Efficient Gene Knockout and Knockdown Systems in Neospora caninum Enable Rapid Discovery and Functional Assessment of Novel Proteins.

The development of molecular genetics has greatly enhanced the study of the biology and pathology associated with parasites of the phylum Apicomplexa. While the molecular tools are highly developed for the apicomplexan Toxoplasma gondii, the closely related parasite Neospora caninum lacks efficient tools for genetic manipulation. To enable efficient homologous recombination in N. caninum, we targeted the Ku heterodimer DNA repair mechanism in the genomic reference strain, Nc-Liverpool (NcLiv), and show that deletion of Ku80 results in a destabilization and loss of its partner Ku70. Disruption of Ku80 generated parasites in which genes are efficiently epitope tagged and only short homology regions are required for gene knockouts. We used this improved strain to target novel nonessential genes encoding dense granule proteins that are unique to N. caninum or conserved in T. gondii. To expand the utility of this strain for essential genes, we developed the auxin-inducible degron system for N. caninum using parasite-specific promoters. As a proof of concept, we knocked down a novel nuclear factor in both N. caninum and T. gondii and showed that it is essential for survival of both parasites. Together, these efficient knockout and knockdown technologies will enable the field to unravel specific gene functions in N. caninum, which is likely to aid in the identification of targets responsible for the phenotypic differences observed between these two closely related apicomplexan parasites. IMPORTANCE Neospora caninum is a parasite with veterinary relevance, inducing severe disease in dogs and reproductive disorders in ruminants, especially cattle, leading to major losses. The close phylogenetic relationship to Toxoplasma gondii and the lack of pathogenicity in humans drives an interest of the scientific community toward using N. caninum as a model to study the pathogenicity of T. gondii. To enable this comparison, it is important to develop efficient molecular tools for N. caninum, to gain accuracy and save time in genetic manipulation protocols. Here, we have developed base strains and protocols using the genomic reference strain of N. caninum to enable efficient knockout and knockdown assays in this model. We demonstrate that these tools are effective in targeting known and previously unexplored genes. Thus, these tools will greatly improve the study of this protozoan, as well as enhance its ability to serve as a model to understand other apicomplexan parasites.

INTRODUCTION

Neospora caninum is a coccidian that belongs to a diverse group of parasitic protozoans of the phylum Apicomplexa. N. caninum is closely related to Toxoplasma gondii, which is perhaps the best studied member of the phylum and an important human pathogen of immunocompromised patients and congenitally infected neonates. While T. gondii infects a wide range of warm-blooded vertebrates, including an estimated one-third of the world’s human population (1), N. caninum has a narrower host range, primarily targeting ruminants and dogs (2). It is one of the leading causes of bovine abortions in the world, inducing significant economic losses to cattle raising (3).

One strategy routinely used to gain a better understanding of the biology of these parasites is based on the development of a robust set of molecular approaches that enable reliable analyses of specific genetic targets. For functional analyses, reverse genetics methods were made possible by the advent of transfection protocols in a number of model apicomplexans (4, 5). This advance led to the development of an array of genetic tools for detailed genetic manipulation of these parasites. These tools are best developed and easiest to use in T. gondii, which includes an assortment of promoters, selectable markers, fluorescent proteins, genome editing strategies, and knockout and conditional knockdown approaches (6).

For N. caninum, however, this toolkit lags substantially behind its more prominent, closely-related parasite, T. gondii. Although transfection of Neospora has been available for decades, most genetic approaches took advantage of vectors developed specifically for T. gondii (7, 8). While those tools are useful for the stable insertion of heterologous genes of interest (GOI) (9–11), further development of the model is needed. For example, while a background strain with a single point mutation on the hypoxanthine-xanthine-guanine phosphoribosyl transferase gene (Nc1Δhxgprt) has been developed as a selectable marker (12–14), it was generated by chemical mutagenesis, which likely generated many mutations and is only available in the NC-1 strain, not the sequenced reference strain NcLiv.

HXGPRT has been used successfully as a selectable marker for stable transformation in T. gondii for over 20 years, due to its safety and versatility conferred by the positive or negative selection with mycophenolic acid/xanthine (Mpa/X) or 6-thioxanthine (6TX), respectively (15). Other available selectable markers include chloramphenicol acetyltransferase (CAT) and a mutated version of dihydrofolate reductase-thymidylate synthase (DHFR) driven by either Toxoplasma or endogenous Neospora promoters. These markers have been effectively used for gene insertions and endogenous tagging, based on selection strategies using resistance to chloramphenicol and pyrimethamine, respectively (16, 17).

The advent of genome editing tools based on clustered regularly interspaced short palindromic repeats-associated gene 9 (CRISPR/Cas9) system has dramatically changed each system in which it has been developed. The CRISPR/Cas9 system was adapted for T. gondii (18, 19) and has contributed to the field by promoting fast and precise targeted gene disruption, endogenous epitope tagging, and site-specific insertion of selectable markers. Without further adaptation, plasmids with CRISPR/Cas9 components used in T. gondii were satisfactorily used to disrupt GFP in N. caninum tachyzoites stably expressing the fluorescent marker, as well as disrupt the NcGRA7 gene by insertion of a selectable marker (20). Since then, pU6 plasmids—adapted or not with N. caninum promoters—have been employed for endogenous epitope tagging and gene disruption, especially of genes encoding dense granule proteins (21–25). Those strategies typically rely on large homology flanks (over 800 bp) for homology-directed repair or disruption by insertion of selectable markers, which is a setback compared with the stage of development of the CRISPR/Cas9 system found in the Toxoplasma model.

The key difference in Toxoplasma is the combination of an efficient CRISPR/Cas9 system in the background of a knockout of the Ku80 gene, which enables highly efficient recombination (26, 27). The Ku heterodimer is conserved in eukaryotes and is formed by two subunits, Ku70 and Ku80, that maintain genomic integrity through binding to DNA double-strand breaks, making repairs by the non-homologous end joining (NHEJ) pathway (28). In its absence, random integration is greatly diminished, which allows for a more precise manipulation of the targeted loci through homologous recombination. In this work, we aimed to further improve the molecular toolbox of the Neospora model system. To do this, we disrupted Ku80 in the NcLiv strain and subsequently disrupted the HXGPRT selectable marker used for Ku80 disruption, thereby recycling this marker for downstream selections. We examined the effect of the absence of Ku80 on its partner Ku70 and demonstrated that CRISPR/Cas9 based endogenous gene tagging and knockouts become efficient similar to that observed in T. gondii. We additionally modified this strain for conditional knockouts by adapting the auxin-inducible degron system. We demonstrate the utility of these strains by disrupting two secreted GRA proteins and conditionally disrupting a novel nuclear protein. Thus, these new tools dramatically improve the molecular toolkit of N. caninum.

RESULTS

Disruption of HXGPRT in the NcLiv strain using CRISPR/Cas9.

To aid in the genetic manipulation of N. caninum, we first sought to disrupt the selectable marker HXGPRT using CRISPR/Cas9 (Fig. 1A). To do this, we generated sequences encoding a guide RNA (gRNA) against Neospora’s HXGPRT (NcHXGPRT, NCLIV_038170) in the pU6 universal plasmid and transfected it into wild-type NcLiv parasites (18, 19). Knockouts of NcHXGPRT were negatively selected with 6TX and cloned by limiting dilution, and a clone was chosen that failed to grow in positive selection (Mpa/X). Sequencing of the coding region showed a single nucleotide deletion at the gRNA site within the HXGPRT locus (Fig. 1A), and this strain was designated NcLivΔhxgprt.

FIG 1

Disruption of HXGPRT and Ku80 in NcLiv strain parasites. (A) Diagram showing CRISPR/Cas9 targeting of HXGPRT and resulting deletion of a single nucleotide rendering parasites sensitive to Mpa/X. (B) Schematic and IFA of 3xHA endogenous gene tagging of Ku80. Ku80 localizes to the nucleus as assessed by Hoechst costain, and IMC12 is used to mark the periphery of the parasites. (C) Disruption of tagged Ku80 using CRISPR/Cas9 and a homology template with the selectable marker HXGPRT. IFA shows the lack of staining in the NcLivΔku80 (+HXGPRT) strain. (D) PCR verification of the Ku80 knockout shows the selectable marker is in the Ku80 locus (primers P47/P48) and the coding region is eliminated in the knockout (primers P47/P49; P50/P51). (E) Diagram and (F) PCR verification (primers P47/P52) showing the elimination of HXGPRT using negative selection with 6TX to generate parasites lacking both Ku80 and HXGPRT (denoted NcLivΔhΔk).

Epitope tagging and disruption of Ku80 in N. caninum.

To improve the efficiency of homologous recombination in the NcLivΔhxgprt strain, we then targeted the Ku80 gene (NCLIV_056045), which has been shown to dramatically improve recombination in T. gondii (26, 27). We first used CRISPR/Cas9 to add a 3xHA epitope tag at the C-terminus of Ku80 as well as the selectable marker DHFR, which resulted in parasites with nuclear staining as expected (NcLivKu803xHA; Fig. 1B).

To disrupt the Ku80 gene in NcLivKu803xHA background, we created a gRNA against the beta-barrel domain of Ku80 and a homology directed repair template that consisted of the Ku80 5′ and 3′ flanking regions, cloned upstream and downstream of a Toxoplasma HXGPRT selection cassette (Fig. 1C). These were transfected into the tagged parasites and selected, and clonal isolates were screened for the absence of Ku80 staining. One isolate was selected and the correct replacement of Ku80 with HXGPRT was verified by PCR (Fig. 1D). To improve the utility of this NcLivΔku80 (+TgHXGPRT) strain for downstream analyses, we used a similar strategy of a gRNA and homologous recombination to delete the T. gondii HXGPRT selectable marker from the Ku80 locus using negative selection (6TX; Fig. 1E). The deletion of TgHXGPRT was confirmed by PCR and the strain lacking both Ku80 and HXGPRT was designated NcLivΔhΔk (Fig. 1F).

Ku70 is destabilized in NcLivΔhΔk parasites.

Ku80 works in conjunction with its partner protein Ku70 to facilitate nonhomologous end joining (28). To assess whether disruption of Ku80 had an effect on its partner, we epitope tagged Ku70 in both NcLivΔhxgprt and NcLivΔhΔk strain parasites (Fig. 2A). Immunofluorescence assay (IFA) of the NcLivΔhxgprt strain containing wild-type Ku80 showed a nuclear localization for Ku70, as expected (Fig. 2B). However, in NcLivΔhΔk strain parasites, the staining of Ku70 was dramatically decreased (Fig. 2C) and in many parasites was undetectable (not shown). The impact on Ku70 was confirmed by Western blot analysis, which showed a nearly complete loss of detectable signal (Fig. 2D). These data indicate that disruption of Ku80 results in the concomitant destabilization and degradation of Ku70.

FIG 2

Disruption of Ku80 destabilizes its partner Ku70. (A) Diagram and IFA of Ku70 tagged with and without Ku80. Ku70 is nearly undetectable in the Δku80 parasites. Parasites were costained with Hoechst to stain the nucleus and anti-IMC12 to label the periphery of the parasites. (B) Western blot showing that Ku70 is dramatically reduced in parasites lacking Ku80. IMC12 is used as a loading control.

Highly efficient gene tagging in Δku80 parasites identifies novel secreted and nuclear factors.

Disruption of Ku80 in T. gondii results in a dramatic increase in the efficiency of CRISPR/Cas9 epitope tagging due to increased homologous recombination (26). To determine if this was also the case in N. caninum, we compared the efficiency of CRISPR/Cas9-mediated epitope tagging of endogenous genes in strains with and without the protein (Fig. 3). We selected three genes for epitope tagging; the Neospora orthologue of the T. gondii secreted protein GRA48 (NcGRA48, NCLIV_069360), a novel Neospora-specific protein that contains a signal peptide and a downstream series of tandem repeats (NCLIV_066730; Fig. S1 in the supplemental material), and a novel protein (NCLIV_049870) whose orthologue in T. gondii has a strongly negative fitness score in the genome-wide CRISPR/Cas9 screen (TGGT1_235420, -4.8), suggesting it is either important for fitness or essential (Fig. 3A). For epitope tagging, the NcLivΔhxgprt and NcLivΔhΔk strains were transfected with a gRNA construct for each gene as well as a homology repair template with 40 bp regions of homology to the GOI as is frequently used in T. gondii (29). As expected, NcGRA48 tagging showed that this protein localizes to the parasitophorous vacuole (Fig. 3B). Similarly, NCLIV_066730 also localized to the vacuole, indicating that this is indeed a novel Neospora-specific GRA protein (Fig. 3C). To our knowledge, NCLIV_066730 is the first Neospora-specific GRA protein that has been identified, and thus we named it Neospora-specific GRA protein 1 (NSG1) to differentiate it from the other GRA proteins with orthologues in the family. In contrast, NCLIV_049870 localized to the nucleus of the parasite (Fig. 3D). In each case, the efficiency of tagging was dramatically increased from around 20% with Ku80 present to as much as 95% tagging efficiency in Δku80 parasites (Fig. 3E). These data indicate that the NcLivΔhΔk strain has significantly improved recombination, verifies dense granule localization for NcGRA48, and identifies two completely new proteins in N. caninum.

FIG 3

Efficient CRISPR/Cas9 mediated gene tagging in NcLivΔhΔk parasites identifies new secreted and nuclear proteins. (A) Gene model diagram of endogenous tagging of NCLIV_069360/NcGRA48, NCLIV_066730/NSG1, and NCLIV_049870. GRA48 contains two internal predicted transmembrane domains (TM), and NCLIV_066730 contains a predicted signal peptide (SP). Models are not drawn to scale. (B) IFA showing that NcGRA48 localizes to the vacuolar space surrounding the parasites, as expected for a dense granule protein. IMC12 is used to label the periphery of the parasites, and Hoechst staining is used to detect the parasite’s nucleus. (C) IFA showing that 3xHA tagged NCLIV_066730 localizes to the vacuole. As NCLIV_066730 is Neospora-specific, it was named Neospora-specific GRA protein 1 (NSG1). (D) IFA showing that NCLIV_049870 localizes to the parasite’s nucleus as shown by colocalization with Hoechst staining. (E) Diagram showing the frequency of endogenous gene tagging with and without Ku80. The frequency of gene tagging is dramatically increased in parasites lacking Ku80.

Disruption of Ku80 results in efficient gene knockouts with minimal flanking regions.

In T. gondii, disruption of Ku80 enables highly efficient CRISPR/Cas9-mediated whole gene deletions using a CRISPR/Cas9 gRNA targeting the GOI plus a homology-directed repair template consisting of 40 bp flanking homology regions and a selectable marker (29, 30). In the absence of a Ku80 deficient line in Neospora, gene knockouts have been shown to require substantial flanking regions of homology (21–23). Our knockout attempts using 40 bp homology flanks in a Ku80 positive line appeared to generate disrupted genes by the insertion of the cassette containing the selectable marker inside the GOI’s coding region, rather than the desired full gene deletions (not shown). To determine if precise deletions of genes could be obtained in NcLivΔhΔk parasites, we attempted to disrupt NcGRA48 and NSG1 in the epitope tagged strains using the identical approach with a Toxoplasma HXGPRT cassette as the selectable marker (Fig. 4). As predicted, we were able to efficiently disrupt both genes, as shown by a lack of staining for the epitope tag (Fig. 4A and B). We then PCR verified that the entire gene was properly deleted and replaced by HXGPRT (Fig. 4C to E). These data demonstrate that whole gene deletions can easily be obtained in NcLivΔhΔk parasites using minimal flanking regions, similar to that seen in T. gondii (29).

FIG 4

Deletion of Ku80 enables efficient gene knockouts with minimal flanking regions for homology-directed repair. (A, B) IFA showing that deletion strategies eliminated the 3xHA-tag from NCLIV_069360/NcGRA48 and NCLIV_066730/NSG1 tagged strains. IMC12 is used to stain the periphery of the parasites. (C) Diagram showing strategy to PCR verify the gene of interest (GOI) deletions in NcLivΔhΔk parasites. (D, E) Agarose gel analyses of the gene knockouts showing integration of HXGPRT into the correct locus (NcGRA48: primers P53/P48; NSG1: primers P56/P48) as well as the absence of the GOI coding region in the knockout parasites (NcGRA48: primers P54/P55; NSG1: primers P57/P58).

Development of the auxin-inducible degron system for N. caninum.

We additionally wanted to utilize the NcLivΔhΔk parasites for the study of essential genes; thus, we modified the strain with an auxin-inducible degron (AID) system for conditional protein depletion (Fig. 5A) (31). To do this, we first attempted to express the Tir1 protein using a previously described T. gondii construct (31), but were unable to obtain stably expressing parasites (not shown). We thus replaced the promoter with the NcGRA7 promoter (Fig. 5B), which yielded stable expression of the protein localized to the cytoplasm of the parasite, as expected (Fig. 5C).

FIG 5

Stable expression of Tir1 in NcLivΔhΔk parasites. (A) Diagram of the AID system showing an auxin-inducible degron translational fusion is produced by endogenous gene tagging. Addition of indoleacetic acid (IAA) results in proteosomal degradation of the fusion protein. (B) Diagram of Tir1 driven from the Neospora GRA7 promoter, with Toxoplasma HXGPRT as a selectable marker. (C) IFA of NcLivΔhΔk parasites expressing Tir1FLAG. As expected, Tir1 localizes to the cytoplasm of the parasites. IMC12 is used to show the periphery of the parasites.

Knockdown of a novel nuclear protein identifies an essential nuclear factor in N. caninum and T. gondii.

To validate the AID system in N. caninum, we targeted the nuclear protein NCLIV_049870 as we predicted it would be either very important or essential due to the fitness score of its T. gondii orthologue (TGGT1_235420). As a control, we first AID tagged TGGT1_235420 in Tir1-expressing T. gondii and found that TGGT1_235420AID similarly localized to the nucleus (Fig. 6A and B). Knockdown of the protein with indoleacetic acid (IAA) resulted in a dramatic growth arrest of the parasites with the presence of mostly single parasites in a vacuole, some of which were in the process of endodyogeny (Fig. 6B). Many of the parasites also contained additional space within the vacuole. These parasites failed to grow further during elongated IAA treatment, as shown by plaque assays (Fig. 6C), demonstrating that TGGT1_235420 is essential in T. gondii. We then AID tagged NCLIV_049870 in Tir1-expressing N. caninum and found that the AID tagged protein also localized to the nucleus (Fig. 6D and E). IAA treatment of the NCLIV_049870AID parasites resulted in an identical lethal growth arrest with swollen parasites and increased vacuolar space (Fig. 6E and F). This demonstrates the AID system is functional in N. caninum and shows that this newly identified nuclear factor is essential in both parasites.

FIG 6

Conditional knockdown of a novel nuclear protein demonstrates essentiality in T. gondii and N. caninum. (A) Diagram of endogenous AID-tagged TGGT1_235420 in Tir1-expressing T. gondii. Tagging includes a 3xHA tag for detection and a DHFR cassette for selection. (B) IFA showing that 24-h addition of IAA results in loss of the nuclear HA signal and growth arrest of the parasites. Multiple vacuoles are shown +IAA to highlight that the parasites arrest at a single parasite per vacuole with some in the process of replication. Expanded vacuoles are also observed. (C) Plaque assays and quantification of plaque numbers showing that no plaques are formed upon TGGT1_235420 knockdown, confirming the protein is essential in T. gondii. (D) Diagram of endogenous AID-tagged NCLIV_049870 in Tir1-expressing N. caninum. (E) IFA showing that addition of IAA results in a similar growth arrest of the parasites with swollen vacuoles. (F) Plaque assays and quantification of plaque numbers showing that NCLIV_049870 is essential in N. caninum.

DISCUSSION

The close similarity between T. gondii and N. caninum provides an excellent opportunity to explore the factors that are unique to each parasite and are likely to control parasite virulence, their dramatically different host range, and the ability to perform the sexual cycle in distinct definitive hosts. To improve genetic manipulation in N. caninum, we disrupted the NHEJ enzyme Ku80 and found that we were able to obtain highly efficient CRISPR/Cas9-based gene tagging and knockouts. The frequency of tagging and ease of generating full gene knockouts (rather than just insertions) with a guide RNA plus a homology directed repair template are similar to that regularly observed in T. gondii. We additionally exploited the positive and negative selection of HXGPRT to generate parasites lacking this selectable marker (in addition to Ku80) for rapid tagging or knockout approaches. Together, these data suggest that recombination in NcLivΔhΔk parasites is similar to that seen in T. gondii and demonstrate the utility of this newly described strain for genetic manipulation in N. caninum.

We also demonstrated that loss of Ku80 results in the dramatic reduction of its dimer partner Ku70, presumably via destabilization and degradation of the protein. In many Δku80 parasites, Ku70 is undetectable, suggesting the Ku80 knockout is essentially a double knockout of the dimer (which requires both partners for function). This may be surprising as Ku70 was previously reported to be essential in T. gondii (27). However, these studies were done with early knockout tools that were just under development, and the Ku70 fitness score (–0.75) in the genome-wide CRISPR/Cas9 screen suggests that it is dispensable (32). A similar degradation of a partner has been seen in the AC9/AC10/Erk7 complex, where knockdown of AC10 results in the complete loss of AC9 from this complex, which is essential for maintenance of the conoid and release of apical secretory organelles (30). This may indicate that T. gondii has a robust system for elimination of misfolded proteins within the parasite, and that ability seems preserved in N. caninum.

Two of the genes that we tagged for localization purposes were secreted proteins. As expected, the Neospora GRA48 localized to the vacuole, indicating it is a dense granule protein as is seen in T. gondii, in which it was described to be phosphorylated in the parasitophorous vacuole by another secreted Toxoplasma protein, the PV-resident kinase WNG1 (33). Although TgGRA48 has been localized and shown to be activated by another GRA, its function in tachyzoite development and/or host interaction is still unknown. Complementary studies for NcGRA48 may further elucidate the role of GRA48 in parasite biology and enable a comparison between the orthologues of the protein.

The other predicted secreted protein was also demonstrated to be a dense granule protein. We initially selected NSG1 as a target because it was uncharacterized and is Neospora-specific, without any known orthologues within the Sarcocystidae family of Apicomplexa. Neospora appears to lack a number of the GRAs present in T. gondii that are known to modulate host functions, including GRA15, GRA24, TEEGR, TgIST, and TgNSM (34–39). Those Toxoplasma GRAs that are missing in Neospora may be important molecules controlling parasite virulence or host range. However, NSG1 is the first Neospora GRA that is absent in T. gondii, suggesting that this effector protein may modulate Neospora-specific activities in the host or carry out some other function in the Neospora parasite vacuole.

The mature NSG1 protein is largely composed of tandem repeats, which are also present in several secreted Toxoplasma proteins (e.g., GRA16, MAG2, ROP1, TLN1). While these repeats differ in their sequences and are typically of unknown function, they have been hypothesized to have a function in host cell invasion, immune evasion, and ultimately virulence (40–42). It will be particularly interesting to assess whether there is a strong immune response to NSG1, as this may indicate a role on immunity. In addition, NSG1 may serve as a unique antigen for diagnostic purposes, which have largely centered on the detection of surface antigens or secreted proteins present in both species that have diverged substantially to generate parasite-specific responses (43–45). In that sense, it is remarkable that NSG1 is predicted to have linear B cell epitopes throughout its whole sequence, with high scores for hydrophilicity, accessibility, and flexibility exactly in the region where one of the sets of the tandem repeats are concentrated (Fig. S1B). These features have been well characterized for that purpose in other protozoan parasites such as Leishmania and T. cruzi (46, 47).

We additionally modified the NcLivΔhΔk strain with the AID system for the rapid analysis of essential genes. We initially expressed Tir1 using the T. gondii construct that is driven from the tubulin promoter, but the parasites had variable levels of expression, even within parasites in a single vacuole. This variable expression from Toxoplasma promoters in Neospora has been observed previously in the study of rhoptry proteins (9), and was resolved by using the Neospora GRA7 promoter. The variable expression suggests that it may be best practice to use Neospora promoters whenever possible, although Toxoplasma promoters have been shown to be successful for many uses (9, 10, 20) The AID system in the Δku80 background thus provides a new instrument for rapid conditional knockdown at the protein level in N. caninum.

This usefulness was proven during the characterization of NCLIV_049870 and its orthologue TGGT1_235420. Although yet to be thoroughly characterized, the nuclear proteins in both parasites are essential for parasite growth inside the vacuole. An initial analysis of these genes shows orthologues in other species within the family, and the predicted amino acid sequence shares 60% of identity between T. gondii and N. caninum. While there are no definitive conserved motifs in the proteins that would denote a specific function, analysis of the sequence hints to its nuclear role, probably chromatin-remodeling related. Evolutionary analysis by OrthoMCL indicates that both proteins originated from the ancestor Sphaeroforma arctica, which contains a helicase-SANT-associated (HSA) domain that acts as chromatin remodelers by binding nuclear actin-related proteins (48). Although this domain is not preserved in T. gondii, N. caninum, and the closest phylogenetic species, orthologues in Cyclospora cayetanensis, Plasmodium spp and Cryptosporidium spp still contain the HSA domain. In addition, NCLIV_049870 is annotated as “GH22120, related” in ToxoDB, referring to a Drosophila grimshawi protein with repeated mucin-like domains. Human proteins containing such domains as MUC-1C are also involved in chromatin remodeling and are indicators of cancer progression and poor clinical outcome (49, 50). Finally, String database prediction for functional partners of these proteins are all chromatin-related: TGME49_267800 (DRPA), TGME49_203950 (BDP), TGME49_289730 (Pep3/Vps18/deep orange family).

In conclusion, it is well known that molecular genetics provides powerful investigative approaches that can greatly improve our understanding of intracellular pathogens and their interactions with host cells. This study dramatically improves the ability for genetic manipulation of N. caninum, similar to the tools available in T. gondii. The development of molecular genetics for N. caninum will enhance knowledge of neosporosis, enable a better comparison of N. caninum and T. gondii, and provide an efficient heterologous expression system to unravel the processing, targeting, and function of T. gondii proteins.

MATERIALS AND METHODS

Host cells and parasite culture.

Parental NcLiv and modified strains of N. caninum, as well as RHΔhxgprtΔku80 and modified strains of T. gondii, were maintained on confluent monolayers of human foreskin fibroblast (HFF) cells, cultured in complete Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 5% Fetal Bovine Serum (Gibco), 5% Cosmic Calf Serum (HyClone), 2 mM glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin, at 37°C in 5% CO2 atmosphere, as previously described (51, 52).

In silico analysis of POI.

All general information on proteins of interest (POIs) such as sequences, essentiality, genomic location, orthology profile, among others, were obtained at ToxoDB (www.toxodb.org). Orthologous grouping of evolutive and expanded families of POIs were predicted by OrthoMCL (orthomcl.org) and Eukaryotic Pathogen, Vector and Host Informatics Resource (VEuPathDB, veupathdb.org), based on amino acid sequences. We used SignalP 6.0 (services.healthtech.dtu.dk/service.php?SignalP-6.0) for the prediction of signal peptides and cleavage sites of the POI sequences and TMHMM Server v. 2.0 (http://www.cbs.dtu.dk/services/TMHMM/) for the prediction of transmembrane helices. Prediction of structure and/or function of the POIs was performed on the Protein Homology/analogy Recognition Engine V 2.0 (PHYRE2, www.sbg.bio.ic.ac.uk/phyre2). For detection of tandem repeats in the studied POI, we used Rapid Automatic Detection and Alignment of Repeats in Protein Sequences (RADAR, www.ebi.ac.uk/Tools/pfa/radar), along with manual curation of the data. B cell immunogenicity and linear epitope predictions were obtained at Immune Epitope Database and Analysis Resource (IEDB, www.iedb.org), through the following tools: Bepipred Linear Epitope Prediction 2.0 (53); Parker Hydrophilicity Prediction (54); Emini Surface Accessibility Prediction (55); and Karplus & Schulz Flexibility Prediction (56). String (string-db.org) was used for the prediction of potential protein interactions and functional enrichment analysis.

Epitope tagging.

We used CRISPR/Cas9 for endogenous tagging of genes of interest—NCLIV_056045/NcKu80 (primers P1-P4), NCLIV_064150/NcKu70 (primers P5-P8), NCLIV_069360/NcGRA48 (primers P9-P12), NCLIV_066730/NSG1 (primers P13-P16), NCLIV_049870 (primers P17-P20)—with 3xHA, as previously described (29). Oligonucleotides used for this and other purposes are listed in Table S1. Briefly, sequences encoding gRNAs were ligated into the pU6 Universal plasmid and prepared along with a PCR product from the pLIC:3xHA:DHFR vector with 40 bp flanking regions for recombination at the 3′ end of each gene. Transfection was performed in 2 mm gap cuvettes with 400 μL of cytomix containing 107 freshly lysed tachyzoites, 100 μg of each construct, and CaCl2 (150 μM), using single pulses of 1.5–2.0 ms at 1.5 kV, 25 Ω and 25 μF in an electroporator (Bio-Rad), as previously described (19). Parasites were then selected in medium containing 1 μM pyrimethamine and cloned by limiting dilution. Tagged clones were screened by IFA using the HA.11 monoclonal antibody (BioLegend). Cross-reactive rabbit antibodies against Toxoplasma IMC12 (Back et al., unpublished data) were used to stain the periphery of the parasites. Epitope tagging efficiency was calculated by counting the HA positive vacuoles versus IMC12 positive vacuoles at 14 days posttransfection. At least 100 vacuoles per strain were counted to determine the percentage of parasites successfully tagged.

Gene disruption.

CRISPR/Cas9 and homologous recombination were used to disrupt selected targets, as previously described (29). For NcHXGPRT (NcLivΔhxgprt), a single sgRNA (primers P21/P22) was used to disrupt the GOI’s coding region. For NcKu80 (NcLivΔhΔk), flanking sequences from the 5’UTR (forward, 559 bp) and 3’UTR (reverse, 528 bp) were amplified from N. caninum genomic DNA (primers P23-P26), inserted into a cassette containing T. gondii HXGPRT driven by the N. caninum GRA7 promoter (pJET1.2:pNcGRA7:TgHXGPRT), and transfected into NcLivΔhxgprt parasites with a gRNA (primers P27/P28) directed to the beta-barrel domain of the gene. To remove the selectable marker, NcLivΔhΔk parasites were transfected with a gRNA (primers P29/P30) directed to the TgHXGPRT coding region and an AccI truncated version of the original cassette as a repair template (–1633 bp). This eliminated most of the NcGRA7 promoter and TgHXGPRT coding region, although retaining approximately 2000 bp for efficient homologous recombination. To generate NcLivΔ069360/NcGRA48 and NcLivΔ066730/NSG1 strains, epitope tagged NcLivΔhΔk parasites were transfected with sgRNA directed to the GOI’s coding regions, along with a PCR product from the pJET1.2:pNcGRA7:TgHXGPRT with 39 bp homology sequences of its UTR regions (NcGRA48: primers P31-P34; NSG1: primers P35-P38).

Auxin-inducible degron modifications to N. caninum.

To produce N. caninum stably expressing the transport inhibitor response 1 (Tir1) auxin receptor from Oryza sativa, the original plasmids generated and codon optimized for T. gondii (pTgTUB1:OsTIR1-3xFLAG, TgSAG1:CAT) (31) were adapted to the N. caninum system by Gibson assembly. Briefly, the Toxoplasma TUB1 promoter was replaced by the Neospora GRA7 promoter (pNcGRA7, primers P39-P42). In addition, the selectable marker used was also changed to TgHXGPRT instead of chloramphenicol acetyltransferase (CAT) used in the original description (primers P43-P46). These changes generated the plasmid pNcGRA7:OsTIR1-3xFLAG, TgSAG1:TgHXGPRT, used in this work. NcLivΔhΔk parasites were transfected with this linearized construct, selected with Mpa/X, cloned by limiting dilution, and a clone was chosen with strong cytoplasmic staining with the M2 anti-Flag Mab (Sigma). The original pTgTUB1:YFP-mAID-3xHA, HXGPRT plasmid, containing the mAID sequence for proteasomal degradation, was altered by changing its selectable marker to DHFR, generating the pTgTUB1:YFP-mAID-3xHA, DHFR plasmid (31). For depletion of AID fusion proteins, the parasites were treated with 500 μM indoleacetic acid (IAA) added at the time of infection for the specified time periods.

Plaque assays.

Six-well plates were seeded with confluent monolayers of HFFs and infected with approximately 250 parasites (±500 μM IAA) and allowed to form plaques for 6 days for T. gondii and 8 days for N. caninum. The monolayers were then fixed in 100% methanol for 3 min, washed with phosphate-buffered saline (PBS), and stained with crystal violet for visualization. The number of plaques were quantified from the samples and the plaque assays were performed in triplicate.

Immunofluorescence assay and fluorescence microscopy.

For IFAs, tachyzoites were used to infect coverslips of HFFs for 24 h. The coverslips were then fixed in 3.7% formaldehyde/PBS for 15 min. The coverslips were washed in phosphate-buffered saline (PBS) and blocked and permeabilized in a solution containing PBS/3% bovine serum albumin (BSA)/0.2% Triton X-100 for 30 min. Samples were then incubated with mouse monoclonal anti-HA antibodies (HA.11, BioLegend) and rabbit anti-IMC12 (Back et al., unpublished data) diluted in PBS, along with BSA (3%) and Triton X-100 (0.2%) for 1 h. The samples were then washed in PBS and treated with species-specific secondary antibodies conjugated to Alexa 594/488, and diluted 1:2000 in PBS/3%BSA/0.2%TX-100 Hoechst staining was performed using Hoechst 33342 (5 μg/mL) in PBS for 5 min. Following a new washing cycle, coverslips were mounted onto microscope slides with Vectashield (Vector Labs) mounting media, and the fluorescence was observed using a Zeiss Axio Imager Z1 microscope. Images were processed via deconvolution using the Zeiss Zen software.

Western blot analysis.

Whole-parasite lysates were separated by 10% SDS-PAGE. Samples were transferred to nitrocellulose overnight and probed with primary antibodies. For all secondary antibody incubations, horseradish peroxidase (HRP)-conjugated goat anti-mouse or goat anti-rabbit antibodies were used at a 1:2000 dilution. Following secondary incubation, a chemiluminescent substrate was used for the detection of HRP activity.

Features of NCLIV_066730/NSG-1. (A) Amino acid sequence of NCLIV_066730/NSG1 highlighting with the predicted signal peptide (green) as well as a series of repeats (pink and red) in the mature protein. Repeats were identified using the RADAR protein repeat finder and curated manually. (B) The presence of antibody binding sites in the amino acid sequence of NCLIV_066730/NSG1 was predicted using four different algorithms at the Immune Epitope Database and Analysis Resource (IEDB), as follows: Bepipred Linear Epitope Prediction 2.0; Parker Hydrophilicity Prediction; Emini Surface Accessibility Prediction; and Karplus & Schulz Flexibility Prediction. Download FIG S1, TIF file, 1.0 MB.

Oligonucleotide primers used in this study. All primer sequences are shown in the 5′ to 3′ orientation. Download Table S1, PDF file, 0.02 MB.