Hammondia hammondi has a developmental program in vitro that mirrors its stringent two host life cycle

Hammondia hammondi is the nearest relative of Toxoplasma gondii, but unlike T. gondii is obligately heteroxenous. We have compared H. hammondi and T. gondii development in vitro and identified multiple H. hammondi-specific growth states. Despite replicating slower than T. gondii, H. hammondi was resistant to pH-induced tissue cyst formation early after excystation. However, in the absence of stress H. hammondi spontaneously converted to a terminally differentiated tissue cyst stage while T. gondii did not. Cultured H. hammondi could infect new host cells for up to 8 days following excystation, and this period was exploited to generate stably transgenic H. hammondi. Coupled with RNAseq analyses, our data clearly show that H. hammondi zoites grow as stringently regulated life stages that are fundamentally distinct from T. gondii tachyzoites and bradyzoites.

155 is yet another trait that distinguishes this parasite from T. gondii, and provides further evidence 156 that the H. hammondi developmental program is hard-wired and very difficult to disrupt. phenotype, we subcultured H. hammondi sporozoites (by needle passage) at multiple times 165 post-excystation (Fig 5A-B). We found that when H. hammondi zoites were mechanically 166 released from their host cells, they were capable of infecting and replicating within new host 167 cells for up to 8 days post-excystation (Fig 5C-F), after which we were unable to observe any 168 evidence of parasite replication. In agreement with previous studies (2, 14, 15) we also 169 observed that the number of visible H. hammondi vacuoles seems to disappear the longer the 170 parasites were grown in culture. Additionally, we observed that vacuoles began to disappear

184
hammondi cultivated for a short time in vitro could also be used to infect mice. HhCatAmer or 185 HhCatEth1 sporulated oocysts were excysted and grown in HFFs for 4 days, at which time 186 monolayers were scraped, syringe lysed, and filtered. We infected mice intraperitoneally with 187 50,000 zoites of either HhCatAmer (4 mice) or HhCatEth1 (1 mouse) and harvested spleens, 188 peritoneal lavage fluid, and cells (PECs) on 4 (3 mice) or 9 (2 mice) DPI. All 5 spleen samples and 1 PEC sample had detectable H. hammondi DNA based on PCR using H. hammondi specific primers (Fig 6A)

247
To determine if our transgenic approach to disrupt the UPRT gene and insert a dsRED            Table 2 for Gene Sets).   304 gene set that were detectable, 18 of them were ranked in the top 308/4146 (7.4%) and 305 174/4146 (4.2%) of H. hammondi-high genes at D4 and D15, respectively ( Fig 8D). These data 306 suggest that H. hammondi has aspects to its transcriptional profile that are independent of its 307 spontaneous conversion to bradyzoites and that a subset of genes originally thought to be 308 restricted to merozoites may only be merozoite-specific in some species and not others.

319
One caveat with these data is that the number of reads mapping to the H. hammondi 320 transcriptome was, on average, an order of magnitude lower than those that mapped to T.

321
gondii. While this is most certainly due to the dramatic differences in replication rate between 322 these species (Fig 2), we validated a subset of transcripts that were of greater abundance in H.

342
The ability to infect multiple hosts is a remarkable feature of many parasite species.  Table   380 S1 for normalized RNAseq data).

381
One surprising finding was that we repeatedly harvested significantly higher sporozoite 382 yields from T. gondii compared to H. hammondi (~2-fold overall; Fig 1A), and this was independent of oocyst age and the preparation of H. hammondi (Fig 1B). One caveat is that the 384 same prep of TgVEG oocysts was used for all 10 extractions, so this observation could be due

Quantification of sporozoite viability and replication rate 509
After 10 minutes incubation on HFFs at 37°C in 5% CO 2 , monolayers containing freshly

Characterizing limits of in vivo infectivity of H. hammondi grown in vitro 578
Sporozoites were obtained using the excystation protocol described above. After 24 579 hours, HFF monolayers infected with excysted sporozoites were scraped, syringe lysed 5X with

Transfection of Hammondia parasites and selection of recombinant parasites 615
Excysted sporozoites were prepared as described above and incubated overnight in a T25 flask with confluent monolayer of HFFs. After 24 hours, the monolayer was scraped, x g for 10 min, and the pellet was resuspended in 450 µl of cytomix with 2 mM ATP and 5 mM 620 glutathione. Resuspended parasites were transferred to a cuvette, electroporated at 1.6 KV and 621 a capacitance of 25 µF, and used to infect confluent HFF monolayers on coverslips. The 622 coverslips were fixed 5 DPI as described above and mounted using ProLong® Diamond

634
Transfected parasites were then transferred to HFFs and grown for 2 days in cDMEM, then 635 selected for 3 days by incubation in cDMEM containing 10 μM FUDR (Fig 5B). Parasites were 636 again scraped, syringe lysed and filtered, and dsRED-expressing parasites were collected in PBS using flow cytometry. Sorted parasites were injected intraperitoneally into 2 BALB/c mice.

638
After 3 weeks of infection, mice were euthanized, skinned, and the intestines were removed 639 before feeding to specific pathogen-free cats. The oocysts were collected and purified as described above, and oocysts, sporozoites and replicating parasites were evaluated for dsRED

FUDR resistance of transgenic parasites 644
To test for UPRT resistance in transgenic H. hammondi, sporozoites of wild type and infected coverslips containing confluent HFFs with 50,000 parasites, and parasites were  Raw count data per transcript (generated by featureCounts above) were loaded into R 687 statistical software and analyzed using the DESeq2 package (44). Comparisons of D4 and D15 688 read count data were used to identify transcripts of different abundance at each time point, and 689 differences were deemed significant at P adj <0.05. Data were log 2 transformed and normalized 690 using the rlog function in DESeq2 for use in downstream analyses. Log 2 transformed, normalized data from DESeq2 (hereafter referred to as Log 2 (FPM)) were analyzed for enrichment using Gene Set Enrichment Analysis (GSEA; (24)). Since read count overall from 7372 genes matched based on the previously published gene-by-gene annotation (12) and 695 selected only those that had at least 1 day 4 sample and 1 day 15 sample with >5 reads. In 696 total, 4146 genes passed these benchmarks and were used in subsequent analyses. We used 697 this approach to identify gene sets that were significantly different between days 4 and day 15 in 698 culture in both species and those that were different between species at both day 4 and day 15.

699
We compared these data using previously curated gene sets (434 total; as published in (25)) as 700 well as 7 additional gene sets that we curated ourselves. These are listed in Table 2 Table S2 for primer sequences