA, short hairpin RNA.

Figure 2

miGFPi unravels a requirement for Tif-1a in GSC maintenance. (A) miGFPi strategy for Tif-1a loss-of-function and rescue experiments (Tif-1a wild-type, S610A, and S610D sequences are shown at the C-terminus). (B) Relevant miGFPi genotypes for loss-of-function experiments. (C) Tif-1a is required for germline stem cell maintenance. Compared to the wild type (right), miGFPi-mediated depletion of Tif-1a results in a loss of all germline cells and empty ovarioles (white arrowheads). Blue (gray lower panels), DNA; Red, 1B1 [fusome and spectrosome marker (undifferentiated cells)], Green: Vasa (germline cell marker). CRISPR, clustered regularly interspaced short palindromic repeats; eGFP, enhanced green fluorescent protein; GSC, germline stem cell; iGFPi, in vivo GFP interference; miGFPi, minimal in vivo GFP interference; shRNA, short hairpin RNA.

Figure 3

Rescue of the miGFPi-induced phenotypes. (A) Outline of the miGFPi genotypes for the Tif-1a rescue experiments. (B–D) Expression of a Tif-1a wild-type and Tif-1a S610D construct in the miGFPi Tif-1a background fully rescues the Tif-1a loss-of-function phenotype, whereas the miGFPi Tif-1a S610A ovaries contain empty ovarioles (white arrowheads) (for details see text). iGFPi, in vivo GFP interference; miGFPi, minimal in vivo GFP interference; shRNA, short hairpin RNA; WT, wild type.

Figure 4

MESR4 controls differentiation in the GSC lineage. (A) miGFPi strategy for MESR4 loss-of-function. (B) Loss of MESR4 function in the germline is associated with a decreased size of the ovaries and differentiation defects of germline cells [pseudo egg chambers containing undifferentiated cells (red arrowheads), polyploid nurse cells (yellow arrowheads), and maturing oocytes (white arrowhead)]. (C and E) The number of 1B1 positive cells in the germarium is expanded in MESR4 miGFPi. (C) Green, 1B1; magenta, DNA. (E) Red: manual outline of nuclei of 1B1 positive cells of a MESR4 miGFPi and wild-type germarium. (D) The oocyte-accumulating protein Orb does not enrich in specific cells in MESR4 miGFPi as compared to the control (yellow arrowheads). (F) One copy of the V5::miGFPi::MESR4 over MESR4⁷⁹ and nanos-Gal4-driven iGFPi shRNA results in the same phenotype as MESR4 miGFPi. (G) Identification of a CRISPR/Cas9-generated MESR4 mutant. MESR4 mutants die at pupal stages. The genomic region flanking the gRNA was PCR amplified from homozygous MESR4⁷⁹ larvae and sequencing revealed a deletion of the start codon. The genotypes are: left larva, MESR4⁷⁹/MESR4⁷⁹; right larva, MESR4⁷⁹/CyO- Kr-Gal4 UAS-GFP. (H) MESR4 is expressed in the germline. V5 staining of homozygous V5::miGFPi::MESR4 reveals low expression of MESR4 in GSCs and early cystocytes. MESR4 is strongly induced in 16-cell cystocytes and remains highly expressed at later stages of differentiation yellow arrowheads: GSCs, white arrowheads: differentiating cells. (I) Endogenously, N-terminally GFP tagged MESR4 in Drosophila S2 cells localizes predominantly to the nucleus and is weakly detectable in the cytoplasm (blue, DNA; green, GFP::MESR4). CRISPR, clustered regularly interspaced short palindromic repeats; eGFP, enhanced GFP; GFP, green fluorescent protein; GSC, germline stem cell; iGFPi, in vivo GFP interference; miGFPi, minimal in vivo GFP interference; RNAi, RNA interference; shRNA, short hairpin RNA.

Figure 5

miGFPi allows for simultaneous knockdown of multiple genes with one validated shRNA line. (A) Outline of the miGFPi genotypes for the Tif-1a, MESR4 double mutant experiments. (B–D) Combined loss of Tif-1a and MESR4 function in the germline using miGFPi results in the Tif-1a phenotype and a loss of all germline cells. Expression of Tif-1a S610A in the Tif-1a, MESR4 double miGFPi background results in the Tif-1a phenotype, whereas expression of Tif-1a WT results in the MESR4 loss-of-function phenotype, demonstrating that miGFPi effectively depletes both genes. Red, 1B1; blue/gray, DNA. (E and F) qPCR validation of miGFPi mediated knockdown of Tif-1a and MESR4 using gene-specific primers (E) and miGFPi allele-specific (F) primers. Error bars = SD. *** P < 0.001; n.s., not significant. Student’s t-test. GFP, green fluorescent protein; iGFPi, in vivo GFP interference; miGFPi, minimal in vivo GFP interference; qPCR, quantitative PCR; shRNA, short hairpin RNA; WT, wild type.

Figure 6

miGFPi allows for isoform-specific loss-of-function analysis. (A) miGFPi strategy for lola-all and simultaneous lola-PB and lola-PD loss-of-function. (B) lola is expressed in the larval brain. V5 staining of V5::miGFPi::lola-all, V5::miGFPi::lola-PB, and V5::miGFPi::lola-PD, and HA staining of V5::miGFPi::lola-PB and HA::miGFPi::lola-PD. (C) lola function is required for NSC differentiation. Depletion of all lola-isoforms or the simultaneous loss of the two most abundant NSC lola-isoforms, lola-PB and lola-PD (right), results in a NSC overproliferation phenotype at the posterior side of the larval central brain compared to the wild type (left). Individual loss-of-function of lola-PB and lola-PD does not induce NSC overproliferation, suggesting a complex interaction of these isoforms in NSCs. lola can be detected in the cells of the inner optic anlagen (IOA), where wor-Gal4 is not expressed. Red, Miranda; blue, lola-all; gray, V5. eGFP, enhanced GFP; GFP, green fluorescent protein; iGFPi, in vivo GFP interference; miGFPi, minimal in vivo GFP interference; NSC, neural stem cell; shRNA, short hairpin RNA.