Optimizing Systems for Cas9 Expression in Toxoplasma gondii

Abstract

CRISPR-Cas9 technologies have enabled genome engineering in an unprecedented array of species, accelerating biological studies in both model and nonmodel systems. However, Cas9 can be inherently toxic, which has limited its use in some organisms. We previously described the serendipitous discovery of a single guide RNA (sgRNA) that helped overcome Cas9 toxicity in the apicomplexan parasite Toxoplasma gondii, enabling the first genome-wide loss-of-function screens in any apicomplexan. Even in the presence of the buffering sgRNA, low-level Cas9 toxicity persists and results in frequent loss of Cas9 expression, which can affect the outcome of these screens. Similar Cas9-mediated toxicity has also been described in other organisms. We therefore sought to define the requirements for stable Cas9 expression, comparing different expression constructs and characterizing the role of the buffering sgRNA to understand the basis of Cas9 toxicity. We find that viral 2A peptides can substantially improve the selection and stability of Cas9 expression. We also demonstrate that the sgRNA has two functions: primarily facilitating integration of the Cas9-expression construct following initial genome targeting and secondarily improving long-term parasite fitness by alleviating Cas9 toxicity. We define a set of guidelines for the expression of Cas9 with improved stability and selection stringency, which are directly applicable to a variety of genetic approaches in diverse organisms. Our work also emphasizes the need for further characterizing the effects of Cas9 expression.IMPORTANCE Toxoplasma gondii is an intracellular parasite that causes life-threatening disease in immunocompromised patients and affects the developing fetus when contracted during pregnancy. Closely related species cause malaria and severe diarrhea, thereby constituting leading causes for childhood mortality. Despite their importance to global health, this family of parasites has remained enigmatic. Given its remarkable experimental tractability, T. gondii has emerged as a model also for the study of related parasites. Genetic approaches are important tools for studying the biology of organisms, including T. gondii As such, the recent developments of CRISPR-Cas9-based techniques for genome editing have vastly expanded our ability to study the biology of numerous species. In some organisms, however, CRISPR-Cas9 has been difficult to implement due to its inherent toxicity. Our research characterizes the basis of the observed toxicity, using T. gondii as a model, allowing us to develop approaches to aid the use of CRISPR-Cas9 in diverse species.

Keywords: CRISPR; Cas9; Toxoplasma gondii; genome editing.

FIG 1

A construct for improved stability of Cas9 expression. (A) Schematic of Cas9 expression constructs. The triangle indicates the site of 2A-mediated peptide bond skipping. Expected molecular weights of processed and unprocessed proteins are indicated. Following transfection and chloramphenicol selection of these constructs in wt parasites, the resulting populations were subcloned. Populations after drug selection and clones were further assayed. (B) Representative immunofluorescence images of methanol-fixed intracellular parasites. Cells were stained for FLAG. ALD provides a cytosolic parasite stain, and Hoechst stain was used to stain host and parasite DNA. The white triangle marks a vacuole with aberrant, cytosolic FLAG localization. The bar graph shows the corresponding quantification of parasite vacuoles positive for FLAG staining. Bar, 10 μm. CAT, chloramphenicol acetyltransferase. ALD, aldolase. Mean ± SEM, n = 3 biological replicates. Data were arcsine transformed prior to performing an unpaired t test. ***, P ≤ 0.001. (C) Immunoblots showing processing of transcriptionally linked constructs into separate polypeptides in isolated Cas9⁺ clones. Actin serves as a loading control. (D) The efficiency of SAG1 disruption in wt and isolated Cas9⁺ clones was measured via immunofluorescence microscopy following different treatments. Mean ± SEM, n = 2 independent experiments; n.s., nonsignificant; n.d., not detected; pyr, pyrimethamine.

FIG 2

Genome targeting increases the rate of Cas9 construct integration. (A) Experimental outline of the fluorescence-based competition assays used in this study. Representative flow cytometry pseudocolor density plots are shown for parasites assessed at time points before, 1 day after, and 16 days after transfection of the constructs depicted above. (B) Bar graphs display the proportion of mNeonGreen- and mTagBFP-positive cells at day 1 posttransfection and mixing and day 16 following continuous drug selection. A Cas9 expression construct carrying sgRNA #1 outcompeted one lacking an sgRNA. (C) Results of an analogous competition in the presence or absence of sgRNA #1; here, selecting parasites for dCas9 expression. (D) sgRNA #1 is predicted to target Cas9 to the NHE1 3′ UTR in type I strains of T. gondii. The orange hexagon denotes the position of the stop codon. (E) Names and genomic targets of sgRNAs used in this study. (F) sgRNA #1 was highly favorable over an sgRNA with no genomic target (sgRNA #2) when selecting for Cas9 expression. (G) In the presence of sgRNA #1, a construct with Cas9 nuclease activity improved selection. (H) Whole-genome sequencing of clones derived from three independent transfections using a Cas9 construct expressing sgRNA #1 showed integration of the expression construct at the NHE1 3′ UTR. (I) Targeting alternative intergenic loci (sgRNAs #3 and #4) is comparable to targeting the NHE1 locus via sgRNA #1. Mean ± SEM, n = 2 biological replicates. Quantitative data were arcsine transformed prior to performing an unpaired t test. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; n.s., nonsignificant.

FIG 3

Long-term fitness cost of carrying Cas9 in the absence of an sgRNA. (A) Using improved constructs, three independent populations of Cas9-expressing parasites were generated for constructs carrying sgRNA #1 (group A) or lacking an sgRNA (group B). (B) Flow cytometry of drug-selected stable populations showed that parasites expressed the expected fluorescence marker. (C) Expression of functional Cas9 was confirmed by immunofluorescence microscopy, assessing the efficiency of SAG1 gene disruption 24 h after transfection of a SAG1-targeting sgRNA. (D) All drug-selected populations generated in the presence of sgRNA #1 outcompeted those that lacked an sgRNA. (E) Group B populations were transfected with complementation constructs that encode the red fluorescent protein mRuby3 and sgRNA #1 or a nonfunctional, scrambled sequence. Drug-selected populations were competed against their parental population. For competition data, mean ± SEM, n = 2 biological replicates. Data were arcsine transformed prior to performing an unpaired t test. **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001; n.s., nonsignificant.

FIG 4

The fitness cost of Cas9 expression in the absence of an sgRNA persists in clonal populations. (A) Clonal populations were derived from populations of groups A and B. (B) Immunoblot showing Cas9 expression among clones A1c to A3c and B1c to B3c. Cas9 expression was detected by blotting for FLAG. Actin serves as a loading control. (C) Expression of functional Cas9 was confirmed by immunofluorescence microscopy, assessing the efficiency of SAG1 gene disruption 24 h after transfection of a SAG1-targeting sgRNA. (D) Clonal populations largely recapitulated the results of their parental populations, in that group A clones generally outcompeted clones of group B. Only clone A3c was equally as fit as or less fit than group B clones. Mean ± SEM, n = 2 biological replicates. Data were arcsine transformed prior to performing an unpaired t test. **, P ≤ 0.01; n.s., nonsignificant. (E) Whole-genome sequencing was performed on group A and B clones and the parental wt clone. Numbers of SNVs relative to the parental wt clone occurred at a normal rate and were similar between clones of groups A and B. No SNVs were group specific.