Toxoplasma ceramide synthases: Gene duplication, functional divergence, and roles in parasite fitness

Abstract Toxoplasma gondii is an obligate, intracellular apicomplexan protozoan parasite of both humans and animals that can cause fetal damage and abortion and severe disease in the immunosuppressed. Sphingolipids have indispensable functions as signaling molecules and are essential and ubiquitous components of eukaryotic membranes that are both synthesized and scavenged by the Apicomplexa. Ceramide is the precursor for all sphingolipids, and here we report the identification, localization and analyses of the Toxoplasma ceramide synthases TgCerS1 and TgCerS2. Interestingly, we observed that while TgCerS1 was a fully functional orthologue of the yeast ceramide synthase (Lag1p) capable of catalyzing the conversion of sphinganine to ceramide, in contrast TgCerS2 was catalytically inactive. Furthermore, genomic deletion of TgCerS1 using CRISPR/Cas‐9 led to viable but slow‐growing parasites indicating its importance but not indispensability. In contrast, genomic knock out of TgCerS2 was only accessible utilizing the rapamycin‐inducible Cre recombinase system. Surprisingly, the results demonstrated that this “pseudo” ceramide synthase, TgCerS2, has a considerably greater role in parasite fitness than its catalytically active orthologue (TgCerS1). Phylogenetic analyses indicated that, as in humans and plants, the ceramide synthase isoforms found in Toxoplasma and other Apicomplexa may have arisen through gene duplication. However, in the Apicomplexa the duplicated copy is hypothesized to have subsequently evolved into a non‐functional “pseudo” ceramide synthase. This arrangement is unique to the Apicomplexa and further illustrates the unusual biology that characterize these protozoan parasites.

and TgCerS2.Interestingly, we observed that while TgCerS1 was a fully functional orthologue of the yeast ceramide synthase (Lag1p) capable of catalyzing the conversion of sphinganine to ceramide, in contrast TgCerS2 was catalytically inactive.Furthermore, genomic deletion of TgCerS1 using CRISPR/Cas-9 led to viable but slow-growing parasites indicating its importance but not indispensability.In contrast, genomic knock out of TgCerS2 was only accessible utilizing the rapamycin-inducible Cre recombinase system.Surprisingly, the results demonstrated that this "pseudo" ceramide synthase, TgCerS2, has a considerably greater role in parasite fitness than its catalytically active orthologue (TgCerS1).
Phylogenetic analyses indicated that, as in humans and plants, the ceramide synthase isoforms found in Toxoplasma and other Apicomplexa may have arisen through gene duplication.However, in the Apicomplexa the duplicated copy is hypothesized to have subsequently evolved into a non-functional "pseudo" ceramide synthase.This arrangement is unique to the Apicomplexa and further illustrates the unusual biology that characterize these protozoan parasites.

| INTRODUCTION
The Apicomplexa is a phyla of infectious protozoa comprising pathogens of both domestic animals and humans. 1 These include the haemosporidian Plasmodium spp., including P. falciparum the causative agent of severe malaria; the coccidian Cryptosporidium spp.(diarrhea), Eimeria spp.(coccidiosis in poultry and cattle) and Toxoplasma gondii (toxoplasmosis). 2 Toxoplasma is an obligate intracellular parasite with the ability to invade and colonize a broad range of nucleated cells from warm-blooded vertebrates; therefore, toxoplasmosis is one of the most prevalent infections and is estimated to affect 2-3 billion people worldwide. 3Importantly, infection is a significant cause of congenital defects and subsequent abortions in both economically important domestic animals 4 and humans. 5However, treatment regimens are limited and have remained largely unchanged since the 1950s, 6 and there are no therapies available for chronic infection with encysted bradyzoite forms. 7To this end efforts are focused on identifying new drug targets.Recent studies have shown major differences in the lipid, particularly sphingolipid, profiles of apicomplexans compared to the host. 3,8,9][17] In proliferative tachyzoite infection sphingolipids can be scavenged from the host 8,18,19 and Toxoplasma have been shown to possess sphingomyelin, the primary mammalian phosphosphingolipid. 3 However, they have also been demonstrated to harbor a relatively high level of ethanolamine phosphorylceramide, which is only found as a trace phosphosphingolipid in the host cell, 3 and non-mammalian inositol phosphorylceramide. 8In addition, host sphingolipid biosynthesis is non-essential for Toxoplasma proliferation, 8,19 and the parasite retains a fully functional biosynthetic pathway 8,18,20 which is active throughout its lifecycle. 21This indicated that de novo synthesis is important for parasitism and a potential target for therapeutic intervention.
Ceramide, the product of ceramide synthase (CerS or Lag1p [longevity-assurance gene] in yeast), is the central molecule among the sphingolipid precursors which functions in a plethora of biological processes, including apoptosis, growth arrest, and stress responses. 22The first, rate-limiting stage in sphingolipid biosynthesis is mediated by serine palmitoyltransferase (SPT), which catalyzes the condensation of L-serine and, typically, palmitoyl-CoA to form 3-keto-di-hydro-sphingosine (KDS or 3-keto-sphinganine) 23 in the endoplasmic reticulum (ER).Subsequently, KDS is reduced and then N-acylated by CerS to form ceramide, the central metabolic hub of sphingolipid biosynthesis, which is then transported to the Golgi apparatus where it is converted into glycosphingolipids (GSLs) and, in mammals, sphingomyelin (SM). 24,25lthough Toxoplasma and the other apicomplexans maintain this pathway, it is in many ways dramatically divergent when compared to the mammalian cell machinery. 22For example, the first functionally characterized enzyme in the apicomplexan biosynthetic pathway was a functional orthologue of the yeast inositol phosphorylceramide (IPC) synthase, an enzyme with no mammalian equivalent. 8,26Furthermore, the apicomplexan SPT is, uniquely in the Eukaryota, of bacterial origin and presumed to have been acquired via lateral gene transfer. 20n addition, both orthologues of the Toxoplasma SPT (1 and 2) have recently been demonstrated to be involved ceramide regulation in vitro and in both acute and chronic in vivo infection. 27However, the characterization of the apicomplexan CerS, the second key step in eukaryotic sphingolipid biosynthesis (Figure 1), remained absent.Here, we describe and characterize the Toxoplasma Tg-CerS1 and TgCerS2, which also have orthologues in the wider Apicomplexa.Utilizing a combination of biochemical, genetic and phylogenetic tools we demonstrate that while TgCerS1 is a functional CerS, TgCerS2 lacks this activity in vitro.However, while functional TgCerS1 plays little role in parasite fitness, TgCerS2 was indicated to have a particularly important role.Phylogenetic analyses indicated that the apicomplexan "pseudo" ceramide synthase, TgCerS2 and orthologues, may be the result of an ancient gene duplication event and a subsequent divergence that is unique to the Apicomplexa.

DNA and RNA
Genomic DNA was extracted from T. gondii tachyzoites to use as PCR template by pelleting parasites and resuspending in PBS.DNA extraction was then performed using the QIAamp DNA Blood Mini Kit (Qiagen) as per the manufacturer's protocol.RNA was isolated from freshly purified tachyzoites of Toxoplasma gondii using RNeasy Micro Kit (Qiagen) and subsequently reverse-transcribed into firststrand cDNA using Qiagen One Step RT-PCR Kit System.

| Ceramide synthase assay
Following determination of expression, equivalent concentrations of TgCerS1 and TgCerS2, plus controls, were assayed for ceramide synthase activity in LoBind tubes (Eppendorf) as follows.A 100 μL reaction volume with   31 The NBD lipids were detected using a Typhoon FLA 9500 biomolecular imager (GE Healthcare Life Sciences) operated in Cy2 fluorescence mode with 473 nm excitation laser, BPB1 filter, 50 μm pixel size, and PMT voltage setting of 290 V. Data were processed using ImageLab (Bio-Rad Laboratories) to adjust intensity.

| Cloning and subcellular localization of TgCerS1 and TgCerS2
Primers were designed to amplify and FLAG tag the TgCerS1 and TgCerS2 coding sequences: TgCerS1Xpres s/TgCerS1FLAGXpress_R and TgCerS2Xpress_F/TgCe rS2FLAGXpress_R from the synthetic genes (Table S2).The resultant PCR products were introduced into NheI (NEB) linearized ToxoXpress vector (pTXP; designed, created, and validated in-house; Figure S1) using the In-Fusion HD cloning kit (TaKaRa) as per manufacturer's protocol (Table S3) to create pTXP_TgCerS1-FLAG and pTXP_TgCerS2-FLAG.All constructs were verified by sequencing (Eurofins Genomics) prior to further work.Transient transfections were carried out using a 4D Nucleofector (Lonza), protocol FI 158.Briefly, parasites freshly lysed from HFFs monolayer were homogenized by passage through a 25-gauge needle and isolated by centrifugation at 1500 × g for 10 min at 4°C.The pellet was resuspended in P3 buffer with added supplementary buffer P1 (Lonza).20 μL of the parasite suspension (10 7 mL −1 ) were added to a dried pellet of ethanol-precipitated pTXP_TgCerS1-FLAG or pTXP_TgCerS2-FLAG plasmid and electroporated.Subsequently, 100 μL of media was added, and 10 μL or 20 μL were transferred to 24-well plates containing confluent HFF cells grown on glass coverslips before incubation at 37°C, 5.0% CO 2 .Cells were fixed with 4% paraformaldehyde in PBS (pH 7.4) for 15 min and then permeabilized with 0.4% (v/v) Triton X-100 in PBS for 10 min, before incubation in blocking buffer (PBS supplemented with 1% [w/v] BSA (Sigma-Aldrich), 0.1% fish skin gelatin (Sigma-Aldrich) and 0.1% [v/v] Triton X-100) for 15 min at room temperature.Samples were incubated with polyclonal rat anti-TgSPT1_Δ158 (1:200) 20,27 and mouse monoclonal anti-FLAG antibody (1:200; AbCam) in blocking buffer overnight at 4°C.After PBS washing, samples were incubated with Alexa Fluor® 488 anti-rat and Alexa Fluor® 647 anti-mouse secondary antibodies (ThermoFisher) at 1:500 in blocking buffer for 1 h at room temperature.The samples were incubated with Hoechst 33342 (Sigma-Aldrich) in PBS for 10 min, mounted using Vectashield H-1000 (Vector labs) and sealed before imaging.

| Construction of plasmids for CRISPR/Cas9-based manipulation
CRISPR/Cas-9 was utilized to accelerate the generation of the conditional ΔTgCerS1 and ΔTgCerS2 lines and also to create CRISPR/Cas-9 KOs.All CRISPR/Cas9 plasmids designed and used in this study were created by using Q5 Site-Directed Mutagenesis Kit (NEB-E0554S) based on the previously described CRISPR/Cas9 plasmid 32 as primary scaffold, as per the manufacturer's protocol.All constructs were verified by sequencing (Eurofins Genomics) prior to further work.Primers for TgCerS1 and TgCerS2 sgRNAs were predicted and designed using online tools including EuPaGDT, E-CRISP, and CHOPCHOP (Table S4), targeting the 5′ and 3′ of the respective coding regions to ensure efficient gene deletion and subsequent DNA insertion.CRISPR/Cas-9 plasmids were transfected along with supplementary DNA cassettes to allow mutant verification and positive selection.In the case of non-conditional knockouts, a GFP reporter gene was introduced to the corresponding genes TgCerS1 and TgCerS2, respectively, after DNA amplification from a previously described vector 33 using designed primers (Table S5).

| Construction of cassettes for inducible recombinase-based knockout
A pUC19 (NEB) plasmid was used as a backbone for the constructs after it was linearized with FD PvuII to customize the subsequent assembly of the appropriate fragments.All DNA amplicons for plasmids constructs were amplified with high fidelity Phusion Polymerase (NEB) to minimize replication errors.pUC19-5′UTR-loxP-TgCers1-2A-KillerRed-loxP-EYFP-3′UTR-5′UTR-HXGPRT-3′UTR and pUC19-5′UTR-loxP-TgCerS2-2A-KillerRed-loxP-EYFP-3′UTR-5′UTR-HXGPRT-3′UTR were designed to be knocked into the TgCerS1 or TgCerS2 loci respectively.Briefly, four different DNA fragments were amplified using primers sets F01, F02, F03, and F04 (Tables S6 and  S7; Integrated DNA Technologies [IDT]).Fragments were cloned into a PvuII linearized pUC19 using In-Fusion HD cloning kit (TaKaRa) as per manufacturer's instructions.All constructs were verified by sequencing (Eurofins Genomics) prior to further work.FastDigest (FD) LguI and a combination of FD SacI/FD Ajul (for TgCers1) and FD XbaI and combination of FD ScaI and FD Hind III (for TgCers2) were used to digest plasmids at 37°C for 30 min and electrophoresed on a 0.8% agarose gel to verify successful assembly.Finally, FD SphI and FD CpoI were used to linearize TgCerS1 DNA cassette and FD HindIII was used to linearize TgCerS2 DNA cassette before electroporation into freshly egressed RH.diCre (parental) parasites as described above.

| Toxoplasma mutant strain selection and verification
All transfections were performed by a 4D Nucleofector (Lonza) as described above.Transfected parasites were incubated overnight at 37°C and 5% CO 2 with media being replaced with complete DMEM containing 25 μg/ mL MPA (Sigma-Aldrich) and 50 μg/mL xanthine (Sigma-Aldrich) the following day and left to grow for at least 8 days with the selection media refreshed every 48 h.Clonal line selection was then performed with newly acquired mutant strains before molecular verification by diagnostic PCR using GoTaq DNA polymerase (Promega) and multiple sets of primers (Figure S2; Tables S8 and S9).In addition, confocal fluorescence microscopy monitoring the red (585/610 nm) to yellow (513/527 nm) fluorescence switch was utilized as described below.CRISPR/Cas-9 non-conditional KOs were also assessed microscopically monitoring GFP signal post-transfection.

TgCerS2 knockouts
TgCerS1 and TgCerS2 conditional knockouts were obtained after the addition of 100 nΜ rapamycin to RH.diCre:CerS1 and RH.diCre:CerS2 strains, respectively.Briefly, tachyzoites were collected from HFFs by scraping cells and passing through a 25G needle by syringe lysis.Following centrifugation at 50 × g for 5 min to allow removal of cell debris, the supernatant was centrifuged again at 500 × g for 5 min.Parasites were then resuspended and incubated for 4 h at 37°C, 5% CO 2 in complete DMEM with 100 nM of rapamycin before washing twice with warm media to remove excess rapamycin.Following viability assessment by trypan blue (0.4% w/v), the parasites were then used to infect HHF monolayers.Toxoplasma parental line was also tested under the same treatment conditions to exclude rapamycin effects on parasite fitness.

Toxoplasma fitness
Plaque and monolayer disruption assays were employed to analyze parasite fitness.In both cases, HHF cells were grown to confluence in 24 (plaque) and 6 (disruption) well plates.100 and 10 5 parasites per well respectively were added and the plates incubated for 12 days or up to 96 h.Subsequently, wells were gently washed with PBS and fixed with 4% paraformaldehyde (pH 7.4) for 20 min at room temperature.0.05% w/v Crystal Violet (CV) solution then replaced the fixative and, after 30 min at room temperature, the wells were gently rinsed twice with water and left to dry before coating with 0.3% agarose.Plaques and disruption areas were photographed using an Olympus inverted microscope CKX53 and a 5 MP XCAM camera 1080p, under 4X magnification.ImageJ was then used to determine plaque sizes and monolayer disruption areas.Values are expressed as mean values of three independent experiments ± SD.P-values were calculated using unpaired two-tailed Student's t-test unless stated otherwise.P-value significance thresholds were set at: *p < .05,**p < .01,***p < .001.

| Phylogenetic analysis
ToxoDB, VEuPathBD, TriTrypBD, PlasmoDB, UniProt were used in order to identify CerS orthologues in various taxa and species.All accession numbers listed in the text aside from: E. necatrix ENH_00074610.1-p1,E. maxima EMWEY_00008780-t26_1-p1, E. mitis EMH_0037690-t26_1-p1, E. acervuline EAH_00007580-t26_1-p1, and E. brunetti EBH_0045800-t26_1-p1. Multiple amino acid alignment was performed using MAFFT and MUSCLE alignment algorithms. 34The predicted CerS amino acid sequences selected for phylogenetic analyses were aligned using ClustalW, 35 manually edited to remove non-aligned regions, and then re-aligned in ClustalW with the output selected as a ClustalW format.Molecular Evolutionary Genetics Analysis (MEGA-X) 36 was used to reconstruct phylogenetic trees and perform gene duplication analysis.Neighbor Joining, UPGMA Tree and Maximum Likelihood methods were used to allow for equal and unequal rates of evolution between the CerS's of different species and taxa.NGphy logeny.frwas also used to visualize MAFFT alignments and identify CerS consensus regions.Sequence accession numbers with respective organism names are listed in Table S10.

| Identification of Toxoplasma gondii ceramide synthases
In all eukaryotic systems studied to date, ceramide synthases (CerS) are ER-resident integral membrane proteins that N-acylate di-hydrosphingosine (sphinganine) to produce di-hydroceramide which is then desaturated to form ceramide, the simplest sphingolipid and a key secondary messenger in numerous cellular pathways. 37First identified as encoded by the longevity-assurance genes Lag1p and Lac1p in yeast, 38 orthologues were subsequently found to be ubiquitous in the Eukaryota.All eukaryotes studied to date have been found to encode at least two CerS orthologues, 39 with humans expressing six isoforms with each generating ceramides with a defined acyl chain length. 39,40All orthologues harbor the conserved 52 amino acid Lag1 motif.Although the precise functionality of this domain remains unclear, 9 of the 11 canonical residues are essential for CerS activity. 41n initial search of the Toxoplasma genome database (toxodb.org/)revealed a single, putatively encoded 383 amino acid CerS orthologue, TGGT1_316450 (named TgCerS1).However, re-probing the database with the identified sequence yielded a paralog, TGGT1_283710 (named TgCerS2) a predicted 342 long amino acid protein which exhibited only 16.8% sequence identity and 34.5% similarity to TgCerS1.TgCerS2 demonstrated higher homology within the Lag1 motif region (25% identity and 42% similarity, CLUSTALW2 and MAFFT alignment); however, it lacked the essential and canonical arginine (R208), double histidine (H217 H218) and tyrosine (Y259) residues (Figure 2A,B; TgCerS1 numbering).All of these were subject to non-conservative changes in TgCerS2, leucine (L169), proline (P178), cysteine (C179) and aspartic acid (D220) respectively (Figure 2A,B; TgCerS2 numbering).While the structurefunction relationships of CerSs are not well studied, their topology displays a conserved consensus of seven predicted transmembrane domains (TMDs) with the active site facing the lumen of the ER 44,45 and their acyl chain preference determined by a region of 11 residues forming an exo-membranous loop. 46TgCerS1 and TgCerS2 maintain the consensus of seven TMDs as predicted by the protein structure prediction software AlphaFold2 43 (Figure 2C,D) and, increasing the confidence in the prediction, RosettaFold 47 (Figure S3).Notably, while both structure prediction programs revealed the same transmembrane architecture, including the Lag1 motif, the extracellular loop structures were predicted with much lower confidence and differed significantly in secondary structure and overall fold.
The CRISPR/Cas9-based genome-wide data generated by Sidik et al., 2016 48 indicated that both orthologues may have only minor a role in parasite fitness and expression profiling demonstrated their presence throughout the lifecycle. 49In addition, utilizing a spatial proteomic approach, hyperplexed localization of organelle proteins by isotope tagging (hyperLOPIT), Barylyuk et al., 2020 50 indicated that like all studied ceramide synthases, TgCerS2 is located in the ER.However, the localization of TgCerS1 was not assigned (Table 1).

TgCerS1 and TgCerS2
The in vitro study of integral membrane proteins is generally challenging.By utilizing a wheat germ-based cell-free membrane protein expression system, both TgCerS1 and TgCerS2 were expressed as FLAG-tagged proteins in in vitro proteoliposomes.Western blot analyses utilizing anti-FLAG rabbit monoclonal antibodies demonstrated that both proteins are produced (Figure 3A).Subsequently, biochemical analyses of the CerS activity within these proteoliposomes was undertaken utilizing fluorescent NBD-sphinganine (NBD-Sph) and acyl-CoA (C16:0 or C24:0).Lipidomic analyses have indicated that the primary ceramide species in Toxoplasma are C-16 and C-18. 3,8In line with this observation, TgCerS1 demonstrated a robust CerS activity with NBD-Sph and C16:0 acyl-CoA as substrates.No activity was observed toward NBD-Sph and C24:0 acyl-CoA (Figure 3B,C).In sharp contrast, TgCerS2 showed no activity to NBD-Sph combined with either of the CoA substrates (Figure 3B,C), in keeping with the absence of the canonical residues in the Lag1 motif (Figure 2A,B).The human orthologues HsCerS2 and HsCerS5 were analyzed in the same manner and used as positive controls for CerS activity with C24:0 and C16:0 acyl-CoA respectively (Figure 3B,C).Like the active TgCerS1, and other eukaryotic orthologues, the "pseudo" ceramide synthase TgCerS2 localized to the ER, as shown by co-localization with the validated ER marker TgSPT1 20,27 (Figure 3D).ER localization is canonical for CerS and provided further evidence that despite the often unusual nature of the enzymes, 20 the geometry of the sphingolipid biosynthetic pathway is conserved with respect to that seen in other eukaryotes.

TgCerS1 and TgCerS2 in parasite fitness
CRISPR/Cas9-directed knockout allowed isolation of viable, but disabled, clones lacking the TgCerS1 locus (Figure S4).However, equivalent TgCerS2 knockout clones could not be isolated.Therefore, we developed an innovative version of the conditional diCre-System 52 in which the successful excision of loxP-flanked DNA sequences is shown by a red to yellow fluorescence shift in the engineered Toxoplasma (Figure 4A,B; and by PCR in Figure S2).Using this system, we demonstrated that catalytically inactive TgCerS2 played a greater role in parasite proliferation and, subsequently, host cell lysis than the active orthologue TgCerS1 (Figure 4C).More precisely, whereas the induced TgCerS1 (RH.diCre:CerS1) knockout parasites showed only a slight, statistically non-significant reduction in plaque formation, this was greatly, and statistically significantly, reduced in the induced TgCerS2 knockout, from 3.81 ± 0.59 cm 2 to 0.93 ± 0.49 cm 2 .No significant change was noted with the parental line on rapamycin induction or with uninduced RH.diCre:CerS1/2 (Figure 4C,D and S5).It should be noted that the monolayer disruption assay (Figure S4) broadly replicated these data, but also suggested a mild fitness phenotype for TgCerS1 (RH.diCre:CerS1 and stable KO ΔCerS1).
Complementary to these analyses the parasite load and replication was assessed directly by counting the parasites per vacuole and the number of vacuoles in the infected cell population.Matching the plaque assay outcome, the TgCerS2 knockout (RH.diCre:CerS2) exhibited significantly fewer parasites per vacuole compared with parental line, with the majority of the PVs found contain 4 to 8 Toxoplasma after 12 days, whereas >40% of vacuoles in the parental line contained ≥32 parasites (Figure 4E).Furthermore, the number of vacuoles in the infected cell population was also significantly lower in induced RH.di-Cre:CerS2 (Figure 4F).In contrast, in the same timeframe, the induced TgCerS1 knockout (RH.diCre:CerS1) exhibited parasite per vacuole and vacuole per population profiles similar to parental line (Figure 4E,F).Again, it should be noted that the monolayer disruption assay (Figure S4) largely replicated these data.
Collectively these data indicated a clearly reduced parasite replication rate in the case of the induced TgCerS2 deletion.This could explain the difficulty of isolating a TgCerS2 knockout line using the direct CRISPR/Cas9 approach.

| Phylogenetic analyses of the apicomplexan ceramide synthases
Further analyses revealed that TgCerS1 and TgCerS2 have orthologues in the Apicomplexa.Two in Cryptosporidium muris, Cystoisospora suis, Cyclospora cayetanensis, Hammodia hammondi, Plasmodium falciparum, and Sarcocystis neurona, but only one copy in Cytauxzoon felis, Eimeria tenella, Neospora canium, and Theileria A region of the CerS amino acid sequences encompassing the Lag1p domain, including apicomplexan, human and Arabidopsis thaliana, were aligned (Figure S6) and subjected to phylogenetic analyses, including neighbor joining (Figure 5A), maximum parsimony and minimum evolution (Figure S7).In all cases, the sequences separated into two distinct and distant clades.The first is F I G U R E 5 Phylogenetic analyses of the apicomplexan ceramide synthases.(A) Neighbor Joining (N-J) tree of apicomplexan, human and Arabidopsis thaliana ceramide synthases analogs.Bootstrap values are based on 1000 replicates.Toxoplasma TgCerS1 and TgCerS2 are highlighted in red boxes.The canonical and essential ceramide synthase di-histidine residues (HH) are highlighted in the CerS1 group, while proline-cysteine (highlighted -PC) residues are commonly found in the same position in the CerS2 cluster, with the exception of glutamine-histidine (highlighted-QH) in Plasmodium and leucine-isoleucine (highlighted-LI) in Cryptosporidium muris.Gene duplication assessment based on N-J tree analysis and represented by rhombuses.The distinct clades containing TgCerS1 and TgCerS2 are highlighted with green and purple respectively.(B) MAFFT Lag1 motif alignment of TgCerS1, TgCerS2 and orthologues from the Apicomplexa, humans and plants.Consensus is highlighted above.Hydrophobic amino acids are depicted in green, aromatic amino acids in cyan, aliphatic amino acids in red and orange, and large polar acids in purple and blue. 42haracterized by the conserved histidine residues (H217 H218 in TgCerS1) and represents majority of the ceramide synthases, including the subgroups of HsC-erS1-6, AtCerS1-3 and the apicomplexan orthologues of the functional ceramide synthase, TgCerS1.The second clade contains the "pseudo" ceramide synthase, TgCerS2, and its apicomplexan orthologues.All of these possess only one (Q209 H210, e.g., Plasmodium spp.) or no histidines (P178 C179, e.g.TgCerS2) in place of the pair found in the other clade (Figure 5B).Furthermore, the R to L and Y to D changes found in the Lag1 domain of TgC-erS2 (Figure 2; at positions 169 and 220, respectively) are conserved in all members of this clade.Coupled with the evidence presented above showing that TgCerS2 is not catalytical active in vitro, although still ER localized, this strongly suggested that the genes encoding this apicomplexan unique "pseudo" ceramide synthase diverged early in evolution, possibly as a result of gene duplication, where one copy of the gene maintained the original function and the other diverged to a new role.The majority of species that possess a single CerS maintain an orthologue of the functional ceramide synthase TgCerS1.If an early gene duplication event led to the CerS2 orthologues, this indicated that the second copy was lost during evolution.However, surprisingly E. tenella, a coccidian apicomplexan like Toxoplasma, only maintains an orthologue of the "pseudo" ceramide synthase, TgCerS2, indicating that this parasite is unable to synthesize ceramide de novo. 4 | DISCUSSION Due to the essentiality of sphingolipids, 22 ceramide synthases (CerS) have been investigated as potential targets for therapeutic intervention in human disorders 53,54 and microbial infections. 55However, functional analyses to define druggability have been lacking in the Apicomplexa.To address this deficit in this study, we isolated, analyzed, and functionally defined ceramide synthase (CerS) in the model apicomplexan Toxoplasma gondii.
Two CerS orthologues, which we named TgCerS1 and TgCerS2, were identified in the Toxoplasma genome using F I G U R E 5 (Continued)  bioinformatic approaches.genome-wide, analyses had identified of these gene products as constitutively expressed throughout the lifecycle 49 and of some importance for parasite fitness. 48Here, we define the function, localization and evolutionary origin of these unusual enzymes in Toxoplasma.TgCerS1 was found to be a fully functional CerS, with preferential use of C16:0 acyl-CoA consistent with previous lipidomic analyses. 3,8Surprisingly, TgCerS2 was catalytically inactive as a CerS in vitro (Figure 3) and contained non-conservative substitutions in the canonical residues defining the putative active site Lag1 motif.Utilizing the recently released AlphaFold2/ RosettaFold artificial intelligence programs 43,56 we established that the structure of both orthologues was canonical, seven transmembrane domains with the substituted residues found in same arrangement in both TgCerS1 and TgCerS2 (Figures 2 and S3).With both programs, multiple predicted structures produced the transmembrane regions consistently and with high confidence levels.In both enzymes, the Lag1 motif is situated in the center of the trans-membrane region forming a hydrophobic pocket perfectly suited to accommodate the lipid substrates, with the putative active histidines located close together in Tg-CerS1.Importantly, the changes in the TgCerS sequence, active TgCerS1 to inactive TgCerS2, lead to no significant structural change in the putative active site indicating that TgCerS2 may still bind the same substrates when the catalytic activity as a ceramide synthase is lost.However, the soluble loop regions are predicted with low confidence and show limited consistent three-dimensional structure.The "Hox-like" domain which was predicted in the human CerSs 57,58 following the first transmembrane helix is not present in either TgCerS1 or 2. This is not surprising as the enzyme is highly unlikely to have any DNA-binding functions.Notably, there are large differences between the AlphaFold2 and RosettaFold predictions of the TgCerS1 and TgCerS2 cytosolic face structures, in explanation the C-terminus could represent different conformationssuch as the opening and closing of loops over the active side.The models for the luminal side, the N-terminus, show even lower confidence levels and are not even consistent with respect to secondary structure.Only experimental structure determination by X-ray crystallography or cryo-Electron Microscopy will be able to resolve the discrepancies between the models and allow the further elucidation of function.
Using an immunofluorescent approach both proteins were identified as localized in the Toxoplasma ER, the canonical localization of CerS (Figure 3).Together, these data indicated that TgCerS2, while non-catalytic, had maintained structure, localization, and, presumably, functionality through evolution.Using an inducible knockout approach, in agreement with data from the genome-wide screen (Table 1), 48 deletion of TgCerS1 exhibited only a modest fitness phenotype in an in vitro tachyzoite infection assay (Figures 4 and S4).However, in contrast, loss of the non-catalytically active TgCerS2 demonstrated a statistically significant effect (Figures 4 and S4).This deviated from the data generated during the genome-wide screen, 48 which suggested that loss of this protein had a minimal effect (Table 1).However, in agreement with the inducible KO phenotype, and unlike with TgCerS1, in an orthologous approach conventional CRISPR/Cas9 KO of TgCerS2 was unachievable in our study.Similarly, in functional studies of TgSPT1 and 2, 27 the former exhibited a major defect in an equivalent in vitro assay despite the genomic screen data suggesting this protein was non-essential for fitness (Fitness Score −0.62 48 ).The reasons for these deviations are unclear; however both this study, and the previous TgSPT1 analyses, 27 indicate that re-examination of gene and protein essentiality in Toxoplasma is warranted.
From where did this "pseudo" enzyme arise?Further bioinformatic analyses demonstrated that orthologues of TgCerS1 could be identified in almost all members of the Apicomplexa, and TgCerS2 orthologues were found in a large subset.Phylogenetic analyses (Figures 5A and S7) clearly demonstrated that the predicted ceramide synthases divide into two clades, one containing the TgCerS1 orthologues and including the Arabidopsis and human enzymes, the other clustered the TgCerS2 "pseudo" enzyme orthologues from the Apicomplexa.The fact that all the Tg-CerS2 orthologues maintain the R-L and Y-D changes (positions 169 and 220 in TgCerS2) and the loss of canonical HH (positions 217 and 218 in TgCerS2), to PC in all aside from Plasmodium spp.(QH) and C. muris (LI), supported early divergence and the maintenance of key functionality (Figure 5).However, despite this putative selective pressure, Cytauxzoon felis, Theileria annulata, and Neospora canium encode only a TgCerS1 orthologue (Figure 5B).While C. felis and T. annulata are of the same family, Theileriidae, N. canium is in the Sarocystidae.Sarcocystis neurona, Cystoisospora suis and Hammodia hammondi are also in this latter family and maintain orthologues of both TgCerS1 and TgCerS2 (T.gondii is also in the Sarocystidae).This suggested that the loss of the TgCerS2 orthologue occurred later in evolution due to unknown environmental stresses and evolutionary pressures.
In contrast, and interestingly, E. tenella maintain a TgCerS2 orthologue but appears to have lost the active ceramide synthase which indicated that this parasite lacks the ability to synthesize ceramide de novo (Figure 5).This further supported the importance of the non-catalytic ceramide synthase, a hypothesis reinforced by the same gene arrangement being found other Eimeria species (see Materials and Methods).However, uniquely in the Eimeria species, E. falciformisi maintains the functional and "pseudo" CerS2 orthologue arrangement (EfaB_MINUS_15758.g1433 EfaB_PLUS_2387.g283,respectively).Interestingly, E. falciformisi is the only Eimeria species in which the sphingolipid content has been extensively analyzed, with a lipidomic approach indicating that de novo ceramide biosynthesis occurs. 9Like E. falciformisi, another member of the family Eimeridae, Cyclospora cayetanensis, also encoded ortholgues for both proteins.Together, these data indicated that loss of orthologues of either protein (TgCerS1 or TgCerS2) is a relatively recent event in evolution and that at least one must be maintained.Apart from in most Eimeria species, the maintained orthologue appears to be a catalytical active ceramide synthase.
However, the maintenance of apicomplexan TgCerS2 orthologues, either alongside a functional enzyme or, in the Eimeria species, alone is mysterious given its apparent lack of catalytic activity.Hypothetically, following gene duplication, one copy could become non-functional due to degenerative mutations.However, alongside the situation in Eimeria species and the key role in Toxoplasma fitness reported here (Figure 4), the conservation of amino acids in the Lag1 motif and of the tertiary structure (Figure 2) in these orthologues make this unlikely.Alternatively, a duplicated copy could acquire a new function that is favorable for the organism.The maintenance of TgCerS2 orthologues, and its seemingly non-catalytic nature, across the Apicomplexa support this hypothesis.However, the true function of the TgCerS2 orthologues remains unknown.Given the proposed dimerization of human CerS2 and CerS5 to modulate activity 59 we investigated the possibility of TgCerS1 and TgCerS2 dimerization using the predicted structures (Figure 2) and those similarly generated for human CerS2 and CerS5.However, no evidence for the formation of dimers was provided for either pair (data not shown), although apparent TgCerS1 and TgCerS2 homodimers were seen by Western blotting (Figure 3A).In addition, the conservation of TgCerS2 alone in the Eimeria spp.renders this explanation unlikely.An alternative hypothesis would be that the non-catalytic isoforms act as ceramide binders and regulate the availability of this highly bioactive lipid.This is a particularly attractive explanation for the Eimeria species which presumably would have to scavenge and regulate host ceramide.However, while functional analyses of these essential proteins in an obligate intracellular pathogen of a host encoding the same activity are highly challenging, future detailed analyses to fully understand the roles of the apicomplexan "pseudo" enzyme could make use of the stable TgCerS1 knockout reported here (KO ΔCerS1; Figure S4).This mutant has a limited fitness deficit (Figure S4), although was recently shown by us to have a distinctive morphological phenotype. 60Hypothetically, this transgenic line is devoid of de novo ceramide synthase activity and therefore reliant of the acquisition of host lipid.In such a background, the host requirement for host ceramide synthesis could be probed using either broad spectrum or isoform specific host ceramide synthase inhibitors 54,61,62 and low serum media. 8owever, the identification and analyses described here detail a system unique to the Apicomplexa, a phylum of parasitic protozoa.Given the importance to parasite proliferation, coupled with the non-mammalian nature of ceramide synthesis and regulation, this could represent a potential point for therapeutic intervention for a range of diseases from malaria to toxoplasmosis.

F I G U R E 1
Substrates and products of ceramide synthase.

F I G U R E 2 T A B L E 1
Abbreviation: NA, not assigned.

F I G U R E 3
Cell-free expression and biochemical analyses of TgCer1 and TgCer2.(A) Western blot of expressed FLAG-tagged TgCer1 and TgCer2and V5-tagged HsCerS5 and HsCerS2 probed with the primary antibodies indicated.(B) TLC separation of ceramide synthase reaction products in solvent phase CHCl 3 :MeOH:H 2 O (80:12:1 v:v:v), HsCerS5 and HsCerS2 are controls and provide markers for sphinganine (NBD-Sph) and ceramide (C 16 -NBD-Cer and C 24 -NBD-Cer, respectively).Acyl-CoA substrate for each reaction shown along top of the plate.The formation of C 16 -NBD-Cer by CerS5 and TgCerS1 is only partially dependent on 16:0 CoA addition, presumably because the reaction mixture contains some wheat germ-derived CoA.(C) Longer exposure of B, red arrow indicates C 24 -NBD-Cer produced through action of HsCerS2.The relative low activity of HsCerS2 is consistent with our laboratory experience and maybe due to a lack of phosphorylation 51 or low in vitro bio-availability of the long chain CoA.The acyl-CoA substrate for each reaction shown along bottom of the plate.(D) Airyscan microscopy showing co-localization of ER marker TgSPT1 (in green; stained with polyclonal rat anti-TgSPT1)20 with TgCerS1_FLAG and TgCerS2_FLAG (in red; stained with monoclonal mouse anti-FLAG).Blue is Hoechst-stained DNA.Yellow is overlap of green (TgSPT1) and red (TgCerS1/2).

F I G U R E 4
TgCerS2 plays a critical role in parasite fitness in vitro.(A) Schematic illustration of the diCre mode of action and formation of TgCerS1 and TgCerS2 conditional knockouts (KOs).(B) Fluorescent identification of successful TgCerS1 and TgCerS2 deletion after 100 nM rapamycin induction.Confocal images were captured 72 h post-infection, scale bars 5 μM.(C) Plaque assay of Toxoplasma (100 parasites per well in 24-well plate) parental and TgCerS1 (RH.diCre:CerS1) and TgCerS2 (RH.diCre:CerS2) knockouts, uninduced or rapamycin induced (4 h), in HHF cells 12 days post-infection.Plaques are distinguishable as clear areas on the background of a Crystal Violet stained HHF monolayer, representative images shown.(D) Area of plaque sizes in cm 2 .(E) Comparison of the intracellular replication rates of the parental and the RH.diCre:CerS1 and RH.diCre:CerS2 KOs 12 days post-infection as determined by the distribution of the number (2, 4, 8,16, ≥32) of Toxoplasma parasites per parasitophorous vacuole (PV); (F) and by the number of PVs in the infected cell population (microscopic field of view).At least 150 vacuoles were analyzed for each line.PVs and intracellular parasites were analyzed from 3 slides (1.54 cm 2 ) for each line, in each experiment.Values are expressed as mean ± SD of three independent experiments.P-value significance thresholds were set at: *p < .05,***p < .01,***p < .001.