Targeted mutagenesis in a human-parasitic nematode

Abstract

Parasitic nematodes infect over 1 billion people worldwide and cause some of the most common neglected tropical diseases. Despite their prevalence, our understanding of the biology of parasitic nematodes has been limited by the lack of tools for genetic intervention. In particular, it has not yet been possible to generate targeted gene disruptions and mutant phenotypes in any parasitic nematode. Here, we report the development of a method for introducing CRISPR-Cas9-mediated gene disruptions in the human-parasitic threadworm Strongyloides stercoralis. We disrupted the S. stercoralis twitchin gene unc-22, resulting in nematodes with severe motility defects. Ss-unc-22 mutations were resolved by homology-directed repair when a repair template was provided. Omission of a repair template resulted in deletions at the target locus. Ss-unc-22 mutations were heritable; we passed Ss-unc-22 mutants through a host and successfully recovered mutant progeny. Using a similar approach, we also disrupted the unc-22 gene of the rat-parasitic nematode Strongyloides ratti. Our results demonstrate the applicability of CRISPR-Cas9 to parasitic nematodes, and thereby enable future studies of gene function in these medically relevant but previously genetically intractable parasites.

Fig 1. A strategy for targeted mutagenesis in S. stercoralis using CRISPR-Cas9.

(A) The life cycle of S. stercoralis. iL3s enter hosts by skin penetration. The nematodes then develop into parasitic adults, which reside and reproduce in the small intestine. Their progeny exit the host in feces and develop into either iL3s or free-living adults. The free-living adults mate and reproduce in the environment, and all of their progeny develop into iL3s. Thus, S. stercoralis can develop through a single generation outside the host [10]. S. stercoralis can also complete its life cycle within a single host [9]. L1-L4 = 1^st-4^th larval stages. Adapted from Gang and Hallem, 2016 [10]. (B) The unc-22 genes of C. elegans and S. stercoralis. The Ss-unc-22 gene structure depicted is based on the gene prediction from WormBase ParaSite [24,47]. The CRISPR target sites tested and their predicted on-target activity scores are indicated [50]. Scale bar = 1 kb. (C) Plasmid vectors for the expression of Cas9 and sgRNA in S. stercoralis. (D) RNP complex assembly. Cas9 protein, crRNA targeting Ss-unc-22, and tracrRNA are incubated in vitro to form RNP complexes [20]. (E) Strategy for targeted mutagenesis in S. stercoralis. Plasmid vectors or RNP complexes were introduced into developing eggs by gonadal microinjection of free-living adult females. F₁ iL3 progeny were screened for unc phenotypes, putatively resulting from mutation of Ss-unc-22.

Fig 2. CRISPR-Cas9 targeting of the Ss-unc-22 gene results in iL3s with an uncoordinated phenotype.

(A-B) Time-lapse images of wild-type iL3s (A) vs. unc F₁ iL3s (B) swimming in a water droplet. Wild-type iL3s showed continuous rapid movement in water; unc F₁ iL3s experienced intermittent bouts of twitching, paralysis, and uncoordinated movement. For A and B, red lines indicate iL3 trajectories. Scale bars = 200 μm. (C) Swimming distance for wild-type iL3s vs. unc F₁ iL3s over a 10-s period. unc F₁ iL3s swam shorter distances relative to wild-type iL3s. ***P<0.001, Mann-Whitney test. n = 21–23 trials for each population. (D) Average crawling speed for wild-type iL3s vs. unc F₁ iL3s over a 20-s period. unc F₁ iL3s showed reduced crawling speeds relative to wild-type iL3s. **P<0.01, unpaired t test with Welch’s correction. n = 30–32 trials for each population. (E) Maximum crawling displacement for wild-type iL3s vs. unc F₁ iL3s over a 5-min period. unc F₁ iL3s traversed less distance than wild-type iL3s. **P<0.01, Mann-Whitney test. n = 26 trials for each population. For C-E, graphs show medians and interquartile ranges. unc F₁ iL3 data for B-E were obtained from plasmid vector delivery of CRISPR-Cas9 constructs at Ss-unc-22 site #1.

Fig 3. Nicotine induces twitching in unc F₁ iL3s.

(A) A nicotine assay for S. stercoralis iL3s. Free-living adult females were injected with CRISPR constructs targeting Ss-unc-22. F₁ iL3s were collected and exposed to 1% nicotine. Wild-type iL3s gradually paralyzed over the course of 8 min, whereas unc F₁ iL3s twitched continuously. Some, but not all, of the F₁ iL3s contained putative Ss-unc-22 mutations and twitched in nicotine. (B) Twitching frequency of S. stercoralis wild-type iL3s and the F₁ iL3s from microinjected females following nicotine exposure. For each condition, the Ss-unc-22 target site and delivery method of the CRISPR-Cas9 constructs are indicated. DNA = plasmid vector delivery; RNP = ribonucleoprotein complex delivery. The twitching frequency of F₁ iL3s for all Ss-unc-22 target sites and delivery methods tested differed from that of wild-type iL3s (P<0.001, chi-square test with Bonferroni correction). Instances where twitching frequency differed between target sites or delivery methods are indicated. ***P<0.001, chi-square test with Bonferroni correction. n = 446–1,314 iL3s per condition. (C) CRISPR-Cas9-mediated mutagenesis of Ss-unc-22 requires a highly specific sgRNA. Plasmid vectors for the expression of Cas9 and a sgRNA targeting S. ratti site #2 were injected into S. stercoralis. The twitching phenotype in S. stercoralis F₁ iL3s was not observed when the S. ratti version of site #2 was used. ***P<0.001, Fisher’s exact test. n = 484-677 iL3s for each condition. The alignment of S. stercoralis and S. ratti site #2 is shown with the PAM underlined. Asterisks indicate nucleotide differences between the S. stercoralis and S. ratti targets.

Fig 4. CRISPR-mediated mutagenesis of Ss-unc-22 results in putative deletion of the target locus.

(A) Representative gel of wild-type iL3s (top) or unc F₁ iL3s from RNP injections at site #3 (bottom). Genomic DNA from each iL3 was split into two reactions: ctrl. = control reaction amplifying 416 bp of the first exon of the Ss-act-2 gene to confirm the presence of genomic DNA; u22 = reaction amplifying 660 bp around site #3. Size markers = 1.5 kb, 1 kb, and 500 bp from top to bottom. (B) The Ss-unc-22 region is significantly depleted in unc F₁ iL3s. Left: relative quantity analysis of PCR products. All control bands and all u22 bands were quantified relative to their respective reference bands, denoted by asterisks in A. Values >1 indicate more PCR product than the reference while values <1 indicate less product. ***P<0.001, two-way ANOVA with Sidak’s post-test. Medians and interquartile ranges shown. Right: relative quantity of the Ss-unc-22 site #3 target region for each unc F₁ iL3 tested, and inferred genotypes. nd = PCR product not detected. (C) Whole-genome sequencing coverage plots for populations of Ss-unc-22-targeted F₁ iL3s or wild-type iL3s. A 4-kb window centered on the predicted cut site is shown [24,47]. Black lines = average coverage depth by position (reads per base); red lines = average genome-wide coverage; blue lines = average coverage for the Ss-unc-22 gene. Coverage around Ss-unc-22 site #3 is significantly depleted in both Ss-unc-22 libraries relative to the Ss-unc-22 gene average (P<0.05; see Methods). No depletion is observed in the wild-type library (P>0.05; see Methods). Gray shaded regions represent stretches of continuous significant coverage depletion around the cut site (Ss-unc-22 library A = 510 bp, Ss-unc-22 library B = 725 bp).

Fig 5. CRISPR-mediated homology-directed repair of Ss-unc-22.

(A) Strategy for HDR at Ss-unc-22 target site #2. unc F₁ iL3s that displayed both the nicotine-twitching phenotype and red fluorescence were selected as candidates for HDR and were genotyped using the primer sets indicated. 5’ and 3’ integration primer pairs amplify only following successful integration of Ss-act-2::mRFPmars into site #2. HA = homology arm. (B) Representative DIC + epifluorescence overlays of unc F₁ iL3s expressing Ss-act-2::mRFPmars. Top, iL3 expressing mRFPmars (sparse expression indicated by the arrow) from an extrachromosomal array. Bottom, iL3 expressing mRFPmars following HDR, showing near-uniform mRFPmars expression in the body wall. For both images, anterior is to the left. Scale bar = 50 μm. (C) Representative genotypes of a wild-type iL3 and unc F₁ iL3s expressing mRFPmars. Genomic DNA from individual iL3s was split into four reactions: ctrl. = control reaction amplifying 416 bp of the first exon of the Ss-act-2 gene to confirm the presence of genomic DNA; wt = reaction for the wild-type locus of site #2 where primer R1 overlaps the predicted CRISPR cut site; 5’ = reaction for insertion of the 5’ border of the integrated cassette; 3’ = reaction for insertion of the 3’ border of the integrated cassette. For genotypes: array = red unc F₁ iL3s that showed no evidence of integration; integrated = red unc F₁ iL3s with successful HDR. Some integrated iL3s had putative homozygous disruptions of Ss-unc-22 site #2 (e.g. iL3s #4 and #7, which lacked the wt band). Asterisks indicate genotypes for iL3s shown in B. Size markers = 2 kb, 1.5 kb, 1 kb, and 500 bp from top to bottom.

Fig 6. The unc phenotype is heritable following host passage.

(A) Strategies for heritable transmission of Ss-unc-22 mutations. Gerbil hosts were infected with either all wild-type iL3s, a 50/50 mix of unc and wild-type F₁ iL3s, or unc-enriched F₁ iL3s. F₂ and F₃ progeny were collected from host feces and screened for unc phenotypes. Note that iL3s collected from host feces can be the F₂ or F₃ generation depending on whether they developed into iL3s directly, or after a free-living generation (Fig 1A) [10]. (B) Twitching frequency of wild-type control progeny and F₂ or F₃ progeny collected from unc infections. The twitching frequency of the F₂ or F₃ iL3s collected from the mixed unc infection differed from that of wild-type iL3s. ***P<0.001, chi-square test with Bonferroni correction. n = 1,908–3,849 iL3s per condition. The twitching frequency of F₂ adults collected from the unc-enriched infection differed from that of wild-type adults. ***P<0.001, chi-square test with Bonferroni correction. n = 164–332 adults per condition. The twitching frequency of F₂ or F₃ iL3s collected from the unc-enriched infection differed from that of wild-type iL3s. ***P<0.001, chi-square test with Bonferroni correction. n = 2,694–3,849 iL3s per condition. (C) Swimming distance for wild-type iL3s vs. unc F₂ or F₃ iL3s over a 10-s period. unc iL3s swam shorter distances than wild-type iL3s. ***P<0.001, Mann-Whitney test. n = 15–16 worms for each population. (D) Mean crawling speed for wild-type iL3s vs. unc F₂ or F₃ iL3s over a 20-s period. unc iL3s showed reduced crawling speeds relative to wild-type iL3s. ***P<0.001, unpaired t test with Welch’s correction. n = 16 worms for each population.