Ethics statement

All experiments that involved live animals were conducted with the approval of local animal welfare authorities. The experiments were reviewed and approved by the Regierungspräsidium Giessen (approval numbers: GI 18/5 A7/2018 and GI 18/5 G32/2018). All animals were euthanized by exsanguination under general anesthesia with isoflurane.

Parasites

Sarcocystis calchasi sporocysts were obtained from an experimentally infected northern goshawk. The goshawk originated from the wild and was presented to the Clinic for Birds, Reptiles, Amphibians and Fish at the Justus Liebig University Giessen with clinical conditions impeding the release back to the wild. The goshawk’s feces from three consecutive days were tested for the absence of Sarcocystis spp. sporocysts by flotation in saturated saline three times over a period of 40 days. The goshawk was infected by feeding the carcass of a pigeon experimentally infected with S. calchasi isolate Giessen16. This was done as part of the regular host passage for maintenance of the isolate. Fourteen days after infection (dpi) the goshawk was euthanized, its intestines scraped and sporocysts isolated from the gut mucosa and intestinal content via trypsin digestion and filtering. The sporocyst suspension was finally treated with sodium hypochlorite, counted in a hemocytometer and stored in Hank’s Balanced Salt Solution (HBSS) supplemented with penicillin/streptomycin and amphotericin B at 4°C.

To obtain sporozoites, sporocysts were excysted via bile incubation. Briefly, sporocysts were washed once with PBS, treated shortly for 10 min with sodium hypochlorite, washed five times and incubated for 4.5 h in Dulbecco’s minimal essential medium high glucose with 3% fetal bovine serum, glutamine and penicillin/streptomycin (DMEM culture medium) containing 15% chicken bile. Released sporozoites were washed three times with DMEM culture medium and counted in a hemocytometer.

Sarcocystis calchasi bradyzoites were isolated from the pectoral muscle of a pigeon experimentally infected with the same sporocysts as used for sporozoite extraction as described by Verma 2017 [40]. Pectoral muscle was homogenized with a commercial blender, digested with HCl-pepsin solution and filtered through sterile gauze. After removal of host cell debris, bradyzoites were washed with PBS and resuspended in HBSS supplemented with penicillin/streptomycin and amphotericin B. Bradyzoites were counted in a hemocytometer and stored in 4°C or liquid nitrogen.

Cell culture

To produce merozoites, cell cultures were infected with sporozoites. After repeated testing of permanent cell lines (Vero cells, obtained from the German Collection of Microorganisms and Cell Cultures GmbH, ACC 33, and CEC-32 [41]) with very inconsistent success rates, primary pigeon liver cells were used for culturing S. calchasi merozoites. Primary cells were produced by incubation of fertilized pigeon eggs in an incubator at 37°C, approx. 60% humidity and regular turning. On day 12, eggs were removed from the incubator. The eggshell was opened aseptically and the embryo removed. After decapitation, the liver was separated from the embryo, washed twice in PBS and digested in trypsin solution for 30 min or until the tissue had dissolved. The suspension was filtered through sterile gauze, centrifuged for 2 min at 2,000 x g and resuspended in 6 ml DMEM culture medium. The cells were incubated at 37°C with 5% CO₂ for 1–3 days until a monolayer had developed.

Sporocysts were excysted as described above with additional removal of remaining sporocysts in a PD desalting column. The final pellet of sporozoites was taken up in 200 µL DMEM culture medium and dispensed in the supernatant of the primary pigeon liver cells. Appearance of merozoites in the supernatant was monitored every 2–3 days. If a considerable number of merozoites was observed, the supernatant was removed, passed through a PD desalting column, and merozoites were counted in a hemocytometer and either processed directly or stored in liquid nitrogen. The cells were covered with fresh DMEM culture medium and re-incubated until no further parasite development was observed.

DNA and RNA extraction

High molecular weight DNA was extracted from 2 x 10⁸ bradyzoites using a phenol/chloroform protocol [42]. RNA was extracted from freshly excysted sporozoites, merozoites from cell culture supernatant, and freshly isolated bradyzoites with the RNeasy Mini Kit (Qiagen, Hilden, Germany) with additional on-column DNase treatment (RNase-Free DNase Set, Qiagen). In total, three replicates of sporozoites (6.8 x 10⁵–3.23 x 10⁶) and both four replicates of merozoites (1.6 x 10⁵–2.11 x 10⁶) and bradyzoites (1.6 x 10⁹) were processed. Extracted RNA amount and quality were assessed photometrically with the NanoDrop 2000c (Thermo Fisher Scientific, Wilmington, DE, U.S.A.), followed by immediate freezing at -80°C.

Whole genome sequencing

The library was prepared using the ultra-low input whole genome sequencing workflow according to protocol version 02 (“Preparing HiFi SMRTbell Libraries from Ultra-low DNA Input (PacBio, Version 02; August 2020)), with adaption to the SMRTbell® Prep Kit 3.0 (PacBio, Pacific Biosciences, CA, U.S.A.). In summary, 20 ng of DNA was sheared using a g-TUBE (Covaris, MA, U.S.A.), with a centrifugation program with incremental speed settings aiming at a mean fragment size of 12–15 kb. The fragment size distribution was quality controlled by running a dilution of the sheared sample on a Femto Pulse (Agilent, Agilent Technologies, CA, U.S.A.) with FP-1002 run protocol (165 kb). After assessing the size distribution, the sheared sample was used as input for the library preparation according to the manufacturers’ instructions. In short, the libraries were purified using 0.8X ProNex beads (Promega, Madison, U.S.A.) and eluted in 49 µl elution buffer after incubation at 37°C for 20 min. A whole genome amplification was performed with the gDNA Sample Amplification Kit (Pacific Biosciences, CA, U.S.A.) according to the above-mentioned protocol with the following modifications: The elongation time for both PCR programs (Reaction mix A and Reaction mix B) were increased to 10 min and used for 15 cycles. The concentration of the amplified library was determined with the Qubit dsDNA HS Assay (Thermo Fisher Scientific, MA, U.S.A.) and size distribution was analyzed on a Fragment Analyzer (Agilent) with DNF-464 run protocol using a 1 ng/µl dilution in TE. The amplified libraries of the two reaction mixes were pooled and a second library preparation was performed with the SMRTbell Prep Kit 3.0 (PacBio) following the protocol version 02 (Procedure-checklist-Preparing whole genome and metagenome libraries using SMRTbell prep kit 3.0, PacBio; Version 01, April 2022) with the modification of increasing the ligation time of the SMRTbell adapters to 60 min. A nuclease treatment was performed to remove non-circular library templates and the library was size-selected on a BluePippin (Sage Science, MA, U.S.A.) with the following settings: 0.75% DF Marker S1 High-Pass 6–10 kb vs3 using a cutoff at 6 kb. The size-selected library was purified with SMRT bell cleanup beads (PacBio) and eluted in 11 µl elution buffer. The concentration and size distribution of the library were assessed with the Qubit dsDNA HS kit (Thermo Fisher Scientific) and a Fragment Analyzer (Agilent), respectively. Sequencing primer (v3.2) and polymerase were bound to the library using the Sequel® II Binding Kit 3.2 (PacBio) and the bound library was sequenced on a Sequel IIe instrument using a Sequel II SMRT Cell (8M) with a final on plate loading concentration of 85 pM, 2 hrs pre-extension and 30 hrs movie time. The circular consensus reads were generated with SMRT Link version 11, using a minimum accuracy of 0.99 and 3 passes.

RNA sequencing

A total of 25 ng of RNA per sample was used to enrich for polyadenylated mRNA followed by cDNA sequencing library preparation utilizing the Illumina^® Stranded mRNA Prep Kit (Illumina, San Diego, CA, U.S.A.) according to the manufacturer’s instructions. After library quality control by capillary electrophoresis (4200 TapeStation, Agilent), cDNA libraries were sequenced on the Illumina NovaSeq 6000 platform generating 150 bp paired-end reads.

Genome assembly and annotation

First, a k-mer spectrum for the PacBio HiFi reads was computed using the Jellyfish 2 [43] modules count and histo with a k-mer size of 25 in canonical mode. According to the single peaked k-mer spectrum, a hifiasm [44] assembly was carried out with the options --primary --n-hap 1 --hom-cov 80. The resulting contigs were screened for assembly duplication using the “comp” module of the k-mer analysis toolkit KAT [45] and BUSCO [46] with the coccidia database (coccidian_odb10).

The genome annotation was done in several steps. First, a repeat library was constructed using RepeatModeler [47], then the assembled contigs were soft-masked with the repeats using RepeatMasker [48]. The RNA-seq reads were merged using FLASH [49] (-M 100 -O) and aligned to the contigs using STAR [50] in two-pass mode. Reference genomes and proteomes of 11 T. gondii strains and 2 S. neurona strains were downloaded from ToxoDB (Release 67) [51]. Then, three different methods for gene model predictions were computed. First, a de novo prediction was computed using BRAKER3 [52] including the RNA-seq read mappings and the ToxoDB proteomes. Second, the ToxoDB annotations were used to compute a homology-based prediction using GeMoMa [53]. For the third prediction, the RNA-seq mappings were assembled separately using StringTie 2 [54] and then merged with the StringTie-merge option. The three gene models were combined using EVM [55] with different weights (BRAKER3 = 4, GeMoMa = 2, StringTie = 1). Quality of the final model was assessed using BUSCO with the coccidia database.

The functional annotation of the gene model was derived from best blast hits against the ToxoDB proteomes using DIAMOND [56] (blastp --iterate --evalue 10^-20) and an InterProScan [57] search including GO terms. Gene families were identified by a best bidirectional blastp approach of the predicted gene models against reference genes from T. gondii.

The Orthofinder algorithm [58] was applied to identify groups of orthologous genes (orthogroups) between S. calchasi, T. gondii (ToxoDB) and S. neurona (ToxoDB). The orthogroups were used to identify syntenic blocks between the three genomes. DIAMOND blastp results of the orthologous genes were re-computed and the MCScanX tool [59] was applied to compute the syntenic blocks. Orthogroups and species-specific genes of S. calchasi, S. neurona and T. gondii were counted with BioVenn [60].

All PacBio and Illumina sequencing datasets and their assembly dataset were deposited at the European Nucleotide Archive (ENA) with the study identifier PRJEB77870. Within this study, the PacBio dataset can be accessed under the accession numbers ERS20600716 (raw data) and ERS20818611 (assembled genome). RNA-seq data of bradyzoites are available under accession numbers ERS20818612, -13, -14, and -86, data of merozoites under ERS20818687, -88, -89, and -90, and data of sporozoites under ERS20818705, -739, and -890.

Differential gene expression analysis

The read counts and TPM (transcripts per million) values were estimated using Salmon (v2) [61]. The Salmon index was constructed using decoy sequences according to the Salmon manual. Differentially expressed genes were detected using pair-wise comparisons of the samples using the DESeq2 algorithm [62]. The appropriateness of the assumptions underlying the DESeq2 normalization was evaluated using the degCheckFactors function of the DEGreport R module [63]. Genes with a log2 fold change greater than 1 or smaller than -1 and an FDR adjusted p-value below 0.01 were considered differentially expressed.

In situ-hybridization

Four SRS mRNA sequences (ScSRS4, ScSRS8, ScSRS10, and ScSRS12) were localized on microscopical tissue sections by in situ-hybridization. Samples were prepared from liver and pectoral muscle of experimentally infected pigeons. Liver tissue was collected from pigeons in the acute (schizogonic) phase of PPE at dpi 10, pectoral muscle from pigeons in the chronic (cyst-bearing) phase of PPE at dpi 59. In situ-hybridization was performed using the RNAscope 2.5 HD Reagent Kit-Red (Advanced Cell Diagnostics, U.S.A.) according to manufacturer´s instructions. In brief, formalin-fixed, paraffin embedded tissue sections of 5 µm were deparaffinized with xylene and rehydrated with graded ethanols. For liver samples, hydrogen peroxide was applied for 10 min at room temperature. As minor adjustment from the protocol, muscle samples were incubated with Levamisole (Vector Laboratories, Burlingame, U.S.A.) instead, followed by washing in distilled water. Slides were boiled in RNAscope Target Retrieval Reagents for 15 min. After applying a hydrophobe barrier, slides were incubated with supplied Protease Plus for 30 min at 40°C in a wet chamber. Each of four custom-designed probes targeting ScSRS8, ScSRS4, ScSRS10, ScSRS12 (Table 1) was hybridized for 2 hrs at 40°C followed by multiple amplification steps. A probe targeting the mRNA of dihydrodipicolinate reductase (DapB) was used to control for unspecific binding. For bright field evaluation a red signal was achieved by using a Fast Red chromogen. Slides were counterstained using Mayer’s hematoxylin.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Probes for ScSAG4, ScSAG8, ScSAG10, and ScSAG12 mRNA.

https://doi.org/10.1371/journal.pone.0322226.t001

qPCR

For confirmation of RNA-seq data, the mRNA expression of selected surface antigens (ScSRS4, ScSRS8, ScSRS9, ScSRS10, ScSRS12, ScSRS15), rhoptry bulb (ScROP9, ScROP20, ScROP29, ScROP30) and rhoptry neck (ScRON2, ScRON2_L1, ScRON9) proteins, and microneme proteins (ScMIC2, ScMIC12, ScMIC13, ScMIC15, ScAMA1a, ScAMA1b, ScAMA4) was quantified by RT-qPCR in RNA isolated from sporozoites, merozoites and bradyzoites. Primers were designed with Primer–BLAST [64] for the selected genes as well as potential reference genes (S1 Table). Two reference genes were selected that showed similar expression levels in all three stages. Templates of cDNA were amplified with HotStarTaq DNA Polymerase (Qiagen). Efficiency of primers was determined by triplicate testing of serial dilutions of S. calchasi gDNA. Each gene was tested simultaneously with 2 reference genes in triplicate in 3 replicates of sporozoite and merozoite RNA and 4 replicates of bradyzoite RNA. Differential gene expression was calculated by dCt analysis with adjustment of PCR efficiency [65] by qPCRsoft software (Analytik Jena, Jena, Germany). The genes were clustered by hierarchical clustering using the Eucledian distance metric and average linkage. Differences between dCt values were statistically analyzed with the unpaired student’s t-test.

Sarcocystis calchasi replicates in primary pigeon embryo liver cells

Propagation of S. calchasi in permanent cell lines was successful in Vero cells once but, after passages of the parasite through its natural hosts, further attempts were unsuccessful. Likewise, the avian cell line CEC-32 (quail fibroblasts) could not be infected by S. calchasi sporozoites. In contrast, primary pigeon embryo liver cells and their surrounding fibroblasts were readily and reproducibly infected. Successful infectious doses for 25 mL flasks ranged from 1.04 x 10⁵ to 1.7 x 10⁶ sporozoites. Large numbers of free merozoites in the supernatant (6.77 x 10⁵–2.11 x 10⁶) were observed from dpi 5 onwards and could be maintained until dpi 13 the longest. Thereafter, the release of merozoites declined and around dpi 21 the monolayer of primary cells dissolved. Attempts to passage free merozoites onto fresh primary cells were unsuccessful.

The S. calchasi genome resembles the S. neurona genome with few genes unique to the genus Sarcocystis

PacBio platform Sequel IIe generated 34.72 Gb sequence data from 3,046,204 reads encoded in 75 contigs. The assembly N50 value was 6 Mbp with a maximum contig length of 12.50 Mbp (mean length: 1.52 Mbp). The k-mer spectrum showed a single peak, which indicates a haploid or a highly homozygous genome. The size of the assembled genome of 113.89 Mbp is similar to the S. neurona SN1 genome (127 Mbp) and roughly twice the size of the T. gondii genome (ME49: 65.67 Mbp). Like in S. neurona, large parts of the S. calchasi genome consist of repetitive elements (30.11 Mbp) whereas the T. gondii genome includes as few as 2.5 Mbp. Quality of genome assembly was assessed with KAT and BUSCO analysis. Both analyses did not show any significant duplication. Of 502 BUSCO groups in the lineage dataset of 20 coccidia genomes, 495 BUSCOs (98.6%) were complete indicating an almost complete assembly. Gene annotation comprised several methods (see Material and methods) and included RNA-seq data from 11 S. calchasi samples (4 x bradyzoites, 4 x merozoites, 3 x sporozoites) with an average size of 80.69 Mio. reads. From these data, 9,716 coding sequences (CDS) were predicted, 118 rRNA sequences and 89 tRNA sequences. For 6,730 CDS homologs were found in the ToxoDB database and 4,609 InterPro hits were generated. In total, 6,508 CDS were functionally annotated which is similar to S. neurona with 6,938 genes annotated in the SN3.E1 S. neurona genome. Quality of annotation was again assessed by BUSCO analysis in protein mode. Of 502 BUSCO groups, 497 (99.0%) were complete. Orthofinder identified a large core genome in the family Sarcocystidae with 4,712 shared orthogroups in Toxoplasma strain ME49, S. neurona strain SN3 and S. calchasi isolate Giessen16. The number of orthogroups shared by T. gondii and S. calchasi (712) is more than twice the number shared by S. neurona and S. calchasi (320) and the number of genes unique to S. calchasi (3,409) is more than twice the number of unique genes in S. neurona (1,446) and T. gondii (1,665) (Fig 1). To evaluate the accuracy of gene prediction in the different sets of unique and shared genes, the percentage of S. calchasi genes with TPM <1 was assessed, averaged over all samples (Table 2). This revealed a large number of genes that were either low expressed or erroneously predicted in genes unique to S. calchasi (40.00%) or shared by S. calchasi and T. gondii (34.02%). Regarding the ordering of homologous genes, S. calchasi and S. neurona genomes share large syntenic blocks suggesting a high level of common chromosome arrangement (Fig 2). In contrast, S. calchasi and T. gondii genomes are much less syntenic as it has been observed for S. neurona and T. gondii [23].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Percentage of S. calchasi genes with TPM <1, averaged over all samples.

https://doi.org/10.1371/journal.pone.0322226.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Core genome of Sarcocystis spp. and T. gondii.

Euler diagram of orthogroups shared by S. calchasi (red), S. neurona (green) and T. gondii (blue). Figures refer to the numbers of orthogroups with unique or overlapping presence.

https://doi.org/10.1371/journal.pone.0322226.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Synteny between S. calchasi, S. neurona and T. gondii genomes.

S. calchasi (top) and S. neurona (middle) scaffolds are in large parts syntenic reflecting their close relationship. In comparison, T. gondii (bottom) chromosomes do not share large syntenic parts with Sarcocystis spp. scaffolds.

https://doi.org/10.1371/journal.pone.0322226.g002

Large parts of the S. calchasi transcriptome are differentially expressed in sporozoites, merozoites and bradyzoites

Differential gene expression analyses with DESeq2 require the majority of genes not to be differentially expressed. The analysis of the S. calchasi RNA-seq samples showed a bias between the samples, as shown in Fig 3. It is expected to see all samples peak at zero, whereas some of the S. calchasi samples peaked at around -0.5 or 0.5. This might be an indicator that the DESeq2 analysis overestimated the number of differentially expressed genes. Additional methods were applied to support the findings of differential gene expression. To confirm RNA-seq data, 20 selected SPD genes that were differentially expressed according to DESeq2 analysis were tested by qPCR. For dCt analysis, 18SrRNA and GAPDH-1 genes were selected as reference genes as they showed best congruence of Ct values in S. calchasi sporozoite, merozoite and bradyzoite cDNA. Hierarchical clustering of these 20 genes confirmed TPM values by revealing two clusters of genes being expressed in higher levels in sporozoites and merozoites and one cluster of genes with higher expression levels in bradyzoites (Fig 4). Although the difference in expression in the last cluster was not as pronounced, most bradyzoites showed significantly higher expression levels by student’s t-test in all tested genes (S1 File). Additionally, in situ-hybridization of four SRS encoding genes (ScSRS4, 8, 10, and 12) in host tissues again confirmed RNA-seq and qPCR data regarding differential expression in merozoites and bradyzoites. In conclusion, although RNA-seq data needed to be interpreted carefully, TPM values could be confirmed by two further methods of measuring gene expression and therefore were considered to be reliable.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Quality assessment of RNA-seq data from S. calchasi bradyzoites, merozoites and sporozoites.

The gene ratios between each gene and the average gene peak around zero if the majority of genes are not differentially expressed. Most S. calchasi samples show a bias from the expected peak at zero. This could affect the results of the differential gene expression analysis. The bandwidth was selected using the rule-of-thumb method. The number of genes in each sample ranged from 8,203 in Mero_2 to 8,884 in Mero_1.

https://doi.org/10.1371/journal.pone.0322226.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Heatmap of dCt analysis of selected SPD encoding genes in the S. calchasi genome.

dCt values are calculated as mean (Ct[reference genes]) - Ct[gene of interest]). Gray cells mark PCRs with no measured Ct value. The clustered genes revealed three major clusters. Clusters #1 (bottom) and #3 (middle) contain genes being expressed at higher levels in sporozoites and merozoites compared to bradyzoites with a more pronounced difference in cluster #3. Cluster #2 (top) contains genes with slightly higher expression in bradyzoites.

https://doi.org/10.1371/journal.pone.0322226.g004

Of 9,716 genes detected in the S. calchasi transcriptome, 4,162 were significantly differentially expressed (p < 0.01) in bradyzoites compared to merozoites. In bradyzoites and sporozoites, 5,404 genes showed different transcript levels. Least expression differences were observed in merozoites compared to sporozoites with 3,356 differentially expressed genes. Approximately the same number (4,111 of 6,784) was differentially expressed when comparing S. neurona merozoites to schizonts [24] indicating a high level of stage-specific gene expression in Sarcocystis species. The number of significantly up- or down-regulated genes in bradyzoites, merozoites and sporozoites are summarized in Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Differentially expressed genes in S. calchasi bradyzoites, merozoites and sporozoites. DESeq2 analysis on Salmon-estimated counts.

https://doi.org/10.1371/journal.pone.0322226.t003

SRS encoding genes are expressed either in sporozoites and merozoites or in bradyzoites

In the S. calchasi genome 11 SRS encoding genes were identified. The 11 SRS sequences included homologs of genes encoding immunodominant proteins in S. neurona, SnSAG1 (SnSRS10), SnSAG2 (SnSRS8), SnSAG3 (SnSRS12), and SnSAG4 (SnSRS4). Transcripts of the corresponding homologs in S. calchasi (ScSAG1, ScSAG2, ScSAG3, and ScSAG4) were abundant in merozoites and sporozoites but not or very low expressed in bradyzoites. Expression levels were determined by TPM values from RNA-seq and dCT values from qPCR analysis (S2 Table; Fig 4). Highest expression levels were detected for ScSAG3, followed by ScSAG2 and lowest expression was measured for ScSAG4 and ScSAG1. RNA-seq and qPCR data agreed with each other except for the expression of ScSAG1 in sporozoites detected by qPCR but not by RNA-seq. This order of expression intensity was contrasted by in situ-hybridization data from liver tissue with schizonts and free merozoites (Fig 5). The strongest signal was detected in the slides hybridized with the ScSAG1 and ScSAG2 probes. ScSAG1 and ScSAG2 mRNA was localized to schizonts as well as free merozoites. Weaker signals were obtained with the ScSAG3 probe mainly in schizonts while the ScSAG4 probe yielded weak signals from free merozoites. In bradyzoites, ScSAG1, ScSAG2, ScSAG3, and ScSAG4 expression was not or at a very low level detected by RNA-seq, qPCR and in situ-hybridization. Of the remaining S. calchasi SRS encoding genes, ScSRS11 was also found expressed by sporozoites and merozoites while ScSRS9 and ScSRS15 were predominantly expressed in bradyzoites. Four further SRS sequences (ScSRS14, ScSAG16, ScSAG17, ScSAG20) were not expressed in all three invasive stages (Fig 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Microscopical tissue localization of ScSAG1-4 mRNA via in situ-hybridization.

A - D: Sarcocystis calchasi merozoites and schizonts in the liver at dpi 10. ScSAG1 (A) and ScSAG2 (B) were strongly expressed in schizonts (arrows) as well as in merozoites (arrowheads). Weaker signals were detected for ScSAG3 mainly in schizonts (C). In contrast, ScSAG4 was predominantly expressed in merozoites (D, arrowheads). E – H: While Sarcocystis mRNA encoding ScSAG2 was weakly detected in few bradyzoites within muscle cysts at dpi 59 (F), mRNA encoding ScSAG1, -3 or -4 was not detected in parasitic cysts at all (E, G, H). All: in situ-hybridization with Fast Red as chromogen (red) and hematoxylin as counterstain (blue). Scale bars, 40 µm.

https://doi.org/10.1371/journal.pone.0322226.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Expression patterns of S. calchasi SPDs.

The majority of SRS and MIC encoding genes were expressed predominantly in sporozoites and merozoites while genes of rhoptry proteins were mainly expressed in bradyzoites. The expression level of RON encoding genes was generally low although detected mostly in bradyzoites. Green: genes mainly expressed in bradyzoites; orange: genes mainly expressed in sporozoites and merozoites; purple: genes not or very low expressed in bradyzoites, sporozoites and merozoites; N = Nucleus; AP = Amylopectin; Created in BioRender. Maier-Sam, K. (2025) https://BioRender.com/e35w735.

https://doi.org/10.1371/journal.pone.0322226.g006

Genes associated with host cell invasion are well conserved in S. calchasi

In addition to SRS encoding genes, sequences of numerous factors of the host cell invasion machinery of apicomplexan parasites were identified in the S. calchasi genome. Two paralogs of AMA1 (ScAMA1a and ScAMA1b) as well as homologs to RON2, RON4, RON5, and RON8 were identified in S. calchasi suggesting that the invasion process is conserved. Of the sporozoite and bradyzoite associated T. gondii paralogs, AMA4, RON2_L1, and RON2_L2 homologs were identified in the S. calchasi genome. Stage-specific expression patterns of S. calchasi genes associated with host invasion differed from Toxoplasma and S. neurona. ScAMA1a and ScAMA1b whose S. neurona orthologues had not been found expressed in S. neurona merozoites (23) were differently expressed in S. calchasi invasive stages (Fig 4). ScAMA1a showed high expression levels in sporozoites and merozoites but no expression in bradyzoites. Expression of ScAMA1b could not be evaluated conclusively as RNA-seq data differed slightly from qPCR data. While RNA-seq implied exclusive expression of ScAMA1b in bradyzoites, qPCR detected ScAMA1b transcripts in sporozoites and merozoites without significant difference to the higher expression level in bradyzoites. ScAMA4 was expressed predominantly in bradyzoites. Sarcocystis calchasi RON sequences showed very low expression levels in bradyzoites with some expression of ScRON5 and ScRON2 also in sporozoites and merozoites (Figs 4 and 6).

Sarcocystis calchasi homologs of all RON encoding genes described in T. gondii were identified including homologs to RON1 and RON10 which had not been found in S. neurona [23].

In S. calchasi, homologs to CRMPA and CRMPB as well as to MIC15 and TSP-1 were annotated in its genome. While CRMPB and MIC15 were mainly expressed in S. calchasi sporozoites and merozoites, CRMPA and TSP-1 were also expressed in bradyzoites. Two other factors associated with rhoptry secretion and therefore essential for host cell invasion in T.gondii are micronemal proteins MIC8 [66] and the claudin-like apicomplexan microneme protein (CLAMP) [67]. Homologs of genes encoding for both proteins in S. calchasi had very low expression levels in all three invasive stages.

MIC encoding genes show a similar expression pattern to SRS genes

From 20 MIC sequences described in T. gondii, 14 MIC sequences were identified in the genome of S. calchasi. The set of MIC encoding genes resembles that in S. neurona except for M2AP and MIC10 (absent in S. calchasi) and MIC11, MIC17A and MIC17B/C (absent in S. neurona). Regarding expression of MIC encoding genes in S. calchasi invasive stages, a clear distinction was evident between MIC genes being expressed predominantly in sporozoites and merozoites (ScMIC2, ScMIC4, ScMIC7, ScMIC8, ScMIC13, ScMIC15, ScMIC16, ScMIC17A) and those MIC genes that were expressed mainly in bradyzoites (ScMIC9, ScMIC11, ScMIC12, ScMIC14) (Fig 6). Expression levels obtained from selected MIC sequences (ScMIC2, ScMIC12, ScMiC13, ScMIC15) as detected by qPCR (Fig 4) confirmed RNA-seq data. Two further MIC genes (ScMIC3 and ScMIC17B/C) which showed low sequence homology to their T. gondii counterpart were not expressed in any of the three stages, possibly suggesting a potential role of these MICs in intracellular development.

Few ROP and GRA encoding genes were annotated in the S. calchasi genome

In S. calchasi, 14 ROP homologs were identified. Nine of these homologs (ScROP17/41, ScROP20, ScROP21/27, ScROP29, ScROP30, ScROP33, ScROP35, ScROP38, ScROP39) are classified in T. gondii as known or likely catalytic kinases, one (ScROP22) is classified as pseudokinase and one (ScROP24) as kinase with potential nonchanonical mechanism [68]. ScROP18/36 shares equal homology to ROP18 (a kinase in T. gondii) and ROP36 (a pseudokinase in T. gondii). The function of the T. gondii homolog of the remaining two ROPs (ScROP9, ScROP14) remains unknown [68]. In S. neurona, 18 homologs of ROP encoding genes have been described [23,24], five of which being absent from S. calchasi (ROP19, ROP26, ROP28, ROP37, ROP40). The sequence of another soluble rhoptry protein in T. gondii, the metalloprotease Toxolysin-1 (TLN1), was annotated in the S. neurona and the S. calchasi genome. Overall, the total number of ROP genes in Sarcocystis spp. appeared rather low compared to T. gondii or N. caninum [68]. Transcripts of most ROP genes (n=8) were found predominantly in bradyzoites although two (ScROP35, ScTLN1) had very low transcript levels. Three ROP genes (ScRO9, ScROP17/41, ScROP20) were mainly expressed in sporozoites and merozoites. Four ROP genes were not expressed in either of the three stages (Fig 6).

In S. neurona, 4 homologs to GRA encoding genes have been identified [23,24], two of which (GRA10 and DG32 syn. GRA23/17) were annotated in the S. calchasi genome. Expression levels of ScGRA10 were equally moderate in sporozoites, merozoites and bradyzoites while DG32/GRA23/17 was expressed almost exclusively by bradyzoites.

Genes associated with the inner membrane complex are well conserved and do not exhibit a distinct expression pattern

In S. calchasi homolog genes of most alveolins were identified. In total, homolog genes of 33 IMC proteins and all four IMC sub-compartment proteins (ISP) were present in the S. calchasi genome. Putative homolog genes of glideosome-associated proteins GAP40, GAP45, GAP50 and GAP70 were annotated in the S. calchasi genome. Other IMC proteins that are not associated to either alveolins or the glideosome are PhIL1, which is located at the apical end of the IMC [69], and MORN1 which links the IMC to the cytoskeleton [35]. Homolog genes to PhIL1 and MORN1 were also identified in the S. calchasi genome. The occurrence of genes encoding for IMC proteins in S. neurona and S. calchasi seems to be congruent with few exceptions (e.g., absence of GAP70 in S. neurona). Expression levels of genes encoding IMC proteins in S. calchasi sporozoites, merozoites and bradyzoites varied from genes being highly expressed in all three stages (e.g., IMC20, IMC39), genes highly expressed predominantly in a single stage (e.g., ScIMC13 in merozoites or ScPhIL1 in bradyzoites) and genes with overall very low TPM values (e.g., ScIMC5, ScIMC15).