The coding sequence of dnase2_1590 was amplified from N. brasiliensis cDNA omitting the signal peptide and stop codon and cloned into into pPICZalpha-A downstream of the coding sequence for the Saccharomyces cerevisiae a-mating secretion factor and in frame with an N-terminal myc and 6xHis Tag. PCR was carried out using Q5 polymerase (New England Biolabs) according to manufacturer’s recommendations with 500 nM of the following primers (lower case indicating nucleotides added for cloning purposes, restriction site underlined):

F-5´-aagctGAATTCGGTCTGAGTTGCAAGAACATGGAGG-3´

R-5-ttttgtTCTAGAGCGGTTTTGTTTGTCTTCTTGCTCG-3´

Following transformation of Pichia pastoris X-33, protein expression was optimised for single colonies and scaled up following the EasySelect Pichia expression protocol (Invitrogen). His-tagged proteins were purified from yeast supernatants by nickel chelating chromatography and protein concentration determined by Bradford assay.

A polyclonal antiserum to N. brasiliensis recombinant DNase II was raised by subcutaneous immunisation of a rat with 100 µg protein emulsified in alum, followed by 3 boosts of 50 µg protein via the same route at weeks 4, 6 and 8, and the animal bled at week 9. Parasite proteins were prepared from larval and adult parasites by homogenising in phosphate buffered saline, 0.25 mM n-dodecyl-β-D-maltoside and protease inhibitors for somatic extracts, whereas secreted proteins were collected by in vitro culture in serum-free RPMI-1640 plus antibiotics and concentrated in vivaspin concentrators with a 3 kDa cutoff, as previously described (Huang et al., 2010). Western blotting was performed via standard procedures following resolution of 5 µg parasite proteins by SDS-12% polyacrylamide gel electrophoresis and blotting to polyvinylidene difluoride membrane. Rat anti-DNase II was used at 1:500 dilution, rabbit anti-rat IgG-horseradish peroxidase (Sigma) used as secondary antibody and the blot visualised using enhanced chemiluminescence western blotting detection reagents (Amersham Bioscience).

For facilitated cloning of synthesised sgRNA oligos, a NheI restriction site was introduced into the RGR expression vector (pL-5LTR-RGR(DMD#1)-AmCyan-A, Addgene plasmid #138482). To do this, the entire sequence between two SpeI restriction sites flanking the RGR region was amplified by PCR adding an NheI and AvrII restriction site to the 3´ end and used to replace the original sequence via SpeI and AvrII to allow for cloning of gRNA coding regions via KpnI and NheI. PCR was carried out in 50 μl reactions using Q5 polymerase (New England Biolabs) according to manufacturer’s recommendations with 1 ng of the RGR plasmid as template and 500 nM of the following primers (lower case indicating nucleotides added for cloning purposes and restriction site underlined): F-5´-GCTTGCATGCCGACATGGATTATTGACTAGTCCC-3´; R-5´-attgaCCTAGGGCTAGCTCTAGAGCGGCCGTCCCATTCGCCATGC-3.

The CRISPR/Cas-derived RNA-guided endonucleases (RGEN) algorithm was used to predict Streptococcus pyogenes (Sp) Cas9 gRNA targets with a 5´-NGG-3´ Protospacer Adjacent Motif (PAM) in exon 3 of the dnase-2 gene and subsequent off-target screening against the N. brasiliensis genome. Oligonucleotides encoding gRNAs for subsequent integration into the RGR plasmid via KpnI and NheI were synthesized (GeneArt, Thermo Fisher Scientific) with the following structure: 5´-KpnI-inverted first 6 nt of the guide RNA-HH ribozyme-guide RNA-gRNA scaffold-HDV ribozyme-NheI-3´ (complete sequence Supplementary Figure 3 ). Plasmids were maintained in NEB stable Escherichia coli (New England Biolabs). Positive transformants were selected on LB agar plates containing 50 μg ml^-1 ampicillin. Constructs were verified for error-free integration of transgenes by routine Sanger sequencing (Eurofins Genomics).

NanoMEDIC were produced in HEK293T cells maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) at 37°C, 10% foetal calf serum (FCS), 2 mM L-glutamine, 100 units ml^-1 penicillin and 100 μg ml^-1 streptomycin, as described previously (Gee et al., 2020) with some modifications. In brief, per well of a 6-well plate, 3 x 10⁶ cells were transfected with 1.25 μg of gRNA-encoding plasmid, 1.25 μg of pHLS-EF1a-FRB-SpCas9-A (Addgene plasmid #138477), 1.25 μg of pHLS-EF1a-FKBP12-Gag^HIV (Addgene plasmid #138476), 250 ng of pcDNA1- Tat^HIV (Addgene plasmid #138478) and 500 ng of pMD2.G (VSV-G) (Addgene plasmid #12259), using Lipofectamine 2000 at a ratio of 1:2.5 (Life Technologies). After 16 hours, the transfection medium was replaced with 2 ml of reduced serum culture medium (5% FCS) supplemented with 300 μM A/C heterodimerization agent (formerly AP21967, Takara Bio), 20 mM HEPES and 10 μM cholesterol (balanced with methyl-β-cyclodextrin, Sigma) (Chen et al., 2009). VSV-G-EVs were produced in HEK293T cells following transfection of 3 x 10⁶ cells with 1.25 μg of pMD2.G per well of a 6-well plate with Lipofectamine 2000 at a ratio of 1:2.5. The cell culture supernatant was changed after 18 hours as for NanoMEDIC, omitting the heterodimerisation agent. After an additional incubation of 48 hours at 37°C and 10% CO₂, the EV-containing cell supernatant was harvested, centrifuged at 2,000 x g for 20 min at 4°C and passed through a 0.45 μm Acrodisc syringe filter. NanoMEDIC and VSV-G-EVs were generally concentrated using 10 kDa vivaspin columns and washed twice with serum-free growth medium (NanoMEDIC) or trehalose buffer (10% trehalose in PBS) (VSV-G-EVs). Following centrifugation at 4,000 x g for 15 min at 4°C, the flow-through was discarded and EVs gently resuspended in 2 ml of the respective buffer before aliquoting into cryovials and storage at -80°C. For enrichment by polymer-based precipitation, 4 ml of Lenti-X concentrator (Takara Bio) was added to 12 ml of cell supernatant and the mixture incubated at 4°C with gentle agitation for 18 hours. Precipitated EVs were then pelleted by centrifugation at 3000 x g for 45 min and 4C and resuspended in 2 ml of trehalose buffer.

The composition of EVs was analysed by Western blot and Nano-flow cytometry. Western blotting was performed via standard procedures following resolution of 12 µl concentrated EV preparations by SDS-12% polyacrylamide gel electrophoresis under reducing (Cas9, VSV-G) or non-reducing (CD63, CD81) conditions. Primary antibodies were used at 1:1000 dilution: mouse anti-human CD63 (clone H5C6, Biolegend); mouse anti-human CD81 (TAPA-1) (clone 5A6, Biolegend); rabbit anti-VSV-G (Covance); mouse anti-CRISPR (Cas9) (clone 7A9, Biolegend).

Concentrated NanoMEDIC preparations were analysed for their nanoparticle content by nano-flow cytometry using a NanoAnalyzer calibrated against trehalose buffer. Before acquisition, samples were diluted 1:100 and 1:200 in trehalose buffer. The concentration of particles with diameters larger than 100 nm was determined following gating using the NanoFCM™ Silica Nanospheres Cocktail (S16M-Exo, diameter:68~155 nm) as a standard.

Oligonucleotides encoding the HDR templates ( Supplementary Figure 3 ) were synthesized (Invitrogen) and reconstituted in nuclease-free water at a concentration of 1 µg µl^-1. VSV-G-EVs were then loaded with ssDNA as described previously (Lamichhane et al., 2015). In brief, 5 µg of ssDNA were added to 95 µl of VSV-G-EVs in trehalose buffer and EVs then transferred to a 1 mm electroporation cuvette (BioRad) placed on ice. EVs were electroporated by exponential decay with two pulses at 400 V and 125 µF using a GenePulser Xcell electroporator (Bio-Rad). Cuvettes were placed on ice immediately after electroporation and incubated on ice for 20 min. EVs were then transferred to fresh microfuge tubes and the cuvette washed with one volume (100 µl) of RPMI and added to the tube. To alleviate aggregation, EDTA was added to a final concentration of 1 mM and EVs incubated at room temperature for 15 min, gently resuspended several times during incubation (Lamichhane et al., 2015).

N. brasiliensis were maintained in male SD rats, and infective larvae isolated from faecal cultures using a Baermann apparatus. Larvae were activated to feed as previously described (Huang et al., 2010) for 48 to 72 hrs in RPMI1640, 0.65% glucose, 2 mM L-glutamine, 100 U ml^-1 penicillin, 100 μg ml^-1 streptomycin, 100 μg ml^-1 gentamicin, 20 mM HEPES, 2% rat serum (worm culture medium), then washed twice in serum-free medium prior to exposure to EVs. Per well of a 12-well plate, approximately 3,000 - 4,000 activated L3 were exposed to 200 µl NanoMEDIC and/or 200 µl of VSV-G-EVs, volumes adjusted to 1 ml with serum-free medium and 10 μg ml^-1 polybrene (Sigma) and 200 µg ml^-1 gentamicin added. EV preparations in control worms were substituted with HEK293T cell supernatant from untransfected cells. Following incubation for 18 to 24 hrs, worms were transferred to a 15 ml tube and washed twice in 10 ml of warm serum-free worm medium containing 1 mM EDTA with centrifugation at 150 x g for 1 min between washes. Worms were then resuspended in 2 ml complete worm medium (or serum-free medium when testing for ES products) and incubated for another 24 to 48 hours at 37°C, 5% CO₂.

Homology directed repair (HDR) was effected utilising a single stranded oligodeoxynucleotide with ~50 nt homology arms, which introduced a series of 6 stop codons following a single (sgRNA91) or double (sgRNA46 and 91) stranded DNA break. Detection of HDR was performed by PCR analysis of genomic DNA following exposure of activated L3 to NanoMEDIC with or without the addition of HDR-containing EVs. Worms were washed twice in 10 ml of PBS at 37°C, 72 hrs post exposure to EVs, and genomic DNA isolated using the DNeasy Blood and Tissue DNA extraction kit (Qiagen). PCR was performed using Q5 polymerase with an annealing/extension temperature of 72°C. Primer pairs were designed to amplify fragments from the adjacent exon (2 or 4) and the HDR region (6xstop). The following primers were used with resultant fragment sizes indicative of editing. Exon 2 forward primer (5´-GATTCGGCTATTGGTGCAACTGTTAAGC-3´) and 6xstop reverse primer (5´-ATCCCCGTGCTACTCAGTTACCTAGTCACTTA-3´) gave rise to a 240 bp fragment with a single stranded DNA break (sgRNA91) and a 200 bp fragment with a double stranded DNA break (sgRNA46 and 91). The 6xstop forward primer (TAAGTGACTAGGTAACTGAGTAGC-3´) and exon 4 reverse primer (5´-TCGAGCCTGATTCGGGGTAGTCG-3´) gave rise to a 165 bp fragment in both cases. Exon 3 forward primer (5´-ACCTCAAAATTGCCTACAACGAC-3´) and 6xstop reverse primer gave rise to a 125 bp fragment with a single stranded DNA break (sgRNA91) and an 85 bp fragment with a double stranded DNA break (sgRNA46 and 91). PCR was also performed with primers specific to exon 3 flanking the HDR integration site: exon 3 forward primer as indicated above and exon 3 reverse primer (5´-GGAATCTTGGCACACTGTGTACCAGC-3´). Amplification of eif-3c (eukaryotic translation initiation factor 3 subunit C) was performed to control for genomic DNA integrity (forward primer 5´-GAACACGTTGTAGCTGCGTCA-3´ and reverse primer 5´-AATAGGTTCTCAGCGATTCCGTT-3´).

PCR was performed using Q5 polymerase with an annealing/extension temperature of 72°C using exon 3 forward (5´-ACCTCAAAATTGCCTACAACGAC-3´) and reverse (5´-GGAATCTTGGCACACTGTGTACCAGC-3´) primer to amplify the WT allele and the edited allele, and the product was purified using gel extraction. DNA libraries for deep sequencing were made using the NEBNext Ultra II DNA library preparation kit (New England Biolabs) according to the manufacturer’s instructions. Sequencing was performed by the London Institute of Medical Sciences Genomics facility. Reads were mapped to the predicted PCR products from wild type and edited alleles using Bowtie2 (Langmead and Salzberg, 2012). SNPs and indels were mapped using samtools mPileup (Li et al., 2009). We then used Varscan2 (Koboldt et al., 2012) using the commands pileup2snp and pileup2indel to identify putative SNPs and indels respectively, requiring a minimum average quality of 20 and setting a minimum frequency of 1 in 10000 to enable low frequency alterations to be detected. Data were read into R using the read.table function and comparison of the different files enabled SNPs that were present in the untransfected control to be identified and removed as likely sequencing errors or PCR artefacts. Line plots indicating the frequency of SNPs at each position divided by total reads aligned to that position were constructed to illustrate the distribution of mutations along the template.

Total RNA was extracted using TRIreagent (Sigma) and converted to cDNA using an iScript cDNA kit (Biorad) following removal of contaminating genomic DNA by DNAse I. Reverse Transcription PCR (RT-PCR) was carried out using Q5 DNA polymerase (New England Biolabs) according to the manufacturer’s recommendations. An edited dnase2 transcript was confirmed with an exon 2-specific forward primer (5´-GATTCGGCTATTGGTGCAACTGTTAAGC-3´) and a reverse primer binding the 6xSTOP region in the HDR template (5´-ATCCCCGTGCTACTCAGTTACCTAGTCACTTA-3´), resulting in a product of approximately 140 bp. RT-PCR amplification of eif-3c was performed to control for RNA integrity (forward primer 5´-GAACACGTTGTAGCTGCGTCA-3´ and reverse primer 5´-AATAGGTTCTCAGCGATTCCGTT-3´).

Real-Time quantitative PCR (RT-qPCR) was carried out with a Step-One PLUS Fast Real-time PCR cycler (Applied Biosystems) under standard fast cycling conditions using PowerUP SYBR Green PCR Master Mix (Applied Biosystems) and 250 nM target gene specific forward and reverse primers. PCR amplification efficiencies were established for each primer pair (Pfaffl, 2001) and ranged between 1.9 and 2.1. Cycle threshold (Ct) values of target genes were normalised to the geometric mean of eif-3C (NBR_0001150401; eukaryotic translation initiation factor 3 subunit C) and idhg-1 (NBR_0000658601; isocitrate dehydrogenase gamma) (Hagen et al., 2021) and calibrated to the mean untreated control (wild type) samples for relative quantification by the comparative Ct method (Livak and Schmittgen, 2001). The primers were (forward (F), reverse (R)): Nb-nuc-1 F: 5´- TGACGAACCATACAACGGCA-3´, R: 5´- TGGAACACTGTGGATCAGCC-3´; eif-3C F: 5´-GAACACGTTGTAGCTGCGTCA-3´, R: 5´-AATAGGTTCTCAGCGATTCCGTT-3´; idhg-1 F: 5´-CAGAAATTGGGAGACGGCCT-3´, R: 5´-CCGAGAAACCAGCTGCATAGA-3´; dnase2_1590_E6F: 5´-TGGAAACTTGGAGAAACGGTGCTG-3´; dnase2_1590_E7R: 5´-ACATCTCCGATACAAACTAGGGGCTCC-3´.

Supernatant collected from worms cultured in serum-free medium was thawed and all reactions prepared on ice. Per test sample, 500 ng of a plasmid DNA substrate in 10 µl nuclease free water was placed in PCR tubes and a 10 µl droplet of worm supernatant transferred to the side of the tube. Reactions were initiated by centrifugation and samples placed immediately into a PCR cycler pre-warmed to 37°C and incubated for 2 to 10 min. After heat inactivation at 75°C for 10 min, 4 µl of agarose loading dye was added and 10 µl of the sample separated on a 1% agarose gel.

To assess the dynamics of DNase secretion by L3, worms were extensively washed in serum-free medium, counted and resuspended at 2500 L3 ml^-1. Per time point tested, 80 µl of worm suspension (200 L3) was added to 20 µl of plasmid DNA solution (2 µg in 20 µl RPMI). Following incubation for 15, 30, 60 or 120 min at 37°C, 5% CO2, 80 µl of supernatant was carefully aspirated and transferred to a PCR tube. Samples were immediately frozen at -20°C. For analysis, tubes were placed directly from the freezer into a PCR cycler prewarmed to 75°C and heat inactivated for 10 min. Loading dye was added and the samples separated on a 1% agarose gel.

Treatment groups were analysed for significant differences with the Kruskal-Wallis test and Dunn’s post-hoc test in relation to the control group. Values are expressed as the median with range or the mean ± SEM, and significant differences were determined using GraphPad Prism. P values of <0.05 were considered significant, *p<0.05.

To assess susceptibility of CRISPR-mediated gene editing in N. brasiliensis, we chose secreted DNase II as a target gene (Bouchery et al., 2020), as endonuclease activity in secreted products provides a means for functional analysis of gene expression. While the DNase II has been described to be secreted from infective larvae (Sotillo et al., 2014), a profile of expression in different stages has not been determined. Because a shift in temperature to 37°C acts as a cue for exsheathment and initiation of feeding in infective larvae (L3) (Huang et al., 2010), we first analysed levels of DNase II transcripts and secretion of active enzyme in L3 prior to and during activation by culture at 37°C in the presence of rat serum. Incubation at the elevated temperature led to a sharp increase in DNase II transcripts relative to those of unactivated L3, i.e. isolated directly from faecal cultures in a state of diapause, peaking at an approximate 1000-fold increase after 2-3 days ( Figure 1A ). Transcript levels were drastically reduced in adult worms compared with activated L3, but remained approximately 5-fold above those in non-activated L3 ( Figure 1A ). Western blot analysis of parasite secreted products confirmed that the DNase II was released predominantly by activated L3 ( Figure 1B ). Given the detection of low levels of DNase II transcripts in non-activated L3, there was a possibility that the protein might be pre-synthesised and stored in secretory vesicles allowing for immediate release upon entering the host. To assess this, we exposed non-activated and activated L3 to plasmid DNA during culture at 37°C. However, non-activated L3 did not secrete sufficient amounts of DNase II for effective degradation of DNA within the first 2 hours of incubation at 37°C, although complete degradation was achieved following overnight incubation ( Figure 1C ). In contrast, degradation of DNA in the culture supernatant was observed immediately following exposure to previously activated L3.

The genomic DNA sequence of DNase II was identified in NBR_scaffold_0001590 following alignment of the published cDNA sequence (GenBank: M938457) (Bouchery et al., 2020) to the genomic N. brasilensis database PRJEB511 available on WormBase ParaSite. Nine exons were defined. Exon 3 and exon 7 both encode HxK motifs; this is characteristic of DNase II as the two motifs are required to form the active site (Schafer et al., 2007) ( Figure 2A ). Notably, while the corresponding amino acid sequence derived from NBR_scaffold_0001590 and M938457 was mainly conserved, some differences were detected in the cDNA sequences which could affect the prediction of effective sgRNAs ( Supplementary Figures 1A, B ). Expression of NB_Dnase2_1590 was confirmed in our laboratory strain of N. brasiliensis following cloning of the cDNA sequence into a yeast expression vector and subsequent sequencing such that sgRNA design was based on NB_Dnase2_1590.

For transient delivery of Cas9/gRNA complexes, we adapted a recently developed NanoMEDIC approach established in mammalian cells using VSV-G-pseudotyped extracellular vesicles (Gee et al., 2020). To predict sgRNA target sequences, the Cas-Designer (RGEN) algorithm was used, which allows for off-target screening against the N. brasiliensis genome (Bae et al., 2014a; Park et al., 2015). We chose exon 3 as a target region as it encodes one of the two HxK motifs of functional DNases (Schafer et al., 2007) ( Figure 2A ). While some gRNAs were predicted that covered the HxK region, these did not achieve a frameshift prediction score over 66, which is recommended in order to minimise the possibility of unwanted in-frame deletions (Bae et al., 2014b). The two highest scoring gRNAs, guide 46 and 91, flanking the HxK motif in exon 3, were cloned between the self-cleaving Hammerhead (HH) and hepatitis delta virus (HDV) ribozymes of the transfer plasmid (ribozyme-guide-ribozyme, RGR) ( Figure 2B ). If the deletion was insufficient or frameshift not achievable using either guide, then the EV approach would allow for multiplexing to include both guide RNAs deleting the entire HxK motif coding sequence. Corresponding VSV-G-pseudotyped EVs containing RNP complexes (NanoMEDIC) were produced in HEK293T cells following transfection with the RGR and four packaging plasmids encoding Cas9, Gag, VSV-G and Tat ( Figure 2C ) (Gee et al., 2020).

The NanoMEDIC delivery system utilises two distinct homing mechanisms to recruit Cas9 and a sgRNA into extracellular vesicles in HEK293 producer cells. Streptococcus pyogenes (Sp)Cas9 is selectively packaged into extracellular vesicles by chemical-induced dimerisation. This is mediated by the FKB12 and FRB (rapamycin-binding domain) protein dimerisation system (Putyrski and Schultz, 2012). Cas9 is modified with FRB, whereas FKB12 is fused to HIV Gag protein, and addition of the rapalog (rapamycin analogue) ligand A/C mediates heterodimerisation of FRB-Cas9 with FKB12-Gag and packaging into budding extracellular vesicles. Long terminal repeat (LTR)-driven transcription of the RGR transfer plasmid produces a virus-like mRNA encoding the sgRNA and containing a packaging signal sequence (Ψ, Figure 2B ) which directs RNA incorporation into budding vesicles. The sgRNA encoded in virus-like mRNA is liberated in NanoMEDIC by self-cleavage of the flanking ribozymes. Dissociation of the complexes occurs after dilution of the A/C heterodimerizer once the vesicles fuse with recipient cell membranes, in this case those of N. brasiliensis.

To test whether homology directed repair (HDR) could be achieved following induction of double strand breaks, we generated a single-strand oligonucleotide (ssODN) template encoding a series of 6 stop codons interspaced by single nucleotides to allow for all possible open reading frames and ~50 nt homology arms ( Figure 2A ), as previously described (Ittiprasert et al., 2019). The introduction of premature stop codons allows for degradation of modified transcripts by nonsense-mediated decay and/or premature termination of translation (Hug et al., 2016). Overexpression of VSV-G in HEK293T cells leads to an increased production of VSV-G-expressing EVs, termed ‘gesicles’, that can be utilised for transfer of membrane, cytoplasmic and nuclear proteins (Mangeot et al., 2011). We therefore loaded VSV-G-pseudotyped EVs with the ssODN template and tested whether they could be transferred to N. brasiliensis L3 ( Figure 2D ) (Lamichhane et al., 2015; Meyer et al., 2017).

In our first series of experiments, activated L3 were exposed to NanoMEDIC containing sgRNAs binding at nucleotide 46 and/or 91 of exon 3 in the presence or absence of EVs containing the STOP_ssODN for HDR. Parasite genomic DNA was then assessed by PCR for modifications introduced by HDR ( Figures 3A, B ). PCR with primers binding the stop codon region and adjacent exon 2 or 4 led to amplification of expected PCR products ( Figure 3C ), indicating that NanoMEDIC in conjunction with EV-delivered ssODN templates could be used for site directed mutagenesis in N. brasiliensis. Furthermore, the presence of modified dnase2 transcripts was confirmed by RT-PCR with an exon 2-specific forward primer and a reverse primer binding the 6xSTOP region, resulting in a product of approximately 140 bp ( Figure 3D ). PCR from genomic DNA with the same primer pair produced an amplicon of approximately 200 bp ( Figure 3C ). Importantly, the presence of modified cDNA provided evidence that NanoMEDIC and EV-delivered ssODN can reach tissues actively expressing DNase II following ingestion.

In previous studies with Strongyloides stercoralis, non-homologous end joining (NHEJ) of CRISPR/Cas9-induced double-strand breaks following microinjection of RNP complexes into the gonad resulted in large deletions, while smaller insertions or deletions (indels) appeared to be absent. Similarly, PCR of genomic DNA carried out with an exon 2-binding forward and an exon 4-binding reverse primer resulted in a single fragment of ~350 bp in all samples tested, which could not resolve possible indels ( Figure 3C ). To compensate for the possibility that deletions were insufficient, or frameshift not achievable using either guide, further experiments were carried out using a combination of both guide RNAs, with or without the addition of a ssODN allowing for deletion of the entire motif. To facilitate detection of smaller changes, a further PCR was carried out with primers flanking the double stranded break sites of exon 3 ( Figure 3E ). While multiple PCR fragments were generated in the NanoMEDIC + 6xSTOP_ssODN sample indicating integration of the HDR template, PCR with exon 3-binding forward and reverse primers still failed to resolve possible indels in worms exposed to NanoMEDIC ( Figure 3E ).

To examine potential low frequency errors induced by CRISPR/Cas9-mediated editing, we performed deep sequencing of PCR products derived from worms transfected with CRISPR/Cas9 either with or without the repair template. In worms transfected with CRISPR/Cas9 alone, after removing potential sequencing errors and PCR artefacts (see methods) single nucleotide polymorphisms (SNPs) at low frequency were detected mapping to the wild type allele of the DNase II gene, indicating that the enzyme could introduce breaks triggering error prone repair ( Figure 4A ). Interestingly, addition of the repair template led to an altered pattern of mutations in the wild type allele, without changing the frequency ( Figure 4B ). Similar patterns were seen for small indels mapping to the wild type allele ( Figures 4C, D ). However, the total frequency of SNPs and indels mapping to the edited allele was extremely low and similar to the frequency of mutations mapping to the wild type allele ( Figures 4E, F ), confirming that the homology directed repair was mostly accurate.

We next assessed whether modifications introduced into the DNase II gene resulted in measurable reduction of secreted enzyme. Reduced DNase activity was observed in supernatants collected for 3 days following exposure of activated L3 to NanoMEDIC and ssODN ( Figure 5A ). While complete degradation of donor DNA was observed 5 min after exposure to supernatant from control worms, this required longer (10 min) incubation with supernatant of worms exposed to NanoMEDIC and the STOP_ssODN. Delayed DNA degradation was only observed following delivery of the STOP_ssODN via VSV-G-EVs (ssODN-EV), while direct electroporation of NanoMEDIC with the STOP_ssODN (Nano+ssODN) was unsuccessful. Reduced nuclease activity was also not observed in supernatants from worms exposed to NanoMEDIC and Mock-EVs electroporated with a ssODN encoding an irrelevant sequence of the same length as the HDR template.

To assess the timeframe of delayed DNA degradation by modified worms, secreted products were assessed for their nuclease activity by adding a substrate plasmid DNA to the culture medium alongside the worms and following degradation over time ( Figure 5B ). No intact DNA was detected after 30 min, and complete degradation recorded 60 min after culture with worms exposed to ssODN_EVs only. In contrast, co-culture of NanoMEDIC-exposed worms revealed the presence of some intact plasmid DNA after 30 min. This was more pronounced when worms were exposed to NanoMEDIC + ssODN_EVs, with some larger DNA fragments still present 60 min after co-culture ( Figure 2B ). Furthermore, RT-qPCR revealed a reduction in dnase2 transcript levels by ~25% (mean log2 ± SEM = -0.45 ± 0.12) in NanoMEDIC + ssODN_EV samples, compared to the ssODN_EV only group (-0.003 ± 0.09) (p=0.048, Kruskal-Wallis test with Dunn’s post-hoc; Figure 5C ), indicating nonsense-mediated decay of modified transcripts. No significant decrease of dnase2 transcripts (-0.17 ± 0.09) was recorded in the NanoMEDIC-only group (p=0.74). These data, together with PCR analysis of genomic DNA, indicate that the majority of NanoMEDIC-induced gene disruption can be repaired by non-homologous end joining, and that editing is enhanced by homology-directed repair.

Because N. brasiliensis DNase II has been demonstrated to degrade neutrophil extracellular traps, it has been suggested to facilitate migration of L3 through the skin and lung tissues of their mammalian host (Bouchery et al., 2020). The moderate silencing of the dnase2 gene achieved in this study did not lead to a reduction in worm numbers in the intestines of infected mice ( Figure 5D ). Nevertheless, these data provide proof of principle that CRISPR/Cas9-induced gene editing can be achieved in infective larvae of N. brasiliensis by harnessing extracellular vesicle-mediated delivery of RNPs and HDR templates, providing a new route for genetic manipulation of parasitic nematodes.

As reduction of secreted DNase II activity did not result in altered recovery of adult worms from infected mice, we hypothesised that additional DNase II activities may be encoded in the parasite genome providing some degree of redundancy. A search of WormBase ParaSite revealed that NBR_0000088201 encoded an orthologue of lysosomal DNase II from Caenorhabditis elegans termed NUC-1 (Hedgecock et al., 1983; Wu et al., 2000). The predicted amino acid sequence is shown in Figure 6A , aligned with that of the N. brasiliensis secreted DNase II (NBR_00001590) (Bouchery et al., 2020) and C. elegans NUC-1, revealing 65% identity between the mature N. brasiliensis and C. elegans NUC-1 proteins. Examination of transcript levels by real-time RT-PCR (qPCR) revealed that Nb_nuc-1 is expressed at fairly constant levels through to adult worms ( Figure 6B ). Transcript levels for the secreted DNase II were 10-fold higher than Nb_nuc-1 in activated L3, and at least 10-fold lower than Nb_nuc-1 in resting L3 and adult worms ( Figure 6C ). This suggests that Nb_nuc-1 has more of a housekeeping role, consistent with a lysosomal function, and it is notable that no nuclease activity was detected in culture supernatants of L3 within the first 2 hours following activation ( Figure 1C ) despite appreciable levels of Nb_nuc-1 transcripts. In contrast, expression of the secreted DNase II was almost exclusively associated with larval activation, suggesting that this is the major or sole enzyme released into the mammalian host and primarily responsible for degradation of extracellular/environmental DNA in the early stages of infection.