A unique dynamin-related protein is essential for mitochondrial fission in Toxoplasma gondii

Abstract

The single mitochondrion of apicomplexan protozoa is thought to be critical for all stages of the life cycle, and is a validated drug target against these important human and veterinary parasites. In contrast to other eukaryotes, replication of the mitochondrion is tightly linked to the cell cycle. A key step in mitochondrial segregation is the fission event, which in many eukaryotes occurs by the action of dynamins constricting the outer membrane of the mitochondria from the cytosolic face. To date, none of the components of the apicomplexan fission machinery have been identified and validated. We identify here a highly divergent, dynamin-related protein (TgDrpC), conserved in apicomplexans as essential for mitochondrial biogenesis and potentially for fission in Toxoplasma gondii. We show that TgDrpC is found adjacent to the mitochondrion, and is localised both at its periphery and at its basal part, where fission is expected to occur. We demonstrate that depletion or dominant negative expression of TgDrpC results in interconnected mitochondria and ultimately in drastic changes in mitochondrial morphology, as well as in parasite death. Intriguingly, we find that the canonical adaptor TgFis1 is not required for mitochondrial fission. The identification of an Apicomplexa-specific enzyme required for mitochondrial biogenesis and essential for parasite growth highlights parasite adaptation. This work paves the way for future drug development targeting TgDrpC, and for the analysis of additional partners involved in this crucial step of apicomplexan multiplication.

Fig 1

Conservation of the fission machinery in T. gondii: TgFis1 is not essential for mitochondrial fission A) Illustration of the main events that lead to mitochondrial fission in higher eukaryotes. ER and actin enwrap the mitochondrial membrane at future sites of fission, and an “adaptor complex” recruits Drp1/Dnm1, which forms spirals around the MOM and constricts it. (B) Scheme of the promoter replacement strategy adopted for the conditional KD of TgFis1; red and blue arrows indicate oligonucleotides designed to verify promoter exchange via genomic PCR. (C) Genomic PCR confirms the replacement of the endogenous promoter of TgFis1 with the inducible promoter Tet07-Sag1. A clonal parasite line was used for genomic PCR; primer positions and amplicon length are specified. (D) Confirmation of protein depletion upon addition of ATc for 48h in Fis1KD. In absence of ATc, western blot analysis shows the expected protein size of 17 kDa, tagged with c-MYC; the induction of TgFis1 knockdown with ATc shows that at 48 hours the protein is no longer detectable by western blot. (E) Plaque assay on parasites grown on HFF cells for 7 days in presence and absence of ATc shows that depletion of TgFis1 is not deleterious for parasite fitness. The images shown are representative of three experiments. (F) Immunofluorescence analysis shows that TgFis1 is evenly distributed in the mitochondrial outer membrane. In presence of ATc, the signal of Fis1 is no longer detectable, but mitochondria morphology is not affected. The experiment was performed in triplicate; representative images are shown. Scale bar: 2 μm.

Fig 2

Modelling of the GTP binding domain of TgDrpC (A) Monomer model for TgDrpC GTPase domain. The P-loop responsible for tri/diphosphate binding is indicated in yellow, while the switch I region with the putative MG²⁺ coordinating threonine is in blue. (B) close-up of the GDP binding site after molecular dynamics simulation with key residue interacting to the co-substrate shown in ball-and-stick representation, yellow for residues in the P-loop, cyan for residues interacting with the ribose and the guanine.

Fig 3. Endogenous tagging of TgDrpC reveals a mitochondria localisation.

(A) Strategy for the endogenous tagging of TgDrpC with YFP in RHΔku80 parasite strain: The ligation-independent cloning vector LIC-YFP was integrated at the 3’ end of TgDrpC, through a single-crossover mechanism. (B) PCR analyses using the primers indicated in (A) confirms the integration of the tagging construct at 5’ and 3’ ends. (C) Western blot analysis using α-GFP antibody on the TgDrpC-YFP clonal line verifies the expected protein size of ≈160 kDa. (D) Immunofluorescence images showing the co-localisation between TgDrpC and the indicated organelles (Mic4, micronemes; HSP60, apicoplast; Rop2-4, rhoptries; Tom40, mitochondrion). (E) Quantification of colocalisation between TgDrpC and the indicated organelles; the Manders’ coefficient (average of n>20 values) is reported in the y axis. (F) TgDrpC puncta (green) have different shapes and sizes, varying from spirals to smaller dots on the membrane; 3-D reconstruction (G) confirms that the vast majority of TgDrpC aggregates are in close proximity with mitochondria. Scale bar: 5 μm.

Fig 4. Recruitment of TgDrpC at interconnected mitochondria during replication.

(A) Time-lapse analysis of parasites TgDrpC-YFP/TGME49_215430-tdTomato with interconnected mitochondria (in red, TGME49_215430-tdTomato; in green, TgDrpC-YFP). TgDrpC is at the mitochondria periphery and at the basal interconnection; while the puncta at the periphery are more static, the signal at the basal end is highly dynamic. Time is indicated in minutes. Scale bar: 5 μm. (B) SIM microscopy shows that the puncta at the basal end coincide with sites of constriction of the MOM (white arrows). Scale bar: 5 μm.

Fig 5. Inducible Knock-down of DrpC has a specific mitochondrial effect.

(A) Scheme of the genomic locus after integration of the DrpC-U1 construct, as detailed in Pieperhoff et al., 2015. (B) Western Blot analysis showing efficient downregulation of TgDrpC 72 hours post Rapamycin induction. The membrane was probed with α-HA and α-Catalase antibodies. (C) Immunofluorescence analysis shows that TgDrpC signal is no longer observable at 72 hours post induction; at this time, the distribution of micronemes and rhoptries, probed with Mic4 and Rop13, was not affected in Rapamycin-treated parasites; similarly, the morphology of apicoplast (αHSP60), IMC (αGAP45) and Golgi (GRASP-RFP) appeared normal, as quantified in (D). (D) Organelle phenotypes were scored in the following way: for micronemes and rhoptries, an apical distribution was counted as “normal”; for the apicoplast, both location (at the apical part of resting parasites) and number (one per parasite in a vacuole) were assessed; for the Golgi, correct distribution above the nucleus and morphology (i.e., not fragmented) were checked; the IMC stain forming the “banana shape” typical of tachyzoites was scored as normal. For each data set, 100 parasitophorous vacuoles were counted. This was performed in triplicate. In contrast, as shown by Immunofluorescence analysis in (E) and quantified in (F), induced parasites showed severe morphology defects in the mitochondrion, classified as “interconnected mitochondria” (41%), “thick mitochondrial membranes” (39.2%) and “collapsed mitochondria” (7%) Scale bar: 5μm. (G) Complementation experiments through transient transfection of plasmid p5RT70-GFP-DrpC in Rapamycin-treated parasites led to rescue of the mitochondrial phenotype, as quantified in (F). Quantification in (F) was obtained through comparison of three independent experiments, each with at least 300 vacuoles. Error bars show SD. Asterisks indicate significant difference (P<0.001 multiple t-test).

Fig 6. Over-expression of a dominant-negative form of TgDrpC leads to interconnected mitochondria.

A) Scheme of the plasmid DD-GFP-DrpCDN. (B) Western blot analysis with the indicated antibodies shows that Shield-1 induction efficiently regulates the expression of DD-GFP-DrpCDN as soon as 4 hours post induction. (C) Plaque assay shows that induction of DD-GFP-DrpCDN causes parasite death. Parasites were grown for 7 days on HFF cells in presence and absence of 1μM Shield-1. The experiment was performed in triplicate; representative images are shown. (D) The effects of DD-GFP-DrpCDN expression were assessed 24 hours after Shield-1 induction; immunofluorescence analysis of the specific markers Mic4, Rop13, HSP60 and Gap45 shows no effect on secretory organelles, apicoplast or IMC morphology. (E and F) Eight hours after stabilisation, DD-GFP-DrpCDN shows an accumulation at the basal part of the mitochondria, which increasingly appear interconnected (23% of the total); 24 hours after induction, more than half of the parasites show abnormally interconnected mitochondria, and DD-GFP-DrpCDN accumulates at the interconnections (scale bar: 5 μm). At least 100 parasitophorous vacuoles were counted in triplicate. Error bars represent SD from the three independent experiments. Asterisks indicate significant difference (P<0.001 multiple t-test).

Fig 7. Truncated TgDrpC does not localise to the mitochondrion.

(A) and (B) Schematics of the plasmids DD-GFP-DrpCtruncated (amino-acids 407–1174) and DD-GFP-DrpCGTPase only (amino-acids 1–406). (C) Western blot analysis shows efficient DD-GFP-DrpCtruncated stabilisation 24 hours after Shield-1 induction. (D) Plaque assay analysis shows that overexpression of DD-GFP-DrpCtruncated does not impair parasite fitness. Parasites were grown for 7 days on HFF cells in presence and absence of 1μM Shield-1. (E) Immunofluorescence analysis shows that DD-GFP-DrpCtruncated localises to the cytoplasm and does not form puncta. (F) Immunoblot analysis of clonal DD-GFP-DrpCGTPase only parasites in presence and absence of Shield-1 using the indicated antibodies. (G) DD-GFP-DrpCGTPase only parasites grown in presence of 1 μM of Shield-1 for 7 days do not show any growth defect compared to non-induced controls. (H) Immunofluorescence analysis shows that DD-GFP-DrpCGTPase only localise to the cytoplasm. Scale bar: 5 μm.