F-actin and myosin F control apicoplast elongation dynamics which drive apicoplast-centrosome association in Toxoplasma gondii

ABSTRACT Toxoplasma gondii contains an essential plastid organelle called the apicoplast that is necessary for fatty acid, isoprenoid, and heme synthesis. Perturbations affecting apicoplast function or inheritance lead to parasite death. The apicoplast is a single copy organelle and, therefore, must be divided so that each daughter parasite inherits an apicoplast during cell division. In this study, we identify new roles for F-actin and an unconventional myosin motor, TgMyoF, in this process. First, loss of TgMyoF and actin lead to an accumulation of apicoplast vesicles in the cytosol indicating a role for this actomyosin system in apicoplast protein trafficking or morphological integrity of the organelle. Second, live cell imaging reveals that during division the apicoplast is highly dynamic, exhibiting branched, U-shaped and linear morphologies that are dependent on TgMyoF and actin. In parasites where movement was inhibited by the depletion of TgMyoF, the apicoplast fails to associate with the parasite centrosomes. Thus, this study provides crucial new insight into mechanisms controlling apicoplast-centrosome association, a vital step in the apicoplast division cycle, which ensures that each daughter inherits a single apicoplast. IMPORTANCE Toxoplasma gondii and most other parasites in the phylum Apicomplexa contain an apicoplast, a non-photosynthetic plastid organelle required for fatty acid, isoprenoid, iron-sulfur cluster, and heme synthesis. Perturbation of apicoplast function results in parasite death. Thus, parasite survival critically depends on two cellular processes: apicoplast division to ensure every daughter parasite inherits a single apicoplast, and trafficking of nuclear encoded proteins to the apicoplast. Despite the importance of these processes, there are significant knowledge gaps in regards to the molecular mechanisms which control these processes; this is particularly true for trafficking of nuclear-encoded apicoplast proteins. This study provides crucial new insight into the timing of apicoplast protein synthesis and trafficking to the apicoplast. In addition, this study demonstrates how apicoplast-centrosome association, a key step in the apicoplast division cycle, is controlled by the actomyosin cytoskeleton.

The apicoplast is an essential organelle required for fatty acid, isoprenoid, iron-sul fur cluster, and heme synthesis (2)(3)(4)(5)(6)(7)(8).Genetic manipulations or perturbation of the apicoplast with pharmacological compounds results in apicoplast inheritance defects and parasite death (9)(10)(11).Despite containing a small 35-kb genome, the majority of apicoplast localized proteins are encoded in the nucleus.Therefore, trafficking of these proteins to the apicoplast is a critical cellular process, and our understanding of the mechanisms of apicoplast protein trafficking is incomplete.Many apicoplast proteins contain a bipartite N-terminal signal sequence that includes a signal peptide (SP), which directs protein translation to the endoplasmic reticulum (ER), and an apicoplast transit peptide which is revealed after removal of the SP (12,13).It is unclear how these proteins translocate from the ER to the apicoplast.A subset of apicoplast proteins that localize within the apicoplast membranes, including APT1, Atrx1, and FtsH1, do not contain a signal peptide or apicoplast targeting sequence and have been observed in vesicles that are thought to directly traffic these proteins from the ER to the apicoplast in a cell cycle-dependent manner (14)(15)(16)(17)(18)(19).Proteins that localize to both the apicoplast and the mitochondria are trafficked to the apicoplast via the Golgi and trafficking of these proteins is perturbed by the addition of the ER retention motif HDEL (20).Overall, there are numerous mechanisms controlling the trafficking of proteins to the apicoplast, and the molecular details of each system are poorly understood.
Evolutionarily, the apicoplast was acquired by the parasite via a secondary endosym biosis of red algae and contains four membranes (21).The inner two membranes are derived from the progenitor chloroplast (inner most), the third from the algae plasma membrane (periplastid), and outermost membrane from endomembrane system of the Apicomplexan progenitor which first acquired this organelle (Fig. 1A).Protein translo cons Tic and Toc in the inner two apicoplast membranes facilitate protein transport from the periplastid space to the apicoplast lumen (22,23).Repurposed ER-associated degradation proteins translocate proteins across the periplastid membrane (24,25).How proteins traverse the outer membrane is not understood but, in some instances, may involve fusion of the outer membrane with ER-derived vesicles (26).
The apicoplast is a single-copy organelle; thus, apicoplast division and inheritance must occur with high accuracy to ensure parasite survival.Apicoplast division is a multistep process.During interphase, the apicoplast has a circular morphology and is located at the parasite's apical end (Fig. 1A).At the beginning of cell division, the apicoplast interacts with the duplicated centrosomes and elongates (27).This interac tion between centrosomes and the apicoplast is dependent on the autophagy-related protein Atg8 which localizes to the outer membrane of the apicoplast and disruption of Atg8 in both Toxoplasma and Plasmodium results in apicoplast inheritance defects (28)(29)(30)(31).While the apicoplast is elongated, daughter cell construction within the parasite cytosol is initiated (32).The apical tip of the growing daughters is connected to the centrosomes via a fiber containing striated fiber assemblins (SFAs) (33), ensuring that each daughter cell inherits a single centrosome.As the centrosome migrates into the growing daughter scaffolds, the elongated apicoplast then extends to a U-shaped structure, whereupon apicoplast fission is mediated by dynamin-related protein A (TgDrpA) (34).
Three cytoskeletal proteins, actin (TgAct1), myosin F (TgMyoF: an unconventional apicomplexan-specific myosin motor), and the actin nucleator formin-2 also, play major roles in apicoplast division (35)(36)(37)(38).Since TgMyoF has also been implicated in centro some positioning, it was previously proposed that TgMyoF maintains the duplicated centrosomes in close proximity, thus allowing the apicoplast to associate with both structures (37).However, it has also been shown that TgMyoF and actin control the transport of a variety of vesicles, including dense granules and Rab6 vesicles (38,39), thus an alternative hypothesis is that TgMyoF and actin are needed for the trafficking of vesicles to the apicoplast, which could provide new material (proteins and lipids) to the apicoplast prior to division.In this study, we investigate the roles of TgMyoF and actin in apicoplast division.We demonstrate that vesicles containing the apicoplast protein APT1 are formed in a cell cycle-dependent manner consistent with previous work (14).A halo pulse-chase assay indicates that APT1 is expressed in a cell cycle-dependent manner and that vesicles contain newly synthesized, ER-derived protein.Loss of TgMyoF and actin led to an increased number of vesicles in the cytosol implicating this actomyosin system in apicoplast protein trafficking.
Live cell imaging revealed that elongated apicoplasts are highly dynamic, exhibiting linear, branched, and U-shaped morphologies which are driven by both TgMyoF and actin.Loss of this actomyosin system results in reduced apicoplast length and defects in apicoplast-centrosome association.Thus, we conclude that TgMyoF and actin play a vital role in recruiting the apicoplast to the duplicated centrosomes during apicoplast division, a vital step that ensures each daughter inherits a single apicoplast.

Cytochalasin D treatment
For fixed cell imaging, parasites invaded confluent HFF monolayer for 4 h and then treated with 0.2 µM cytochalasin D (CD) or equivalent volume of DMSO for 12 h before fixation with 4% paraformaldehyde.For live cell imaging, parasites were grown for 12 h before treatment with 2 µM CD or equivalent volume of DMSO for 30 min.Imaging was carried out in the presence of DMSO or CD.Left: assay outline.Parasites expressing APT1-Halo were labeled with JF646 or Oregon Green (OG) for 15 min (pulse).After a 6-h outgrowth, the parasites were labeled with Halo-TMR for 15 min (chase) and immediately imaged live.Right: maximum intensity projections of deconvolved images of Halo-APT1 expressing parasites after Halo pulse-chase assay.Interphase: no TMR labeling was observed in interphase parasites, indicated by circular apicoplasts.Thus, APT1 protein synthesis does not occur during interphase.Early elongation: early in apicoplast elongation, the apicoplast is stained with both Halo-JR646 and TMR (green and magenta arrows).An additional ER-like staining of TMR was also observed (green arrowheads).Late elongation: the apicoplast was stained with both fluorophores and the ER-like staining of Halo-TMR is reduced (inset).Dashed

IAA treatment
To deplete TgMyoF protein levels, TgMyoF-mAID:TgAPT1-EmGFP parasites were grown in confluent HFF monolayer for 5 or 12 h and then treated with a final concentration of 500 µM indole acetic acid (IAA) or equivalent volume of 100% ethanol for 6 or 15 h before fixation for immunohistochemistry or live cell imaging (40,41).

Creation of transgenic parasite lines
A list of parasite lines, primers, and gene accession numbers used in this study can be found in Tables S1, S2 and S3, respectively.

Creation of TgAPT1-EmGFP parasite line
In the TgAPT1-EmGFP parasite line, the endogenous TgAPT1 gene is tagged with EmeraldGFP (EmGFP).The pTKOII_APT1EmGFP plasmid was created from pTKOII_MyoF-EmGFP plasmid (39) by digesting with AvrII and BglII restriction enzymes to remove the TgMyoF genomic sequence.Two kilobases of genomic DNA upstream of the TgAPT1 stop site was amplified by PCR using Q5 high-fidelity polymerase (New England BioLabs) using genomic DNA isolated from ΔKu80 parasites as a template and the primers TgAPT1 F1 and TgAPT1 R1 (Table S2).The plasmid backbone and the PCR product were gel purified.Ligation of the plasmid was carried out via Gibson assembly using NEB builder HiFi DNA assembly master mix as per manufacturer's instructions (New England Biolabs; catalog # M5520AA2) and transfected into NEB5α bacteria.Positive clones were screened by colony PCR and verified by sanger sequencing.pTKOII-TgAPT1-EmGFP plasmid was linearized for homologous recombination using the PstI restriction enzyme, and 25 µg was transfected into 1 × 10 7 ΔKu80: ΔHXGPRT parental parasites (43,44).Twenty-four hours after transfection, mycophenolic acid in ethanol and Xanthine in 0.3M NaOH were added to growth media to a final concentration of 25 µg/mL and 50 µg/mL respectively.Drug treatment continued until at least 75% of the parasites were EmGFP positive.Clonal parasites lines were FACS sorted into a single 96-well plate (2 per well) and grown for 7 days to isolate clonal parasite lines resulting in the creation of TgAPT1-EmGFP parasite line.To verify integration of this construct into the APT1 genomic locus, genomic DNA was isolated using Qiagen DNeasy blood and tissue kit as per manufacturer's instructions (Qiagen, Germantown, MD) and confirmed via PCR using the primers TgAPT1-Promoter F1, APT1 CDS R, EmGFP R (Table S2) (Fig. S1).

Creation of TgMyoF-mAID; TgAPT1-Halo parasite line
In the TgMyoF-mAID:TgAPT1-Halo parasite line, the endogenous TgMyoF gene is tagged with HA-mAID (39) and TgAPT1-Halo is expressed under the control of the APT1 promoter from the dispensable UPRT locus.For the creation of 5′UPRT-pAPT1-APT1-Halo-3′UPRT plasmid (which will be used as a template for created of the HR oligo via PCR), the 5′UPRT-pAPT1-APT1-EmGFP-3′UPRT plasmid was digested with AflII and BglII to remove the EmGFP coding sequence.The Halo coding sequence was amplified using Halo-F1 and Halo-R2 primers using pmin-ActinCB-Halo as a template (Table S2).The plasmid backbone and the PCR product were gel purified.Ligation of the plasmid was carried out via Gibson assembly as per manufacturer's instructions (New England Biolabs; catalog # M5520AA2) and transfected into NEB5alpha bacteria.Positive clones were screened by colony PCR and verified by sanger sequencing.

Creation of ectopic ATG8 expression plasmid
Generation of pAtg8-3xTy-Atg8 plasmid for ectopic expression of TgAtg8 under the endogenous promoter was a two-step process.First, we created the pAtg8-tdTomato-Atg8 plasmid by digesting the pTub_tdTomato-Atg8 (28) with SfoI and BglII restriction enzymes to remove the tubulin promoter sequence.TgAtg8 promoter was amplified from the Ku80 parental genomic DNA using the primers Atg8 promoter F and Atg8 promoter R. Plasmid backbone and the PCR product were gel purified.Ligation of the plasmid was carried out via Gibson assembly as per manufacturer's instructions (New England Biolabs; catalog # M5520AA2) and transfected into NEB5alpha bacteria.Positive clones were screened by colony PCR using Atg8 promoter F and Atg8 promoter R (Table S2) and verified by sanger sequencing.In the second step, pAtg8-tdTomato-Atg8 was digested with AvrII and BglII restriction enzymes to remove tdTomato coding sequence.The 3xTy tag was created by PCR using primers 3xTy F and 3xTy R (Table S2) and the pTKOII_MyoL-3xTy-mAID plasmid as a template (unpublished plasmid; Table S1).Ligation of the plasmid with 3xTy insert was carried out via Gibson assembly as per manufac turer's instructions (New England Biolabs; catalog # M5520AA2) and transfected into NEB5alpha bacteria.Positive clones were screened by colony PCR using primers Atg8 promoter F and 3xTy R and verified by sanger sequencing.Twenty-five micrograms of the pAtg8-3xTy-Atg8 plasmid was transiently transfected into the TgMyoF-mAID parasite line and was grown for 12 h before treating the parasites with either ethanol or IAA for 6 hr.

Fluorescence microscopy
All live and fixed images were acquired on a DeltaVision Elite microscope system built on an Olympus base with a 100 × 1.39 NA objective in an environmental chamber heated to 37°C.This system is equipped with a 15-bit scientific CMOS camera and DV Insight solid-state illumination module with the following excitation wavelengths: DAPI = 390/18 nm, FITC = 475/28 nm, TRITC = 542/27, and Alexa 647 = 632/22 nm.Single band pass emission filters had the following wavelengths: DAPI 435/48 nm, FITC = 525/48 nm, TRITC = 597/45 nm, and Alexa 647 = 679/34 nm.Image acquisition speeds for the live cell imaging were determined on a case-by-case basis and are indicated in the figure legends.All cells were imaged in multiple focal planes and Z-series were captured at 0.2 µm (fixed) or 0.4 µm (live) steps.Images presented in the figures represent maximum intensity Z-slice projections.Brightness and contrast of images were optimized for print.

Immunocytochemistry and live cell imaging
Parasites were fixed with freshly made 4% paraformaldehyde (Electron microscopy sciences, Catalog # 15714) in 1× PBS for 15 min at room temperature.Cells were washed three times in 1× PBS and permeabilized in 0.25% Triton X-100 (ThermoFisher) in 1× PBS for 15 min at room temperature.Cells were washed three times in 1× PBS and were blocked in 2% BSA in 1× PBS and incubated for 15 min.All antibodies were diluted in 2% BSA/1× PBS at the concentrations listed in Table S4.Cells were incubated sequentially with primary and secondary antibody at room temperature for 30 min each, with three washes with 1× PBS between incubations.To stain DNA parasites were incubated in 10 µM DAPI diluted in 1× PBS for 10 min and then washed three times in 1× PBS.Coverslips were mounted onto slides using either Prolong Gold (ThermoFisher Catalog # P36930) or Prolong diamond (ThermoFisher Catalog # P36970) anti-fade reagent and allowed to dry overnight before imaging.
For live cell imaging, growth media was replaced with Fluorobrite DMEM (Thermo Fisher) containing 1% FBS and 1× antimycotic/antibiotic pre-warmed to 37°C.Z-step size and imaging frame rates are indicated in the video legends.
The halo pulse chase assay was carried out as per manufacturer's instructions (Promega, Inc).Parasites were grown overnight in confluent HFF monolayers in MetTak dishes.Parasites were incubated for 15-30 min with either 200 nM Halo-ligand JF646 or 1 µM Halo-Ligand Oregon Green (OG) diluted in growth media.(Note: Oregon Green was used for our initial experiments but this fluorophore is not photostable under our imaging conditions.We switched to JF646 for subsequent experiments.)Cells were washed three times in growth media and incubated for 30 min.Cells were washed again two times in growth media and incubated for 5.5 h.Cells were then labeled in 5 µM Halo-ligand TMR for 15 min.Cells were washed three times in growth media and incubated for a further 30 min in growth media.Cells were washed twice in Fluorobrite DMEM with 1% FBS before live cell.For single time point imaging, two-color (JF646/TMR or OG/TMR) Z-stack images were taken.For time-lapse imaging, single-color TMR imaging was performed for 5 min.

Quantification of TgAPT1 vesicle number
To determine the number of APT1 vesicles, cell cycle stage was determined based on apicoplast morphology and presence and size of daughter parasites.Parasites were categorized as interphase if they did not contain daughter parasites and had a single circular apicoplast.Parasites were categorized as elongated if the parasites contained an elongated apicoplast regardless of the presence of daughter parasites since apicoplast elongation occurs before daughter cell construction (47).Parasites were categorized as post-fission if they contained two daughter parasites, each with a single apicoplast.The apicoplast was visualized using anti-Cpn60 antibodies and expression of TgAPT1-EmGFP.Daughter cells were visualized by expressing ptub-mCherry-TubulinA1 a marker for the pellicle of daughter and mother parasites.Number of vesicles were counted manually on average intensity projection images using cell counter plugin in Fiji (Fiji is Just ImageJ; http://imageJ.net).Three independent experiments were performed with 25 parasites per experiment counted for each stage of division.

Quantification of Halo-TMR localization
To quantify the localization of Halo-TMR (chase) in the Halo pulse-chase assay, we categorized parasites as interphase (if the apicoplast was circular) or elongated based on apicoplast morphology.We then quantified the percentage of parasites which had apicoplast and ER-like Halo-TMR staining or had no Halo-TMR signal.N = 65 vacuoles from three independent experiments.We never observed ER-like staining in the absence of apicoplast staining.

Characterization of TgAPT1 vesicle motion
To characterize TgAPT1 vesicle motion, vesicle tracking analysis was performed using Imaris v9.9.0 microscopy image analysis software (Bitplane AG, Oxford Instruments).Z-stack movies that were bleach corrected in ImageJ using the histogram matching algorithm were uploaded to Imaris and brightness and contrast was auto-adjusted using the "normalize time point" extension.Movies were cropped to remove areas that did not contain apicoplast and vesicles.To prevent aberrant tracking inside the apicoplast itself, the apicoplasts were demarcated from vesicles using the surface tool with a lower surface threshold of 1,500 arbitrary units (a.u.).Vesicles were identified using the spots creation wizard.Size parameter (estimated XY diameter) was 0.3 µm.To distin guish between vesicles and background fluorescence, only objects with a fluorescence intensity above 150 a.u.were tracked.Tracks were created using autoregressive motion algorithm.To be considered true vesicles, vesicles must be located within 5 µm from the apicoplast surface, to prevent the aberrant detection of fluorescent puncta outside the parasite, especially at later time points when photobleaching reduced the fluorescence intensity of the vesicles.A maximum gap of two frames was permitted (i.e., maximum number of the consecutive time points that a track segment can be missed to be considered continuous).Tracks with less than 5 data points were excluded from the analysis as it was not possible to accurately assign a straightness value.Two independent experiments with a total of 14-17 videos were used for this analysis.Straightness values were calculated by determining the ratio of track displacement (the distance between the asites were categorizedfirst and last positions of the vesicle) to track length (total distance traveled).

Measuring apicoplast to centrosome distance
Distance between the elongated apicoplast and centrosomes was determined using the straight-line tool in ImageJ by measuring the closest distance from each centrosome to the apicoplast.The distance was considered as 0 when there was an overlap between fluorescence pixels in each channel.Three independent experiments were performed (N = 108 and 93 from control and knockdown, respectively).

Measuring apicoplast length
Elongated apicoplast length was measured using segmented/straight line tool in Fiji.Three independent experiments were performed (N = 60 parasites total).

Daughter cell-apicoplast proximity categories
For determining the apicoplast proximity to growing daughter cells, we visualized apicoplast localization (Cpn60 staining) in growing daughter parasites (based on eGFP-tubulin).Daughter proximity to the apicoplast was determined visually and required overlap between the fluorescent signal of the two structures.Parasites were categorized as being near to both the daughters (2D), near to one of the daughters (1D), or not close to either daughter (none).Three independent experiments were performed (N = 60 total).

TgAtg8 localization
TgAtg8 localization during different stages of apicoplast division was determined using anti-Ty antibody staining (a kind gift from Dr. Chris de Graffenried: Brown University).Two independent experiments were performed (25 parasites per experiment).

Statistical analysis
Statistical analysis was performed using GraphPad Prism software.Statistics on data sets were performed using unpaired t-tests.P-values < 0.05 were considered statistically significant.Superplots were made as described in reference 48.

The number of TgAPT1-EmGFP vesicles in the cytosol increases during apicoplast elongation
To investigate the mechanisms of apicoplast division and apicoplast protein trafficking from the ER, we generated a parasite line where the endogenous TgAPT1 gene was tagged with EmGFP (Emerald green fluorescent protein) (Fig. 1A and B; Fig S1 ) as TgAPT1 localizes to vesicles in the parasite cytosol and the apicoplast (14,16,48).Previously published data indicated that APT1 vesicles are formed in a cell cycle-dependent manner, so we transiently expressed mCherry-TubulinA1 which demarcates the parasite periphery of both the mother and growing daughter parasites to enable quantification of vesicle number at each stage of the parasite cell division cycle.Parasites were also stained with an anti-Cpn60 antibody, a marker of the apicoplast lumen.Based on the apicoplast shape and the presence and size of daughters, parasites were categorized as interphase (no daughter parasites, circular apicoplast), elongating (small daughters, elongated apicoplast), or post-fission (large daughters, circular apicoplast in each daughter), and the number of vesicles per parasite was determined at each cell cycle stage.In interphase parasites, TgAPT1 forms a ring that surrounds the TgCpn60 signal consistent with its localization at the apicoplast membrane (Fig. 1B), and the number of TgAPT1 positive (TgAPT1+) vesicles per parasite is low (1.15 ± 0.07, mean ± SEM) (Fig. 1B and C).During elongation, the number of TgAPT1 vesicles in the cytosol increased sevenfold to 8.08 ± 0.5 vesicles per parasite (Fig. 1B and C).After apicoplast fission and segregation into the daughter cells, the number of TgAPT1 vesicles per parasite decreased to 3.68 ± 0.2 (Fig. 1B and C).Notably, Cpn60, a nuclear encoded protein that localizes to the apicoplast lumen, was not found in TgAPT1+ vesicles, consistent with previous observations that apicoplast lumen proteins are not found in apicoplast vesicles (49) (Fig. 1B, inset).Consistent with the observations and interpretations made by Karnataki et al. (14), we hypothesize that these vesicles are ER-derived and transport newly synthesized proteins to the apicoplast in a cell cycle-dependent manner.
To further investigate the origin of the APT1 vesicles and determine if these vesicles are derived from or trafficking to the apicoplast, we designed a fluorescent pulse chase experiment using a parasite line expressing APT1-Halo under the endogenous APT1 promoter.Parasites were grown overnight and then pulse labeled with either Halo-Ligand JF646 or Oregon Green for 15 min (Fig. 1D).After this, "pulse" labeling parasites were grown without ligand for 6 h, labeled with Halo-Ligand TMR for 15 min (chase) and imaged immediately.Interphase parasites containing a circular apicoplast were labeled with JF646 only indicating that proteins synthesis did not occur after the first pulse labeling (Fig. 1D; left panel).Early in apicoplast division, indicated by short elongated apicoplasts, the apicoplast was labeled with both JF646 and TMR.In addition, Halo-Ligand TMR had an ER-like staining pattern (Fig. 1D "early elongation"; green arrowheads).Later in apicoplast division when the elongated apicoplast was longer, the apicoplast was stained with both fluorophores and the ER-like staining of TMR was reduced (Fig. 1D; "late elongation", inset).We quantified the localization of Halo-Ligand TMR ("chase") at different stages of the cell division cycle (Fig. 1E).There was a clear delineation in the staining.Ninety percent of parasites in interphase did not exhibit any TMR staining indicating that during interphase APT1 protein synthesis does not occur.During apicoplast division, 90% of parasites contained both an apicoplast and ER-like staining pattern indicating that APT1 protein synthesis and apicoplast division are temporally coordinated.Notably, the ER localization of APT1 was never observed in the APT1-EmGFP construct, likely due to inefficient fluorescent protein maturation in the ER.
We did not observe Halo-JF646-labeled vesicles in the cytosol during apicoplast elongation, as would be expected if the APT1 vesicles originated at the apicoplast.Due to the prevalent ER-like localization of Halo-TMR, we often did not observe Halo-TMR vesicles when taking single time point images.However, time-lapse imaging of Halo-TMR-labeled parasites revealed Halo-TMR-labeled vesicles which transiently moved away from the apicoplast and ER-like localizations, thus permitting their visualization in these diffraction-limited images (Fig. 1F; Video S1).Collectively, these results show that APT1 protein synthesis and the appearance of the APT1 vesicles occur in a cell cycle-dependent manner.In addition, it is likely that APT1 vesicles are ER derived and represent a mechanism to traffic nuclear-encoded proteins to the apicoplast.

Actin depolymerization and TgMyoF knockdown lead to an increased number of TgAPT1-EmGFP vesicles in the cytoplasm
TgAct1 and TgMyoF are cytoskeletal proteins with previously established roles in vesicle transport (38,39).Therefore, we wanted to determine if these proteins play a role in vesicle trafficking to the apicoplast.To do this, a previously generated auxin-inducible TgMyoF knockdown parasite line TgMyoF-mAID (39) was modified to express TgAPT1-EmGFP under the endogenous TgAPT1 promoter from the dispensable UPRT locus to create TgMyoF-mAID::TgAPT1-GFP parasite line (Fig. S2A).
To knockdown TgMyoF, TgMyoF-mAID::TgAPT1-EmGFP parasites were treated for 15 h with ethanol (control) or 500 µM IAA, which results in depletion of TgMyoF to undetect able levels after 4 h (39).To depolymerize actin, TgAPT1-EmGFP parasites were treated with DMSO (control) or 0.2 µM cytochalasin D (CD) for 12 h prior to fixation.Parasites transiently expressing mCherry-Tubulin were categorized as interphase, elongation or post-fission as described in Fig. 1.Both actin depolymerization and TgMyoF depletion resulted in an approximately twofold increase in the number of TgAPT1-EmGFP positive vesicles at all stages of the apicoplast division cycle (Fig. 2A through F; Fig. S3 and S4).In both treatment conditions, an apicoplast inheritance defect was observed as expected (37) (Fig. 2A and D; magenta arrow)."Apicoplast-less" parasites contained the highest accumulation of vesicles in comparison to all other stages of apicoplast division (Fig. 2C  and F).

TgAPT1 vesicles exhibit predominately diffusive-like motion
Next, we sought to describe the motion characteristics of TgAPT1+ vesicles, as previous work demonstrated that other vesicle types are transported throughout the parasite cytosol in an actin and TgMyoF-dependent manner (38,39).TgAPT1-EmGFP parasites in the elongation phase of apicoplast division were imaged by live cell microscopy.Visual inspection of these recordings revealed two striking observations.First, TgAPT1-EmGFP vesicles were found in the parasite cytosol, confirming observations from the fixed cell experiments (Fig. 1 and 2).The motion of these vesicles was predominately diffusive (Fig. 3A; inset 2), although a small number of vesicles exhibited directed motion, i.e.,sustained movement in a single direction for continuous frames (Fig. 3A; inset 1) (Video S2).Secondly, elongated apicoplasts were not static, linear structures, but exhibited changes in morphology throughout the 2-5min imaging period ( Video S3).
To quantify vesicle movement in an automated and unbiased manner, vesicle motion analysis was performed using the spot creation wizard in Imaris image analysis software (https://imaris.oxinst.com/)(Fig. 3B).To prevent aberrant tracking of fluorescent signal inside the apicoplast itself, a surface mask was created over the apicoplast (Fig. 3B; raw vs surface).Vesicles were automatically detected based on size and fluorescent intensity (Fig. 3B; surface and vesicles) (see Materials and Methods for details).Vesicles that are transported in an active, motor-driven manner typically exhibit "directed trajectories, " i.e., vesicles move in approximately a straight trajectory (38), while diffusive vesicles exhibit random movements.Thus, to determine the number of vesicles exhibiting diffusive or directed movement, track straightness was calculated by determining the ratio of track displacement (the distance between first and last positions of the vesicle) to track length (total distance traveled).A straightness value close to zero indicates random motion, i.e., track length is greater than the distance between the first and last positions indicating random motion.A straightness value close to 1 indicates that the total distance traveled and the distance between the first and last positions were similar indicating that the vesicle moved in approximately a straight line.Vesicle tracks observed in the first 20 s of imaging are color-coded according to straightness (Fig. 3B).A frequency distribution of the straightness values indicates that over 80% of vesicles have a straightness value less than 0.5 and the mean straightness value was 0.17 ± 0.15 confirming that apicoplast vesicle motion is predominantly diffusive in nature (Fig. 3C).We then compared the track displacement length for trajectories that had a straightness value greater or less than 0.5.Randomly moving vesicles (with a straightness value <0.5) had a mean track displacement value of 0.3 µm.While vesicles with straighter trajectories moved twice as far with a mean displacement of 0.59 µm (Fig. 3D).

Loss of TgMyoF and F-actin has only a minor impact TgAPT1-EmGFP vesicle motion
Next, we sought to determine if actin depolymerization or TgMyoF knockdown altered vesicle motion.TgAPT1-EmGFP parasites were grown for 18 hr prior to treatment with 2 µM CD or equivalent volume of DMSO for 30 min before live cell imaging (Fig. 3E) (Video S4), while to deplete TgMyoF, TgMyoF-mAID::TgAPT1-EmGFP parasites were grown for 15 hr with ethanol or 500 µm IAA before live cell imaging (Fig. 3F) (Video S5).
Vesicle movements were again tracked with Imaris as described above.Actin depolymerization or TgMyoF knockdown had no effect on the mean straightness values which were 0.18 ± 0.18, 0.14 ± 0.13, 0.2 ± 0.17, and 0.17 ± 0.17 (mean ± SE) for DMSO, CD, Ethanol, and IAA-treated parasites, respectively (Fig. 3G and H).Although, upon loss of F-actin and TgMyoF, the number of vesicles in the cytosol was increased compared to controls.In all cases (treatment and control), the percentage of vesicles with a straight ness value <0.5 was ~80%.To determine if the movement of the remaining ~20% of vesicles was TgMyoF and actin driven, vesicles with a straightness value greater than 0.5 were analyzed further.Since vesicles with a straightness value with greater than 0.5 exhibited longer track displacement (mean length was 0.59 µm [Fig.3D]), we defined "directed trajectories" as those with a straightness value >0.5 and a track displacement length of at least 0.5 µm.In no treatment, DMSO-and ethanol-treated ethanol: interphase: 0.8 ± 0.14; elongation: 7.57 ± 0.27; and post-fission: 1.83 ± 0.16.IAA interphase: 9.9 ± 0.52; elongation: 10.7 ± 0.43; post-fission: 6.6 ± 0.24; no apicoplast = 17.1 ± 0.58.The mean from each experiment is indicated with large circles.These values were used to calculate the mean, standard error of the mean, and P-value.Raw data are shown with smaller colored circles.Experiment   parasites, the percentage of vesicles which met these criteria were 9.5%, 10.3%, and 11%, respectively.Upon actin depolymerization with CD or TgMyoF knockdown, this percentage was reduced to 4.7% and 6.8%, respectively (Fig. 3I).Although these changes are statistically significant, overall our data indicate that only a small portion of APT1 vesicles is transported in active, actomyosin-dependent manner.

TgMyoF knockdown results in reduced apicoplast length
If vesicle accumulation was due to either a defect in protein trafficking to the apicoplast or fragmentation of the apicoplast, we predict that this would alter apicoplast morphol ogy.Therefore, we assessed apicoplast length in the elongation phase of apicoplast division in the absence of TgMyoF.In order to compare parasites at the same stage of cell division, parasites were transiently transfected with a GFP-tubulin expression cassette in order to fluorescently label the pellicle of mother and daughter parasites.Parasites were grown for 12 h prior to initiating a 6-h treatment with ethanol or 500 µM IAA.This shortened IAA treatment time compared with previous experiments allowed us to visualize the apicoplast in the first cell division cycle after depletion of TgMyoF.Parasites were then stained with an anti-Cpn60 antibody and DAPI before image acquisition (Fig. 4A).During elongation, TgMyoF knockdown resulted in a ~25% reduction in apicoplast length (1.54 ± 0.05 µm, mean ± SEM) when compared to the control parasites (2.11 ± 0.05 µm, mean ± SEM) (Fig. 4B).Apicoplast morphology was not disrupted in TgMyoF knockdown parasites in interphase and post-fission stages of division (Fig. S5).

The apicoplast undergoes morphological changes during elongation
Live cell imaging of TgAPT1-EmGFP parasites demonstrated that the apicoplast morphology in the elongation phase of apicoplast division was highly variable between individual vacuoles.In some vacuoles, the apicoplast appeared to be highly dynamic exhibiting alternating linear, U-shaped, and branched morphologies through out the ~2-min imaging period (Fig. 4A; top panel) (Video S3), while others were static in the elongation phase of apicoplast division (Fig. 5A; bottom panel) (Video S3).Given this variability, we quantified the percentage of apicoplasts that were highly dynamic and exhibited rapid changes in morphology during the course of imaging (102 s) (Fig. 5A and B)."Dynamic" apicoplasts were defined as those that exhibited one or more of the following morphological changes: extending and/or retracting, bending, branch formation, or branch movement (Fig. 5A; top panel); 63% ± 6.1% (mean ± SEM) of the apicoplasts were highly dynamic (Fig. 5B), whereas 37% ± 6.1% (mean ± SEM) of parasites contained either less dynamic apicoplasts, exhibiting subtle changes in morphology, or were static with no observable changes in the morphology throughout the ~2-min imaging time (Fig. 5A and B).

Apicoplast dynamics requires actin and TgMyoF proteins
Considering the dynamic nature of the apicoplast, we wanted to determine if these dynamics are dependent on actin and TgMyoF.To depolymerize actin, TgAPT1-EmGFP parasites were again treated with 2 µM cytochalasin D or equivalent volume of DMSO for 30 min before live cell imaging (Fig. 5C).Of the apicoplasts in control DMSO-treated parasites, 82% ± 2.9% (mean ± SEM) were highly dynamic, while only 9% ± 6.6% (mean ± SEM) of the apicoplasts in the cytochalasin D-treated parasites were highly dynamic (Fig. 5D).To test the role of TgMyoF in apicoplast dynamics, TgMyoF-mAID: TgAPT1-EmGFP parasites treated with 500 µM IAA to induce TgMyoF knockdown for 15 h prior to live cell imaging (Fig. 5E).Similar to the cytochalasin D treatment, TgMyoF knockdown resulted in reduced apicoplast dynamics.The percentage of parasites with a highly dynamic apicoplast was significantly reduced from 86% ± 4% (mean ± SEM) in control parasites to 22% ± 13% in the TgMyoF KD parasites (Fig. 5F).

The apicoplast dynamically associates with static centrosomes
Apicoplast association with the centrosomes is a crucial step in the apicoplast division cycle.In the first step in the cell division cycle, the centrosome migrates to the basal end of the nucleus and duplicates; then both centrosomes return to the apical side of the nucleus where they associate with the apicoplast (27,50).The dynamics of the elongated apicoplast (Fig. 5) have not previously been described, and it is unclear if or how the apicoplast interaction with the centrosomes, which are embedded in the nuclear envelope (51), is maintained during apicoplast movement.
To further explore this phenomenon, wild-type parasites were transiently transfected with plasmids containing FNR-RFP (a marker for the apicoplast [27])) and Centrin1-GFP (a marker for the centrosome [52]) expression cassettes and visualized using live-cell microscopy.In parasites with highly dynamic, elongated apicoplasts the centrosomes were not always associated with the tips of the apicoplast but were frequently associ ated with the sides of the elongated apicoplast (Fig. 6A, insets 1 and 3) (Video S6).The apicoplast appears to slide in relation to the static centrosome (Fig. 6A, inset 1; Video S7).The apicoplast was observed to undergo branching when the centrosome was associated with the side rather than the tip of the apicoplast (Fig. 6A, inset 3, magenta arrow) (Video S8).The most dramatic morphological changes occurred when the apicoplast was associated with only a single centrosome (Fig. 6A, inset 2; Video S9).The apicoplast was observed extending toward the second centrosome before retracting backward.In contrast to the dynamics observed in the elongation phase of apicoplast division, interphase and post-fission apicoplasts were static.In interphase parasites, the apicoplast was circular and did not change shape over the course of imaging (Fig. 6B; upper panel).After apicoplast fission, the apicoplast remains elongated for a time before retraction into a circular morphology.Post-fission-elongated apicoplasts are static over the course of the imaging period.The apicoplast movement and branching that were observed prior to fission are no longer observed and centrosomes are statically associated with the apicoplast tips (Fig. 6B; lower panel).

TgMyoF knockdown disrupts apicoplast and centrosome interaction
The live cell imaging of apicoplast and centrosome interactions suggest that the actomyosin-driven apicoplast dynamics are important for initiating the interaction between these two structures and positioning of the centrosomes at the apicoplast  tips (Fig. 6).Therefore, we directly investigated if the loss of TgMyoF, which disrup ted apicoplast dynamics (Fig. 5), also disrupted apicoplast-centrosome interactions.TgMyoF-mAID parasites transiently expressing the centrosome marker Centrin1-GFP were grown for 12 h post invasion and then treated with ethanol or 500 µM IAA for 6 h before fixation and staining with an anti-Cpn60 antibody and DAPI.In control parasites, at the early stages of apicoplast division when the length of the elongated apicoplast is short (less than 1.5 µm), the centrosomes and the apicoplast are frequently not associated with one another (Fig. 7A) and the mean apicoplast to centrosome distance is 0.99 ± 0.18 µm (mean ± SEM) (Fig. 7B and C).Later in the apicoplast division cycle, when the length of the apicoplast is greater than 1.5 µm, the mean centrosome-apicoplast distance is significantly shorter at only 0.36 ± 0.05 µm (Fig. 7A through C) and the centrosomes were close with either the sides or the tips of the apicoplast (Fig. 7A).However, in TgMyoF knockdown parasites, the apicoplast to centrosome distances were 1.2 ± 0.12 and 1.0 ± 0.09 µm when the apicoplast length was greater than or less than 1.5 µm, respectively, (Fig. 7A through C) indicating that TgMyoF is needed for recruiting the apicoplast to the centrosomes.
We next assessed how the loss of centrosome interaction affected apicoplast inheritance.In 100% of control parasites, elongated apicoplasts were found to be in close proximity to both of the growing daughter parasites.In early stages of division, the apicoplast was adjacent to the daughter parasites (Fig. 8A; white arrowhead).Later in division, both daughters contain one end of the elongated apicoplast, even when one of the daughters was in the "upside down" orientation (i.e., the apical ends of the daughter parasites have opposite orientations within the mother) (Fig. 7C; magenta arrowhead).Upon depletion of TgMyoF, only 13.3% of elongated apicoplasts were in close proximity to both the growing daughter parasites (2D), while 70% of the elongated apicoplasts were in close proximity to only one of the daughter parasites (1D) (Fig. 8A; white arrow), and 16.6% of the elongated apicoplasts were not found to be in close proximity to either of the growing daughter parasites (none) (Fig. 8B; Fig. S6).Thus, the disruption of the centrosome-apicoplast interaction leads to asymmetric segregation of the apicoplast into a single daughter cell or to the residual body.

TgMyoF knockdown did not alter TgAtg8 localization with the apicoplast
TgAtg8 is a protein that localizes to the apicoplast outer membrane.Loss of TgAtg8 disrupts the interaction between the centrosomes and apicoplast resulting in asymmet ric distribution of the apicoplast to the growing daughter parasites, reminiscent of the phenotype observed after TgMyoF depletion (Fig. 8A and B) (28).To determine if loss of TgMyoF affected the localization of TgAtg8 at the apicoplast, TgMyoF-mAID para sites transiently expressing 3xTy-TgAtg8 under the endogenous TgAtg8 promoter were treated with ethanol or 500 µM IAA for 6 h before fixation and staining with anti-Cpn60 antibody.TgAtg8 localized to the apicoplast at all stages of the apicoplast division cycle and knockdown of TgMyoF did not disrupt TgAtg8 localization (Fig. 8C; Fig. S7) indicating that the loss of interaction between centrosomes and the elongated apicoplast seen in the TgMyoF knockdown parasites is not due to mislocalization of TgAtg8.

DISCUSSION
The apicoplast is an essential organelle in T. gondii.In order to maintain a functional apicoplast, the parasite must execute two distinct and essential cellular processes: division of this single copy organelle once per cell cycle to ensure each daughter parasite inherits one apicoplast and trafficking of nuclear-encoded proteins to the apicoplast.We investigated the role of actin and an associated myosin motor, TgMyoF, in these processes.

Loss of TgMyoF and actin results in the accumulation of TgAPT1 vesicles in the parasite cytosol
Vesicle trafficking is a multi-step process involving vesicle formation and budding from the donor compartment, transport, docking at the target membrane and finally vesicle fusion.While it is recognized that trafficking of nuclear-encoded proteins to the apicoplast is a vital cellular process, our understanding of this trafficking process is incomplete and a number of trafficking pathways have been proposed (53).Using TgAPT1-EmGFP-expressing parasites, our data confirm previously published data (14) demonstrating that vesicle formation is coordinated with the cell cycle, with the number of vesicles increasing sevenfold during the elongation phase of apicoplast division.We were not able to definitively determine the origin of these vesicles, and we considered two possible hypotheses: first, APT1 vesicles originate at the apicoplast and represent a mechanism for trafficking apicoplast-synthesized lipids to other subcellular locations, or second, these vesicles contain newly synthesized nuclear encoded apicoplast proteins that originate at the ER and represent a mechanism for trafficking these proteins to the apicoplast as previously proposed (14,49).A number of lines of evidence led us to favor the latter hypothesis: (1) vesicle accumulation occurs in "apicoplast-less" parasites indicating that vesicle formation is not dependent on the apicoplast itself, as previously observed (Fig. 2) (49).(2) APT1 protein synthesis and vesicle formation both occur in a short window in the cell cycle when the apicoplast is undergoing division (Fig. 1D and  E).(3) Using the Halo pulse-chase assay and time lapse imaging, we observed Halo-TMRlabeled "chase" vesicles in the parasite cytosol (Fig. 1F).One caveat to this observation is that we did not observe vesicles in all parasites or at all time points.We speculate that the pervasive ER-like Halo-TMR signal hampered our ability to visualize APT1-TMR vesicles in the cytosol.( 4) Recent work has shown that the SNARE proteins TgVAMP4-2 and TgSNAP29 localize to the apicoplast and are essential for the trafficking of nuclear encoded proteins to the apicoplast (54,55), which highlight a role for vesicles in this trafficking pathway, and (5) while we cannot exclude the possibility that some of these vesicles are derived from the apicoplast itself, vesicle formation at and budding from the apicoplast has never been described in the literature.Additionally, chloroplast to ER lipid trafficking in plants occurs by Acyl-coA's and does not involve vesicle formation (56).

TgMyoF and actin only play a minor role in APT1 vesicle motion
Given the role of TgMyoF and actin in dense granule transport, we sought to determine if this cytoskeletal system was also required for APT1 vesicle transport.By tracking the motion of individual vesicles in live parasites, we demonstrate that APT1 vesicle motion appears predominately diffusive-like, with ~80% of vesicles moving in a random manner, apicoplast-centrosome distance of 0.75 µm.In control parasites, the apicoplast-centrosome distance is short (90% of distances are less than 0.75 µm) when apicoplast length is above 1.5 µm.However, in TgMyoF-KD parasites when the apicoplast length is above 1.5 µm, 53% of apicoplasts are greater than 0.75 µm away from the centrosome.Experiment 1 in orange, experiment 2 in gray, experiment 3 in blue.(C) Parasites were categorized as having an apicoplast greater or less than 1.5 µm, and the mean apicoplast-centrosome distance was calculated for each category in control and TgMyoF-KD parasites.In control parasites, the apicoplast centrosome distance is significantly shorter in parasites containing a long apicoplast (>1.5 µm) compared with control parasites containing a short apicoplast (<1.5 µm).In TgMyoF-KD parasites, the mean apicoplast-centrosome distance was ~1 µm, regardless of apicoplast length.
i.e., those with a straightness value less than 0.5 (Fig. 3)."Directed" movements were defined as those with a straightness value greater than 0.5 and a displacement length greater than 500 nm.Only ~10% of vesicles in control parasites met these criteria.Actin depolymerization or TgMyoF depletion reduced this percentage by ~50% (Fig. 3).Therefore, we conclude that only a small proportion of vesicles are transported in an actomyosin dependent manner.Given that 90% of vesicles exhibit diffusive motion, it is unlikely that the small change in the percentage of vesicles exhibiting directed motor-driven movement can account for the large accumulation in vesicles seen after CD and IAA treatment.
A 12-h treatment with CD to induce actin depolymerization or a 15-h treatment with IAA to deplete TgMyoF caused to a significant increase in the number of vesicles in the cytosol at all phases of the cell division cycle (Fig. 2).Given the minor role of this actomyosin system in vesicle transport, and our hypothesis that the APT1 vesicles apicoplast not in close proximity to either of the daughter parasites.In control parasites, 100% of apicoplast were in close proximity to both daughter parasites compared to 13.33% ± 6% of TgMyoF-KD parasites.In TgMyoF-KD parasites, 70% ± 2.9% and 16.7% ± 4.4% of apicoplasts were associated with one daughter parasite or neither daughter parasite respectively (mean ± SEM).Bar graph shows quantification of combined data from 3 independent experiments (20 parasites per experiment).Statistical significance was determined by unpaired Student's t test.(C) TgMyoF-mAID parasites ectopically expressing 3xTy-Atg8 under the Atg8 promoter were grown for 12 h before treatment with ethanol (control) or IAA (TgMyoF KD) for 6 h.Cells were stained with anti-Ty (magenta) and anti-Cpn60 antibodies (yellow).Maximum intensity projections of deconvolved images of parasites at interphase, elongation, and post-fission show that loss of TgMyoF does not affect Atg8 localization.Scale bar represents 5 µm.represent a mechanism for trafficking proteins to the apicoplast, we hypothesize that TgMyoF and F-actin play a role in vesicle fusion with the apicoplast.Disruption of vesicle fusion would lead to an accumulation of APT1 vesicles in the cytosol at all stages of the cell division cycle (Fig. 2).Further work is required to mechanistically uncover the role of TgMyoF and F-actin in this process.It should be noted that we cannot exclude the possibility that some or all of the accumulated vesicles are due to vesiculation of the apicoplast itself.

TgMyoF and Actin control apicoplast dynamics which mediate apicoplastcentrosome association
The sequence of events that occur during the apicoplast division cycle had previously been established via fixed cell immunofluorescence assays or live-cell imaging with low temporal resolution (time between frames on the order of minutes) (27,32) (Fig. 1A).This study reveals new insight into how this sequence of events proceeds (Fig. 9).In interphase, parasites contain a single apicoplast and a single centrosome.At the onset of cell division, centrosomes migrate to the basal end of the nucleus and duplicate (47).The apicoplast transitions from a circular to an elongated morphology.Elongated apicoplasts with a length <1.5 µm were on average ~1 µm away from the centrosomes.Thus, the transition from circular to elongated occurs before and is not dependent on centrosome association.This transition also occurred in the absence of both actin and TgMyoF, indicating that an as yet unidentified protein(s) must control apicoplast membrane deformation.Since apicoplast elongation is not dependent on centrosome association, a mech anism must be in place to drive the interaction of these two organelles.Live cell imaging with high temporal resolution provided insight into how this is achieved.Apicoplast branching, extension, and retraction are needed to move the apicoplast in close proximity to the nuclear envelope-embedded centrosomes to facilitate an interaction with either the sides or tips of the apicoplast.Centrosome sliding along the apicoplast side and branch formation and retraction are likely important for position ing of the centrosome at the apicoplast tips.Consistent with this, the localization of TgAtg8 which plays a role in mediating these interactions is not restricted to the apicoplast tips but found on the entire apicoplast surface (Fig. 8).Loss of TgMyoF results in decreased apicoplast dynamics (Fig. 5) and centrosome association (Fig. 7), which ultimately leads to apicoplast inheritance defects (Fig. 8).Despite the defects in centrosome-apicoplast association in the absence of TgMyoF, TgAtg8 remains localized at the apicoplast.Collectively, these data indicate that actomyosin-driven apicoplast dynamics are important to facilitate the association between these two organelles and for positioning of the centrosomes at the apicoplast tips.
There are still a number of important questions that need to be addressed in future studies.Previous work (35) has demonstrated that the actin nucleator formin 2 relocalizes from the Golgi to the centrosomes during apicoplast division; however, it is unclear how F-actin interacts with the apicoplast in order to drive apicoplast dynam ics.Since TgMyoF and actin are not required for apicoplast elongation, another as yet unidentified protein or proteins must drive membrane deformation and the dramatic transition from a circular to elongated morphology that occurs at the start of apicoplast division.
Overall, our results have determined TgMyoF and actin control apicoplast dynamics which is needed for apicoplast-centrosome association, a vital step of the apicoplast division cycle.

FIG 1
FIG 1 TgAPT1 vesicle trafficking occurs in the elongation phase of apicoplast division (A) Schematic of the apicoplast division cycle in T. gondii.During interphase, the apicoplast is circular and located at the parasite's apical end.At the start of the cell division, apicoplast undergoes elongation and associates with a pair of duplicated centrosomes.Centrosome movement into daughter parasites results in a U-shaped apicoplast.Membrane fission is driven by TgDrpA and results in a single apicoplast being inherited by each daughter cell.Inset: Depiction of apicoplast morphology.The apicoplast contains four membranes: outermost, periplastid (green), and two innermost membrane (black).TgAPT1 localizes to the apicoplast membranes.Cpn60 localizes to the apicoplast lumen.(B) Maximum intensity projection of deconvolved epifluorescent images of the parasites in interphase, elongation, and post-fission stages of the apicoplast division cycle.Parasites are expressing TgAPT1-EmGFP (green) and mCherry-TubulinA1 (a marker for mother and daughter parasite pellicle shown in magenta).Parasites were fixed and stained with an anti-Cpn60 antibody (a marker for the apicoplast lumen shown in yellow) and DAPI (cyan).Scale bar represents 5 µm.Inset: brightness and contrast of insets were adjusted to facilitate visualization of dimmer TgAPT1 vesicles (white arrows).Inset scale bar represents 1 µm.(C) Quantification of the number of TgAPT1-EmGFP vesicles per parasite at interphase, elongation, and post-fission from three independent experiments (N = 60 parasites).The mean from each (Continued on next page)

FIG 1 (
FIG 1 (Continued)experiment is indicated with large circles.These values were used to calculate the mean, standard error of the mean, and P-value.Raw data are shown with smaller colored circles.Experiment 1 in orange, experiment 2 in gray, experiment 3 in blue.Statistical significance was determined by unpaired Student's t test.(D) Pulse-chase assay to track APT1 protein synthesis.
white box indicates area used to make inset.Brightness and contrast settings of the inset were adjusted to highlight the dim TMR signal.Scale bar represents 5 µm.Inset scale bar represents 1 µm.(E) Localization of Halo-TMR "chase" labeling during interphase and apicoplast division.N = 65 vacuoles from three independent experiments.Experiment 1 in orange, experiment 2 in gray, experiment 3 in blue.(F) APT1-Halo expressing parasites labeled with Halo-TMR (chase) were visualized using time-lapse imaging.Halo-TMR-labeled vesicles are observed adjacent to the apicoplast.Dashed white box indicates the area used to make the inset.Inset: appearance of Halo-TMR-labeled vesicles at various time points (white arrows).Time point in seconds is indicated in the top right-hand corner of each image.It should be noted that Halo-TMR vesicles were not observed in all parasites or at all time points.The apicoplast and ER-like staining pattern likely obscures the vesicle signal in these diffraction-limited images.Scale bar represents 5 µm.Inset scale bar represents 1 µm.

FIG 2
FIG 2 Actin depolymerization and TgMyoF knockdown results in an accumulation of APT1 vesicles.(A) TgAPT1-EmGFP parasites treated with DMSO (control) or 0.2 µM cytochalasin D for 12 h were categorized as interphase, elongation, or post-fission based on the stage of the apicoplast division cycle and the size and presence of daughter parasites.Magenta arrow indicates parasite that does not contain an apicoplast.White box indicates area used to make inset in panel B. Images are maximum intensity projections of deconvolved images.Scale bar represents 5 µm.(B) Inset panels from panel A. Brightness and contrast of insets were adjusted to facilitate visualization of dimmer TgAPT1 vesicles.Inset scale bar represents 1 µm.(C) Quantification of the number of TgAPT1-EmGFP vesicles per parasite during interphase, elongation, and post-fission stages of the apicoplast division cycle.Parasites with no apicoplast were categorized separately.DMSO: n = 42/apicoplast division stage, n = 126 total; CD: n = 30-42/apicoplast division stage, n = 156 total.TgAPT1 vesicle number per parasite are as follows (mean ± SEM): DMSO interphase: 1.05 ± 0.1; elongation: 8.26 ± 0.73; and post-fission: 4.19 ± 0.39.CD interphase: 3.59 ± 0.15; elongation: 13.30 ± 0.46; post-fission: 6.83 ± 0.35; no apicoplast: 19 ± 0.46.The mean from each experiment is indicated with large circles.These values were used to calculate the mean, standard error of the mean, and P-value.Raw data are shown with smaller colored circles.Experiment 1 in orange, experiment 2 in gray, experiment 3 in blue.Statistical significance was determined by unpaired Student's t test.(D) TgMyoF-mAID:TgAPT1-EmGFP parasites treated with ethanol (control) or 500 µM IAA for 15 h were categorized as either interphase, elongation, or post-fission based on the stage of the apicoplast division cycle and the size and presence of daughter parasites.Magenta arrow indicates parasite that does not contain an apicoplast.White box indicates area used to make inset in panel E. Images are maximum intensity projections of deconvolved images.Scale bar represents 5 µm.(E) Inset panels from panel D. Brightness and contrast of insets were adjusted to facilitate visualization of dimmer TgAPT1 vesicles.Inset scale bar represents 1 µm.(F) Quantification of the number of TgAPT1-EmGFP vesicles per parasite during interphase, elongation, and post-fission in control and TgMyoF knockdown parasites.Parasites with no apicoplast were categorized separately.(EtOH: n = (Continued on next page)

FIG 3
FIG 3 APT1 vesicles exhibit predominately diffusive motion.(A) Localization TgAPT1-EmGFP parasites in live parasites.TgAPT1 is localized to both the apicoplast and vesicles.White boxes indicate areas used to make insets 1 and 2. Inset 1 depicts TgAPT1+ vesicle moving in a directed manner.Inset 2 depicts a TgAPT1+ vesicle moving with diffusive-like movement.Scale bar = 5 µM.Inset scale bar = 1 µM.(B) The Z-stack movie of TgAPT1-EmGFP expressing parasites was (Continued on next page)

FIG 3 (
FIG 3 (Continued) imported into Imaris.Raw data from frame 1 of the movie are shown.Isosurfaces of the apicoplast were reconstructed to prevent aberrant tracking inside the apicoplast itself.Vesicle tracks observed in the first 20 s of imaging are color-coded by straightness value.(C) Frequency distribution of trajectory straightness values of TgAPT1 vesicles.Straightness value close to zero indicates random motion.Straightness value close to 1 indicates directed/linear movement.Mean straightness value is 0.18 ± 0.15.(D) Frequency distribution of displacement lengths of TgAPT1+ vesicles.Vesicles with a straightness value less than 0.5 (black) had a mean displacement length of 0.3 ± 0.27 µm (mean ± SD), while vesicles with a straightness value greater than 0.5 (orange) had a mean displacement length of 0.59 ± 0.3 µm (mean ± SD). (E) Localization TgAPT1-EmGFP parasites in live parasites treated with DMSO or CD for 30 min prior to imaging.Raw data from frame 1 of the movie are shown.Z-stack movies were imported into Imaris, isosurfaces of the apicoplast and vesicles were reconstructed.Vesicle tracks observed in the first 20 s of imaging are color-coded by straightness value.(F) Localization TgAPT1-EmGFP parasites in live parasites treated with ethanol or IAA for 15 h prior to imaging.Raw data from frame 1 of the movie are shown.Z-stack movies were imported into Imaris; isosurfaces of the apicoplast and vesicles were reconstructed.Vesicle tracks observed in the first 20 s of imaging are color-coded by straightness value.(G) Frequency distribution of trajectory straightness values of TgAPT1 vesicles from DMSO-treated (black) and cytochalasin D (CD)-treated (blue) parasites.Mean straightness values are for DMSO and CD treated are 0.19 ± 0.18 and 0.14 ± 0.13 (mean ± SD), respectively.Straightness value close to zero indicates random motion.Straightness value close to 1 indicates directed movement.(H) Frequency distribution of trajectory straightness values of APT1 vesicles from ethanol (black) and IAA (green)-treated parasites.Mean straightness values are for ethanol and IAA treated are 0.2 ± 0.18 and 0.17 ± 0.17 (mean ± SD), respectively.Straightness value close to zero indicates random motion.Straightness value close to 1 indicates directed movement.(I) Bar graph indicating the percentage of vesicles exhibiting directed motion, defined as a straightness value greater than 0.5 and a track displacement length of greater than 500 nm.Actin depolymerization with CD or TgMyoF knockdown with IAA results in a significant (~50%) decrease in the percentage of directed trajectories.

FIG 4
FIG 4 TgMyoF knockdown results in reduced apicoplast length.(A) TgMyoF-mAID parasites expressing eGFP-tubulin (magenta) were grown for 12 h before treatment with ethanol (control) or IAA (TgMyoF KD) for 6 h.Parasites were stained with an anti-Cpn60 antibody (yellow) and DAPI (cyan).Scale bar represents 5 µm.(B) Quantification of the apicoplast length in control and TgMyoF KD parasites from 3 independent experiments (20 parasites per experiment).Mean apicoplast length was 2.12 ± 0.05 and 1.55 ± 0.05 µm for ethanol-and IAA-treated parasites, respectively.The mean from each experiment is indicated with large circles.These values were used to calculate the mean, SEM, and P-value.Raw data are shown with smaller colored circles.Experiment 1 in orange, experiment 2 in gray, experiment 3 in blue.Statistical significance was determined by unpaired Student's t test.

FIG 5
FIG 5 Apicoplast dynamics require actin and TgMyoF.(A) Time lapse images of TgAPT1-EmGFP parasites during the elongation phase of apicoplast division.Upper: the apicoplast is dynamic and exhibits branch formation and movement over the ~2 min imaging period.Lower: an example of a static elongated apicoplast.Apicoplast does not change shape over the imaging period.White box indicates area used to make time-lapse inset.Inset indicates change in apicoplast morphology over time.Time in seconds is indicated in the top right corner of each image.Scale bar represents 5 µm.Inset scale bar = (Continued on next page)

FIG 5 (
FIG5 (Continued) 1 µm.(B) Quantification of apicoplast dynamics.Apicoplast was categorized as dynamic if one of the following was observed: dramatic changes in apicoplast shape such as bending, branch formation, branch movement, or retraction.Apicoplast was categorized as static of there were subtle or no changes in apicoplast shape or position.N = 70 from four independent experiments.63% ± 6.2% of parasites contained dynamic apicoplasts.37 ± 6.2 contained static apicoplasts.Data are represented as mean ± SEM.Statistical significance was determined by unpaired Student's t test.(C) Time lapse images of TgAPT1-EmGFP parasites during the elongation phase of apicoplast division imaged after treatment with DMSO or cytochalasin D. White box indicates area used to make time-lapse inset.Inset indicates change in apicoplast morphology over time.Time in seconds is indicated in the top right corner of each image.Scale bar represents 5 µm.Inset scale bar = 1 µm.(D) Quantification of apicoplast dynamics after DMSO or CD treatment.84% ± 2.9% of DMSO-treated parasites (gray) contained dynamic apicoplasts.9% ± 6.7% of CD-treated parasites (blue) contained dynamic apicoplasts.Data are represented as mean ± SEM.N = 32 and 41 for DMSO and CD respectively from three independent experiments.Statistical significance was determined by unpaired Student's t test.(E) Time lapse images of TgMyoF-mAID:TgAPT1-EmGFP parasites during the elongation phase of apicoplast division imaged after treatment with ethanol or IAA.White box indicates area used to make time-lapse inset.Inset indicates change in apicoplast morphology over time.Time in seconds is indicated in the top right corner of each image.Scale bar represents 5 µm.Inset scale bar = 1 µm.(F) Quantification of apicoplast dynamics after ethanol or IAA treatment.86% ± 4% of ethanol-treated parasites (gray) contained dynamic apicoplasts.22% ± 13% of IAA-treated parasites (green) contained dynamic apicoplasts.Data are represented as mean ± SEM.N = 38 and 40 for ethanol and IAA, respectively.Statistical significance was determined by unpaired Student's t test.

FIG 6
FIG 6 Centrosome-apicoplast interaction is maintained during apicoplast dynamics.(A) Maximum intensity projection of a deconvolved images of parasites expressing eGFP-centrin1 (centrosome marker; magenta) and FNR-RFP (apicoplast marker; green) during the elongation phase of apicoplast division.Parasites were imaged on five focal planes for a total of 10 min.Z-slice step size was 0.4 µm.Dashed boxes indicate areas used to make time lapse insets.Inset 1: The apicoplast is attached to a single centrosome and slides in relation to the static centrosome, associating with both the tips and sides of the apicoplast.Inset 2: The same apicoplast at a later time point undergoes dramatic morphological change, extending to ~3 µm in length toward the second centrosome before collapsing backward.Inset 3: Two centrosomes associated with the sides of the apicoplast.Apicoplast "branching" (magenta arrow) is observed at the site of centrosome association.Imaging time points in minutes are indicated in the top right-hand corner of each image.Scale bar represents 5 µm, inset scale bar represents 2 µm.(B) Interphase (upper) and post-fission (lower) apicoplasts did not exhibit morphological changes throughout the course of imaging.Imaging time points in minutes are indicated in the top right-hand corner of each image.Scale bar represents 2 µm.

FIG 7
FIG 7 TgMyoF knockdown disrupts the apicoplast and centrosome interaction.(A) TgMyoF-mAID parasites expressing eGFP-Centrin1 (green) were grown for 18 h before treatment with either ethanol or IAA for 6 h.Parasites were fixed and stained with an anti-Cpn60 (magenta) antibody and DAPI (gray).Images are maximum intensity projections of deconvolved (Continued on next page)

FIG 7 (
FIG 7 (Continued) images.Based on the length of the apicoplast control, parasites were categorized as in either early or late elongation.Since the apicoplast elongation length was reduced (Fig. 4) in TgMyoF-KD parasites, these parasites could not be categorized into early or late stages of elongation.White box in indicates area used to make inset.Inset: white dashed line and white number indicate apicoplast length.Green dashed line and green number indicate the distance between the apicoplast and the centrosome.Scale bar represents 5 µm.Inset scale bar represents 2 µm.(B) Distance between the centrosome and apicoplast was plotted as a function of apicoplast length in control and TgMyoF-KD parasites.Vertical dashed line indicates an apicoplast length of 1.5 µm which is used as the cut off for categorization in Fig. 7C.The horizontal dashed line represented an

FIG 8
FIG 8 Apicoplast inheritance defects are not due to changes in Atg8 localization.(A) TgMyoF-AID parasites expressing eGFP-tubulin (magenta) were treated with ethanol or IAA for 6 h and then fixed and stained with an anti-Cpn60 antibody (yellow).Control parasites show that the elongated apicoplasts are in proximity to the respective daughter parasites (white arrowheads).Later in division as indicated by larger daughter parasites, the tips of the elongated apicoplast are localized within the growing daughters (magenta arrowheads).Upon TgMyoF knockdown, there is an asymmetric localization of the elongated apicoplast to one of the daughter cells (white arrow) and absent from the second daughter parasite (asterisks).Scale bar represents 5 µm.(B) Quantification of the percentage of daughter parasites in close proximity to the elongated apicoplast in control parasites and TgMyoF KD parasites.Parasites were categorized as follows: 2D, elongated apicoplast near or within both daughter parasites; 1D, elongated apicoplast in close proximity to one of the daughter parasites; and "none, " elongated

FIG 9
FIG9 Illustration of the role of actin and TgMyoF in apicoplast division.In interphase parasites, T. gondii contains a single apicoplast and centrosome.At the onset of parasite division, centrosomes migrate to the basal end of the nucleus and duplicate.Apicoplast elongation begins in a manner independent of centrosome association, and there is an increase in the number TgAPT1+ vesicles in the cytosol, which could supply the apicoplast with additional proteins and lipids prior to duplication.In the elongation phase, the apicoplast is highly dynamic exhibiting bending and branch formation.At this stage, the centrosome is dynamically associated with both the sides and the tips of the apicoplast.Association of the apicoplast tips with the centrosome facilitates inheritance of a single apicoplast by each daughter cell upon apicoplast fission.In the absence of TgMyoF and F-actin, TgAPT1+ vesicles accumulate in the cytosol.These vesicles could represent either vesiculation/fragmentation of the apicoplast or ER-derived vesicles that have failed to fuse with the apicoplast.Also during elongation, the apicoplast is static and the apicoplast frequently fails to associate with one or both centrosomes.This results in the asymmetric division of the apicoplast into daughter parasites.