A dispensable role of mitochondrial fission protein 1 (Fis1) in the erythrocytic development of Plasmodium falciparum

Malaria remains a huge global health burden and control of this disease has run into a severe bottleneck. To defeat malaria and reach the goal of eradication, a deep understanding of parasite biology is urgently needed. The mitochondrion of the malaria parasite is essential throughout the parasite’s lifecycle and has been validated as a clinical drug target. In the asexual development of Plasmodium spp., the single mitochondrion grows from a small tubular structure to a complex branched network. At the end of schizogony when 8-32 merozoites are produced, the branched mitochondrion is precisely divided, distributing one mitochondrion to each forming daughter merozoite. In mosquito and liver stages, the giant mitochondrial network is split into thousands of pieces then daughter mitochondria are segregated into individual progeny. Despite the significance of mitochondrial fission in Plasmodium, the underlying mechanism is largely unknown. Studies of mitochondrial fission in model eukaryotes have revealed that several mitochondrial fission adaptor proteins are involved in recruiting dynamin GTPases to physically split mitochondrial membranes. Apicomplexan parasites, however, share no identifiable homologs of mitochondrial fission adaptor proteins of yeast or human, except for Fis1. Here, we investigated the localization and essentiality of the Fis1 homolog in Plasmodium falciparum, PfFis1 (PF3D7_1325600), during the asexual lifecycle. We found that PfFis1 requires an intact C-terminus for mitochondrial localization but is not essential for parasite development or mitochondrial fission. The dispensable role of PfFis1 indicates Plasmodium contains additional fission adaptor proteins on the mitochondrial outer membrane that could be essential for mitochondrial fission. Importance Malaria is responsible for over 230 million clinical cases and ∼ half a million deaths each year. The single mitochondrion of the malaria parasite functions as a metabolic hub throughout the parasite’s developmental cycle as well as a source of ATP in certain stages. To pass on its essential functions, the parasite’s mitochondrion needs to be properly divided and segregated into all progeny during cell division via a process named mitochondrial fission. Due to the divergent nature of Plasmodium spp., molecular players involved in mitochondrial fission and their mechanisms of action remain largely unknown. We found that Fis1, the only identifiable mitochondrial fission adaptor protein evolutionarily conserved in the phylum of Apicomplexa, however, is not essential for Plasmodium falciparum. Our data suggest that malaria parasites use redundant fission adaptor proteins on the mitochondrial outer membrane to mediate the fission process.

The single mitochondrion of the malaria parasite functions as a metabolic hub throughout the 52 parasite's developmental cycle as well as a source of ATP in certain stages. To pass on its 53 essential functions, the parasite's mitochondrion needs to be properly divided and segregated 54 into all progeny during cell division via a process named mitochondrial fission. Due to the 55 divergent nature of Plasmodium spp., molecular players involved in mitochondrial fission and 56 their mechanisms of action remain largely unknown. We found that Fis1, the only identifiable 57 mitochondrial fission adaptor protein evolutionarily conserved in the phylum of Apicomplexa, 58 however, is not essential for Plasmodium falciparum. Our data suggest that malaria parasites 59 use redundant fission adaptor proteins on the mitochondrial outer membrane to mediate the 60 fission process. 61 62 Malaria is a major cause of human morbidity and mortality globally (1). The clinical symptoms of 63 malaria are mainly resulted from repetitive growth of Plasmodium parasites inside the red blood 64 cells (RBCs) and continuous rupture of infected RBCs. Post invasion of a host cell, the parasite 65 lives inside the parasitophorous vacuole, undergoing growth and division to generate new 66 infective progeny. Unlike many cells that divide via binary fission, inside RBC, the malaria 67 parasite replicates its nuclear DNA 3-5 rounds without concurrent cytokinesis, resulting in the 68 formation of 8-32 daughter cells (merozoites). Division of 8-32 merozoites starts at the very end 69 of the lifecycle. This unique reproduction manner of Plasmodium during the asexual blood stage 70 is termed schizogony (2). In mosquito and liver stages, the parasite's nuclear DNA is replicated 71 13-14 rounds without cytokinesis, producing up to 10,000 progenies all at once (3); this process 72 is termed sporogony. To coordinate this unique cellular reproduction mechanism (schizogony or 73 sporogony), the single mitochondrion of the malaria parasite also undergoes a process of 74 growth and division. In the asexual blood stage of Plasmodium falciparum, the mitochondrion 75 grows from a single small tubular structure to a large branched network over the 48 h 76 intraerythrocytic development cycle (IDC) (4), which is then divided into 8-32 pieces to provide 77 one merozoite with one daughter mitochondrion. Since the mitochondrion is essential for 78 parasite growth and replication and cannot be made de novo, the process of mitochondrial 79 fission is critical to the malaria parasite. 80 Studies on mitochondrial fission in model organisms have suggested that, the fission machinery 81 that "splits" mitochondrial membranes involves at least two classes of molecules, mitochondrial 82 fission adaptor proteins located on the mitochondrial outer membrane (MOM) and fission 83 GTPases (mechanoenzymes) that are recruited from cytosol to the mitochondrial fission sites 84 (5). In mammalian cells, multiple mitochondrial fission adaptor proteins have been identified to 85 recruit the fission GTPase (dynamin related protein 1, Drp1), including Fis1 (mitochondrial 86 fission protein 1), Mff (mitochondrial fission factor), MiD49 and MiD51 (mitochondrial dynamics 87 proteins 49 kDa and 51 kDa) (6). Interestingly, except for Fis1, other mammalian MOM-bound 88 fission adaptor proteins (Mff, MiD49 and MiD51) have not been identified in yeast 89 (Saccharomyces cerevisiae) (7), plants (7), or unicellular protozoans, suggesting that 90 mitochondrial fission adaptor proteins are largely species specific. In budding yeast, Fis1 does 91 not directly bind to the dynamin GTPase (Dmn1) at the fission sites, but needs a cytosolic 92 adaptor protein Mdv1 (or its paralogs Caf4) (8). In Apicomplexan parasites, Fis1 is the only 93 identifiable mitochondrial fission protein via bioinformatics (9); BLAST searches of the malaria 94 parasite genome database (www.PlasmoDB.org) did not find any homologs of other 95 mitochondrial fission adaptor proteins in human or yeast (Mff, MiD49, MiD51 or Mdv1/Caf4). 96 Overall, Fis1 seems to the only evolutionarily conserved mitochondrial fission adaptor protein 97 throughout eukaryotic kingdoms. 98 Most Fis1 homologs are small single transmembrane proteins with ~ 150 amino acids: the C-99 terminus contains a transmembrane domain for anchoring onto the MOM and a short post 100 transmembrane tail, whereas the N-terminus has two tetratricopeptide repeats (TPR1/TPR2) 101 that facilitate protein-protein interactions (10). The hypothetical Fis1 in P. falciparum 102 (Pf3D7_1325600, 141 aa) shares 30% sequence identity to human Fis1. A modeled 3D 103 structure of PfFis1 with the software I-Tasser (11) is also superimposable with the human Fis1 104 crystal structure (PDB, 1PC2) (data not shown). In asexual blood stages, PfFis1 is transcribed 105 at all stages with a peak transcription in the late trophozoite and early schizont stages (12), 106 consistent with its expected role in mitochondrial fission. However, it has remained unknown if 107 Fis1 is essential for mitochondrial fission in malaria parasites. The Fis1 homolog in the rodent 108 malaria parasite Plasmodium berghei was not included in the large scale gene knockout (KO) 109 study (https://plasmogem.sanger.ac.uk/) (13). The recent PiggyBac mutagenesis survey of P. 110 falciparum could not unequivocally assign a phenotype to PfFis1 with statistical confidence due 111 to the short length of the gene (< 500 bp) (14). 112 In this study, we generated a conditional PfFis1 knockdown (KD) line and a PfFis1 KO line via 113 CRISPR/Cas9 mediated genome editing. In both KD and KO lines of PfFis1, parasites grew 114 normally without noticeable defects, indicating that PfFis1is not essential for mitochondrial 115 fission. We also discovered the important role of the short C-terminal tail of PfFis1 in its correct 116 subcellular localization. 117 118 PfFis1 is localized to the parasite mitochondrion but a conditional knockdown of PfFis1 119 does not cause defects in the parasite 120 In order to detect the localization of PfFis1 in P. falciparum, we episomally expressed tagged 121 PfFis1 fusion proteins with small epitopes in D10 wildtype (WT) parasites. In one transgenic 122 line, we tagged PfFis1with 3HA at the N-terminus ( Fig. 1A) whereas in another transgenic line, 123 PfFis1 was tagged with 3Myc at the C-terminus (Fig. 1B). In both parasites lines, episomal 124 expression of PfFis1 was driven by the promoter of a mitochondrial gene, the 5'-UTR of the 125 annotated mitochondrial ribosomal protein L2 (PfmtRPL2, PF3D7_1132700) (15). Similar to 126 PfFis1, PfmtRPL2 also exhibits a peak transcription at the late trophozoite and schizont stages 127 (www.PlasmoDB.org). Expression of tagged PfFis1 at the predicted molecular weight was 128 confirmed by western blotting (Fig. 1A and 1B). To verify the subcellular localization of PfFis1, 129 we performed immunofluorescence assays (IFA) in both transgenic parasites. Tagged with 3HA 130 at the N-terminus, PfFis1 was localized as expectedly to the mitochondrion; however, PfFis1 131 was diffused to the cytosol when tagged with 3Myc at the C-terminus ( Fig. 1A and 1B). It has 132 been shown recently that truncation of the entire transmembrane domain of Fis1in another 133 apicomplexan parasite, Toxoplasma gondii, resulted in mislocalization of the protein in 134 cytoplasm (16), indicating that Fis1 is anchored to MOM via the transmembrane domain. Here, 135 we kept the PfFis1 transmembrane domain intact but merely added 3Myc C-terminally, also 136 resulting in mislocalization of PfFis1. Our data highlights the critical role of the short C-terminal 137 tail (KSFKYF) in protein trafficking and localization of PfFis1 onto the MOM. 138 To determine the role of PfFis1 in parasite survival, we first utilized the CRISPR/Cas9 mediated 139 (17) TetR-DOZI-aptamer system (18) to conditionally knock down the endogenous expression 140 of PfFis1. The conditional KD system is beneficial to evaluate the essentiality of malarial genes 141 as the parasite maintains a haploid genome in the blood stages where knockouts of essential 142 genes are unachievable. In addition, since our data revealed the importance of C-terminus of 143 PfFis1 in its correct localization, we modified our conventional KD systems (19,20) to reduce 144 the expression of PfFis1, but no tags were added to the C-terminus of PfFis1 (Fig. 1C). We co-145 transfected WT P. falciparum (D10 strain) with the template plasmid, pMG75noP-Fis1-8apt, and 146 two gRNA constructs that were expected to guide Cas9 cleavage near the end of PfFis1 locus. 147 The transfected parasites were selected using media containing blastidicin (Bsd) and 148 anhydrotetracycline (aTc), yielding a conditional KD line. The small molecule aTc maintains the 149 expression level of the targeted gene by preventing the negative regulator, TetR-DOZI fusion 150 protein, from binding to the targeted mRNA which has been tagged by RNA aptamers (18). 151 Hence, expression of PfFis1 was expected to be maintained in aTc supplemented media but 152 abrogated upon aTc removal. The genotype of the PfFis1 KD line was confirmed by diagnostic 153 PCRs using site specific primers (Fig.1D). Since no antibodies were available to detect PfFis1 154 protein, we verified the KD efficiency by qRT-PCR to quantify PfFis1mRNA transcripts in 155 parasites upon aTc removal for 2, 4, 6 and 8 days (up to 4 IDCs). In comparison to aTc plus 156 controls, the PfFis1 transcript dramatically decreased after aTc removal for 1 IDC (2 days) and 157 continued to decrease to a negligible level over the KD time course (Fig. 1E). With or without 158 aTc, however, the parasites did not show any defects in the growth rate (Fig. 1F), indicating the 159 dispensable role of PfFis1 in parasite survival. Interestingly, our data also revealed that in the 160 absence of aTc, the TetR-DOZI-aptamer system not only prevents protein translation by pulling 161 the mRNA away from ribosomes (18), but also cause a rapid degradation of the target mRNA. 162 To our best knowledge, this is the first report to show that the TetR-DOZI-aptamer system could 163 also interfere with mRNA stability of the target gene. 164 165

PfFis1 is dispensable for mitochondrial fission 166
Our data thus far has shown that PfFis1 is likely non-essential for the parasite. To rule out the 167 possibility that a low amount of PfFis1 protein in the KD parasite was sufficient to maintain 168 parasite health, we attempted to knock out PfFis1 via CRISPR/Cas9 genome editing. In D10 169 WT, we co-transfected the KO template plasmid carrying two homologous sequences of PfFis1 170 with two gRNA constructs that guide Cas9 cleavage in the middle of the PfFis1 genetic locus 171 ( Fig. 2A). The transfected parasites were selected by WR99210, a specific inhibitor of the 172 Plasmodium dihydrofolate reductase gene (21) and drug resistant parasites were achieved. The 173 genotype of the transgenic parasite was confirmed by diagnostic PCRs using site specific 174 primers (Fig. 2B). PfFis1 was successfully knocked out, yielding a PfFis1 KO line (PfFis1KO). 175 We tightly synchronized both PfFis1KO and WT lines and observed the growth kinetics over 6 176 IDCs through parasitemia counting in blood smears of the cultures. PfFis1KO parasites did not 177 exhibit any noticeable growth defects when compared to WT (Fig. 2C), nor did they appear 178 morphologically abnormal (Fig. 2D), suggesting that a complete removal of PfFis1 did not cause 179 problems for parasite growth and replication. To monitor mitochondrial morphologies of 180 PfFis1KO parasites, we stained them with a fluorescent dye Mito-Tracker and performed live 181 cell microscopy. In comparison to WT controls, the mitochondrion of PfFis1KO displayed normal 182 development during the asexual blood stage; importantly, it underwent fission in the late 183 schizont stage to produce fragmented mitochondria to be distributed into daughter cells (Fig.  184   2E). In addition, we mixed equal numbers of PfFis1KO and WT parasites in one flask, obtained 185 DNA samples every 2 IDCs (4 days) and detected the presence of PfFis1KO parasites by PCR 186 (Fig. 2F). Only PfFis1KO contained the exogenous hDHFR gene (human dihydrofolate 187 reductase gene, the transfection selectable marker) and robustness of the PCR band persisted 188 throughout 16 IDCs (32 days), suggesting that there was no or negligible fitness cost associated 189 with a complete deletion of PfFis1. Collectively, our data suggest that PfFis1 is not essential for 190 mitochondrial fission in the asexual blood stage of P. falciparum. 191 Conclusions. Our data support that PfFis1 relies on its short C-terminal tail for mitochondrial 192 localization, but is not essential for mitochondrial fission or parasite survival in P. falciparum. 193 Our Fis1 KO data in malaria parasites is consistent with the proposed non-essentiality of Fis1 by 194 KD approaches carried out in Toxoplasma gondii (22). Discovered first in yeast in 2000 (23), 195 Fis1 has been evolutionary conserved in most eukaryotes that contain mitochondria. In yeast, 196 adapter proteins in different genus of the Apicomplexa phylum. In particular, the novel and 205 essential mitochondrial fission adaptor protein of apicomplexan parasites could represent 206 potential anti-parasitic drug targets as they are likely absent in the host mitochondria. 207 208

Parasite culture and transfection 210
Plasmodium falciparum D10 WT was used for all transfections in this study. Parasite culture and 211 transfection procedures followed our previously published protocols (19,20). 212

Plasmid construction 213
To tag PfFis1 with 3Myc at the C-terminus, the plasmid pLN-hDHFR-PfFis1-3Myc was 214 constructed by amplifying the PfFis1 gene from P. falciparum genomic DNA using primers 215  For constructing the template plasmid for KD studies without tags, the 5'HR (5' homologous 242 region) of PfFis1 was amplified from genomic DNA by primers P13+P14. The 3'UTR (3' 243 homologous region) of PfFis1 was amplified from genomic DNA by primers P15+P16. The 244 3'UTR was cloned into the pMG75noP-8apt-3HA vector (20) by AflII and BspeI, whereas the 245 5'HR fragment was subsequently cloned by BspeI and ApaI, yielding the plasmid pMG75noP-246 Fis1-8apt for KD. This plasmid was linearized with EcoRV before transfection. For constructing 247 the template plasmid for KO studies, the 5'HR was amplified from genomic DNA by primers 248 P17+P18. The 3'HR was amplified from genomic DNA by primers P19+P20. The two HR 249 fragments were sequentially cloned into pCC1, yielding the plasmid pCC1-5'3'Fis1 for KO. This 250 plasmid was linearized with HincII before transfection. Primers of P21-P24 were used to check 251 the genotype of the KD line; primers of P25-P29 were used to check the KO genotype. 252

Nucleic acid extraction, PCR and qRT-PCR 255
Genomic DNA from late-stage parasites was isolated with DNeasy Blood and Tissue kit 256 (Qiagen). During the KD time course (days 2, 4, 6 and 8), total RNA from parasites of each 257 condition (aTc plus vs aTc minus) was isolated from saponin lysed parasite pellets followed by 258 treatment with Trizol (Thermo) and purification with RNeasy kit (Qiagen). After treated with 259 DNase I (New England Biolabs), 2 µg RNA of each condition was primed with random 260 hexamers and converted to cDNA using SuperScript III reverse transcriptase (Thermo). qRT-261 PCR was carried out in triplicate with SYBR Green Real-Time PCR Master Mixes (Thermo) in 262 the Real-time PCR instrument (Applied Biosystems). Primers used for amplification of PfFis1 263 are listed as P30-P31. A previously reported house-keeping gene, seryl-tRNA synthetase, was 264 used as the internal control (primers P32-P33) (26). Data was analyzed using 2 -ΔΔCt method as 265 previously described (27). For regular PCR, a reaction volume of 25 µL was used and extension 266 temperature of the Taq polymerase was kept at 62°C. 267

Conflict of interest 270
The authors declare no conflicts of interest with the contents of this article. to that of the aTc plus culture (the latter was normalized to 100%). Seryl-tRNA synthetase was 300 used as an internal control. Error bars indicate standard deviations from triplicate samples; this 301 experiment has been repeated two times. (F) Effect of PfFis1 KD on the growth of P. 302 falciparum. To quantify the growth, parasites were enriched by Percoll, equal number of 303 PfFis1KD parasites was grown in the presence and absence of aTc in the medium (+aTc and -304 aTc) for 12 days (6 IDCs). Parasitemia was counted on every alternate day and the 305 parasitemia was multiplied by the parasite dilution factor to produce a cumulative measure of 306 growth. All data points are mean±s.d. of three independent experiments. 307 for 30 min and washed three times. Images were acquired at ring, trophozoite and schizont 321 stages. Scale bars equal 2μm. (F) Growth competition between PfFis1KO and D10 WT 322 parasites. The house-keeping gene seryl-tRNA synthetase was used as an internal control.