Neospora caninum Liverpool was acquired from ATCC (Cat.#: 50845). NcUru1-4 were isolated from local congenitally infected calves and functionally and genetically characterized (7). All strains were cultured and maintained by serial passages in human foreskin fibroblasts (HFF, ATCC #SCRC 1041) in DMEM (Gibco™, Thermo Fisher Scientific, Waltham, MA, United States), supplemented with 10% Fetal Bovine Serum (Invitrogen, Carlsbad, CA, United States), 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA, United States). Cells were incubated at 37°C and 5% CO₂. The infected HFF monolayers were removed with a cell scraper, centrifuged and washed three times with PBS at 500x g for 10 min. Subsequently, the parasites were resuspended in PBS, passed three times through an 18G needle, and filtered through a 3-uM pore-size filter. The number of tachyzoites was determined by using a Neubauer chamber.

For immunoblots protein extracts of N. caninum tachyzoites were analyzed as previously described (17). Briefly, total parasite proteins (30 ug per well) were separated on 12% SDS-polyacrylamide gels, and bands were visualized by Coomassie Blue staining or transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, MA, United States) together with a visible prestained protein marker (PageRuler™ Prestained Protein Ladder, Thermo Scientific™). The membranes were blocked with 5% (w/v) skim milk in PBS for 1 h at 37°C, rinsed with a washing buffer, and incubated with mouse serum (1:100) overnight at 4°C. The blots were washed five times with PBS-T (1% Tween-20), followed by incubation with horseradish peroxidase (HRP)-labeled goat anti-mouse IgG (H + L) (1:5,000, Sigma, United States). Finally, enhanced chemiluminescence reagents (SuperSignal™ West Pico PLUS Chemiluminescent Substrate, Thermo Scientific™) were used to observe the reaction bands after 30 s exposure time.

Immunofluorescence assay of intracellular tachyzoites grown on HFFs was performed using a commercial N. caninum antiserum raised in goats (1:100; VMRD, Pullman, WA, United States). Donkey anti-goat IgG H&L-FITC conjugated (Abcam, Cambridge, United Kingdom) was used at a 1:1000 dilution as the secondary antibody. The acquisition was carried out in an epifluorescence microscope (Olympus Life Science, Tokyo, Japan) using a × 100 oil objective (7).

For in vivo studies, 6–8 week-old BALB/cJ mice (Jax# 000651) were used following approved institutional animal care and use protocols. Mice were bred at the animal facility of the Laboratory Animals Biotechnology Unit of Institut Pasteur de Montevideo under specific pathogen-free conditions in individually ventilated racks (IVC, 1285 L, Tecniplast, Milan, Italy). The housing environmental conditions during the experiment were as follows: 20 ± 1°C temperature, 30–70% relative humidity, negative pressure (biocontainment), and a light/dark cycle of 14/10 h. All procedures were performed under Biosafety level II conditions. Mice were housed in groups of five for each of the six conditions (PBS, NcLiv and NcURU 1–4), and the whole experiment was performed by duplicate, meaning that 10 mice were used per each condition.

Parasites in tissue culture were mechanically extracted from the host cells by repeatedly passing them through a 26G syringe, filtered through a 3 μm pore size polycarbonate filter, and counted in a Neubauer chamber (7). To prevent the selection of in vitro culture phenotypes, all strains used in phenotypic and in vitro assays were passed less than 12 times from the time of isolation to when the assays were performed.

Beforehand, the LD₅₀ was determined for our commercial NCLiv strain in Balb/cJ female mice. The estimated LD50 was 1×10⁷ Liverpool tachyzoites inoculated intraperitoneally.

This dose (10⁷) was used in all assays to determine strain-specific virulence. Chronic disease was induced in 6–8 week old female BALB/cJ mice by intraperitoneal inoculation of a sublethal dose of 10⁶ tachyzoites, as determined previously based on the literature (13). Mice were visually inspected daily for 30 days post-inoculation to detect clinical signs compatible with neosporosis (rough fur, inactivity, anorexia and neurological signs). Seropositivity against N. caninum infection was systematically tested two weeks post-infection via submandibular bleeding. Sera were tested by immunoblotting against N. caninum tachyzoites lysates. Mice that tested positive for ELISA were euthanised using humane methods, and aseptic brain samples were collected for subsequent analysis. PCR was performed using the previously described primers NP6 (5′-CAGTCAACCTACGTCTTC-3′) and NP21 (5’-GTGCGTCCAATCCTGTAA-3′) (18) to detect parasite DNA in the brain samples. Concurrent positivity in both serology and PCR was considered indicative of a chronic infection.

For vertical transmission studies, chronically infected or a control group of non-infected female mice were housed in groups of six for each of the six conditions (PBS, NcLiv and NcURU 1–4) and the experiment was performed by triplicate, meaning that 18 mice were used per each condition.

Ten-week-old mice were synchronized in their estrous by exposing them to male dirty bedding, as previously described (19). After estrous synchronization, three females were placed with a single male for seven nights. The presence of a vaginal plug was corroborated daily in all cases as a positive sign of mating and considered day 0 of pregnancy. Pregnant mice and fetuses were humanely euthanized at day 20 post-coitum, and brain DNA from mothers and viable fetuses was extracted for PCR assays. Sera were collected from the submandibular vein of BALB/cJ mothers for serological assays.

For the intracellular growth assay, confluent monolayers of human foreskin fibroblasts (HFFs) were grown on coverslips and subsequently inoculated with 10³ freshly released tachyzoites. Two hours post-inoculation, cells were washed five times with PBS to remove unattached parasites and fixed with 4% paraformaldehyde 24 h later. Samples were analyzed by immunofluorescence assays, as described by Cabrera et al. (7). The number of parasites per vacuole in 200 vacuoles per coverslip was determined in triplicate. Confluent HFFs grown on coverslips were kept on ice before the invasion assays. Upon inoculation with 1,000 parasites, plates were centrifuged at 300 × g for 10 min at 4°C to ensure all strains contacted the cells simultaneously. Plates were then switched to 37°C for 30 min, washed with PBS three times to remove unattached parasites, and fixed with 4% paraformaldehyde. The number of parasites per field in 50 randomly selected fields was determined for all strains, in triplicate, by immunofluorescence assay.

DNA was extracted using a DNA purification kit from Zymo Research (#D4074). Illumina libraries for N. caninum NcUru1, NcUru2, NcUru3 and NcUru 4 were performed as described (20), and whole genome sequencing of the strains was performed with 2×100 paired-end cartridges. Reads were mapped to the new reference N. caninum Liverpool genome (21) using Burrows-Wheeler Aligner (BWA 0.7.17) (22) mem using default parameters. Samtools v1.9 was used to post-process alignment files (22). SNP and Indel discovery was performed with Samtools (23) and VarScan. v2.3.9 (24) with parameters –min-reads2 3 –min-coverage 3 –min-var-freq 0.8. SNPeff v.5.0 was used to annotate genomic variants and predict their functional effects. IGV was used for alignment and variant visualization (25). Venny 2.1 was used for Venn diagram plots (26). For the NcUru4 isolate the sequencing coverage was not enough for this analysis.

All animal experiments were done following international animal care and use guidelines, as well as respecting National Law 18.611 with pre-approved protocols by Institutional Ethics Committee (CEUA # 010–17) at the Institut Pasteur de Montevideo.

To comparatively assess the phenotypes of NcUru1-4, we first evaluated there in vivo virulence. To do this, we infected immunocompetent Balb/cj mice with 10⁷ freshly purified parasites of each of the four isolates. NcLiv was used as positive and PBS as negative controls. Mice seroconversion was determined by Western blot two weeks after inoculation, considering as positive the presence of the expected discrete bands of 17 kDa, 29–37 kDa and higher than 70 kDa (27, 28) (Figure 1A). Clinical signs of the disease (anorexia, rough coat, and inactivity followed by nervous signs like rounded back, pelvic limb weakness and walking in circles) were monitored for a month post-inoculation.

As expected, 60% of the mice infected with the reference strain NcLiv had to be euthanised due to the infection within a week. In contrast, mice infected with NcUru1-4 showed no obvious clinical signs of the disease and exhibited equal survival as that of the PBS control (Figure 1B). In addition, all strains were detectable by PCR in the brains of the surviving mice, which suggests that they were all able to transition from the acute to the chronic form. However, while a positive PCR was detectable in all the brains of mice infected with NcLiv (both surviving and euthanized due to the infection), NcUru1-4 were consistently detectable in 30, 50, 60 and 40% of mice, respectively. This suggests that the ability to persist in the host chronically varies significantly with respect to NcLiv, but also among NcUru strains (Figure 1C). However, all mice, both surviving and euthanized, seroconverted two weeks post-inoculation (shown only for five mice infected with NcUru1 in Figure 1A).

We next evaluated the strain-specific impact on pregnancy outcomes. To do this, we optimized a murine model of vertical transmission. In short, we infected 6-8-week-old BALB/cJ female mice with 10⁶ parasites in order to generate subclinical chronic infection. We next crossed females and males, whereby subgroups of females infected with the same parasite strain were crossed with the same male. Every experiment was done in triplicate using six female mice per parasite strain per experiment. We note again that all mice effectively seroconverted two weeks after inoculation (Figure 2A). Pregnancy rates (pregnant/mated females) did not show significant differences. However, they varied from 28% for females inoculated with NcLiv (positive control), to 48% for the control group (PBS) (Figure 2B). In the case of NcLiv, chronic infection of the mother was corroborated postmortem in all cases. The NcUru isolates impacted pregnancy rates differently but within the ranges of the positive and negative controls. Pregnancy rates in females infected with NcUru1 were comparable to that of NcLiv (28%). NcUru4 behaved similarly to the negative control (PBS- 48%) (Figure 2B).

Chronic infection of the females was evaluated postmortem by PCR. The results were comparable to those of the higher inoculum infection described above. These results reinforce that the observable differences in the establishment of chronic infection are intrinsic strain-specific characteristics (Figure 2C). Pregnancy, abortion and the vertical transmission rate for each strain were recorded.

Mice were euthanized a day before the expected due date. This allowed us to correlate the impact of given mouse chronicity on the offspring. We observed that virtually all offspring were viable in both negative (PBS) and positive control (NcLiv) groups; 98 and 100%, respectively (Figure 3A). When the progeny of mice infected with NcUru isolates were evaluated, we observed that the output of viable offspring significantly decreased only for mothers displaying a chronic infection with NcUru2. Chronically infected mice with NcUru2 bear only 63% of viable offspring. In contrast, mice inoculated with NcUru2 who did not develop a chronic infection had 100% viable offspring, suggesting that the impact on offspring viability is directly related to the chronic presence of NcUru2. NcUru3-chronically infected mothers yielded 77% viable pups, while those chronically infected with NcUru4 exhibited 75% viable offspring. Interestingly, progeny viability for the NcUru3 was still halved in non-chronically infected mothers, suggesting that both previously acquired acute infections and chronic persistence of this strain have a sizable impact on pregnancy rates (Figure 3A). NcUru1-infected mothers yielded an average of 81% viable offspring (Figure 3A).

To compare strain-specific vertical transmission rates, we determined the percentage of N. caninum positive fetuses by PCR detection of the parasite DNA in the brains of viable offspring (Figure 3B). Consistent with previous reports, we detected a vertical transmission rate of 93% in three independent experiments for NcLiv. Vertical transmission rates for NcUru1 thru 4 were comparatively lower than that of NcLiv. NcUru2 was detectable only in 20% of the progeny, while the vertical transmission rate was 29 and 17% for mice infected with NcUru3 and NcUru4, respectively. Remarkably, NcUru1 was only detectable in 16% of the progeny (Figure 3B). The latter is consistent with NcUru1-infected mothers of all strains, yielding the highest percentages of viable offspring. We note that NcUru1-infected mothers did not get pregnant for the first replicate experiment (see Table 1).

It is well established for N. caninum’s close relative, Toxoplasma gondii, that the capacity to invade cells and velocity of intracellular growth correlates with virulence (29). To approach the molecular basis of the phenotypic differences observed in vivo, we in vitro characterized each strain’s ability to invade a host cell and their intracellular growth rate using NcLiv as a reference. These assays revealed that NcUru1 is half as invasive as NcLiv, while NcUru2 invades roughly 25% fewer cells than NcLiv in the same timeframe. NcUru3 invades host cells with an efficiency similar to the reference (Figure 4A). Remarkably, NcUru4 exhibits a significantly reduced invasion rate (18%) (Figure 4A). Conversely, when the relative intracellular growth of each strain was evaluated, we found that NcUru4 exhibits a similar growth rate to that of NcLiv, suggesting that those parasites which manage to invade are equally able to grow intracellularly. NcUru1 (72%), NcUru2 (53%) and NcUru3 (38%) were comparatively slower growing than NcLiv (Figure 4B). The reduced efficiency of NcUru1 in invasion and intracellular growth are in consonance with its reduced efficiency in generating chronic infection and lower vertical transmission rates.

To determine the mutations responsible for observed phenotypical differences, we sequenced the complete genomes of NcUru isolates. A large number of genetic variations (2591) were identified among the different strains, including heterozygous SNPs/Indels. Of these, 1819 were located in coding regions (Supplementary Table 1). We identified the genes affected by strain-specific SNPs in our newly annotated genome (21). Notably, most of the mutations were strain-specific, with 189 found in NcUru1, 721 in NcUru2 and 455 in NcUru3. Only 143 mutations were common to all three NcUru isolates (Supplementary Table 2), and 35% corresponded to hypothetical proteins (Supplementary Table 3). Of the annotated genes, 9% contain functional domains related to surface proteins (SRS domains), while the rest mainly correspond to translation-related genes (ribosome biogenesis, ribosomal proteins and elongation factors), protein degradation (ubiquitin proteases and proteasome subunits), AP2 transcription factors and regulatory proteins, including mainly kinases and phosphatases (Supplementary Table 3). The 143 shared genes are enriched in GO terms related to pentose-phosphate pathways and proteasome/ubiquitination (Figure 5), whereas GO terms of strain-specific genes show the predominance of pentose phosphate pathway and cation binding related genes in NcUru1, lipid metabolism and membrane traffic in NcUru3. No significant GO terms were found in NcUru2 (Supplementary Table 1).