In an effort to make Capsaspora a tractable system for evolutionary cell biology studies, we developed genetic tools for overexpression and loss-of-function analysis. We first established a method for engineering stable Capsaspora cell lines by transfecting a vector encoding the fluorescent protein mScarlet and resistance to the antibiotic Geneticin (GenR). After transfection, antibiotic selection, and generation of clonal lines by serial dilution, we were able to generate uniformly mScarlet-positive clonal populations of cells (Figure 1—figure supplement 2A). Transfection of Capsaspora with genes encoding resistance to nourseothricin (NatR) or hygromycin (HygR) followed by selection with the respective antibiotic could also generate drug-resistant transformants (Figure 1—figure supplement 2B). These results demonstrate that stable Capsaspora cell lines expressing multiple transgenes can be generated.

We next sought to develop tools for loss-of-function analysis in Capsaspora. After unsuccessful attempts at CRISPR-Cas9-mediated genome editing, which may be due to the lack of nonhomologous end-joining machinery components in Capsaspora, we used PCR-generated targeting constructs to achieve gene disruption by homologous recombination, replacing the putative TBD from each allele of the coYki gene with a distinct selectable antibiotic marker (Figure 1H; see ‘Materials and methods’ for a detailed description of generating a deletion mutant). This strategy allowed us to generate a cell line in which both coYki alleles were disrupted (Figure 1—figure supplement 2C–E), with no detectable expression of the deleted region of coYki (Figure 1I and J). To further confirm gene disruption, we performed PCR across the disrupted region of the coYki gene in WT and mutant cells. In mutant cells, an amplicon of the size expected for the WT allele was absent, whereas amplicons of the size expected for both antibiotic markers were present (Figure 1—figure supplement 3A and B). Sanger sequencing of these amplicons confirmed biallelic integration of antibiotic markers at the coYki locus (Figure 1—figure supplement 3C). These results demonstrate the generation of a homozygous coYki disruption mutant (subsequently coYki -/-) and establish a method for the generation of mutants by gene targeting in Capsaspora.

To determine the effect of coYki on cell proliferation, we examined the proliferation of Capsaspora cells in adherent or shaking culture. In contrast to what may be expected based on previous studies in animal cells, we observed no significant difference in cell proliferation between WT and coYki -/- cells in either condition (Figure 2—figure supplement 1). We next sought to examine cell proliferation within Capsaspora aggregates. Previously, aggregates have been generated by gentle shaking of cell cultures (Sebé-Pedrós et al., 2013). To make aggregates more amenable to extended imaging in situ, we developed a condition that induced robust aggregate formation by culturing Capsaspora cells without shaking in low-adherence hydrogel-coated wells (Figure 1E–G). Under this condition, most cells coalesced into aggregates over 2–3 days, leaving few isolated cells in the culture. Treatment with the calcium chelator EGTA resulted in rapid dissociation of the aggregates (Figure 1—figure supplement 1), suggesting that calcium-dependent cell adhesion mediates aggregate integrity. We used EdU incorporation to label proliferating cells in aggregates and observed no significant difference in the percent of EdU-positive cells between WT and coYki -/- genotypes (Figure 2A and B). Unlike clones of cultured mammalian cells that often display increased cell proliferation at the clonal boundary (Fisher and Yeh, 1967), distribution of EdU+ cells was homogeneous within Capsaspora aggregates (Figure 2A), suggesting that a mechanism akin to contact-mediated inhibition of proliferation is absent in Capsaspora. Together, these results suggest that the rate of proliferation of Capsaspora is independent of coYki or contact inhibition.

While examining Capsaspora aggregates, we observed a profound phenotype in the shape of the coYki -/- aggregates. After aggregate induction by plating cells in low-adherence conditions, WT cell aggregates showed a round morphology (Figure 2C). In contrast, coYki -/- aggregates were asymmetric and less circular than WT aggregates (Figure 2D). Computational analysis of aggregate morphology and size revealed that WT and coYki -/- aggregate size was not significantly different (Figure 2E), whereas coYki -/- aggregates were significantly less circular than WT aggregates (Figure 2F). Optical sectioning of aggregates expressing mScarlet showed that WT aggregates were thicker perpendicular to the culture surface and more spherical than coYki -/- aggregates (Figure 2G). To quantify the difference in 3D morphology between aggregates, we made optical sections of aggregates and then quantified aggregate curvature relative to the culture substrate using Kappa (Mary and Brouhard, 2019). coYki -/- aggregates showed significantly less curvature than WT aggregates (Figure 2—figure supplement 2), indicating that coYki affects the 3D morphology of aggregates. Time-lapse microscopy of aggregation over a period of 6 days showed that, whereas WT cells accreted into round aggregates that fused with other aggregates when in close proximity but never underwent fission, coYki -/- aggregates were dynamically asymmetrical, less circular than WT, and sometimes underwent fission (Video 1). Taken together, these findings uncover a critical role for coYki in three-dimensional (3D) shape of aggregates, but not cell proliferation, in Capsaspora.

Cell–cell adhesion and cell–substrate adhesion affect 3D shape of tumor cell aggregates (Blandin et al., 2016; Saias et al., 2015). We therefore tested whether cell–cell adhesion and cell–substrate adhesion differ in WT and coYki -/- cells. A prediction of differential cell–cell adhesion is that mosaic aggregates of WT and coYki -/- cells may show cell sorting within aggregates (Foty and Steinberg, 2005). To test this prediction, we induced aggregate formation after mixing mCherry-labeled WT or coYki -/- cells with unlabeled WT cells at 1:9 ratio. The distribution of mCherry-labeled WT or coYki -/- cells within such mixed aggregates was similar and no cell sorting was evident (Figure 3A–D), indicating that individual coYki -/- cells are competent to adhere to and integrate within aggregates and suggesting that the aberrant morphology of coYki -/- aggregates is an emergent property of a large number of coYki -/- cells.

To further test whether cell–cell adhesion differs in WT and coYki -/- cells, we vigorously vortexed Capsaspora cultures to separate individual cells, allowed cell clusters to form by cell–cell adhesion under gentle rotation, and then counted the number of cells in each cluster. The number of cells per cluster was similar for WT and coYki -/- cells (Figure 3E), suggesting that the occurrence of cell–cell adhesion is similar in these genotypes. We next tested whether cell–substrate adhesion differed in WT and coYki -/- cells by agitating monolayer cultures of adherent cells and counting the number of cells that disassociate from the substrate. After agitation, significantly more coYki -/- cells remained attached to the substrate compared to WT cells (Figure 3F), indicating that coYki negatively regulates cell–substrate adhesion. To test whether aggregates of coYki -/- cells similarly show increased adherence to a substrate, we generated cell aggregates in 24-well plates, gently pipetted all nonadherent aggregates into a new well, and then counted the number of transferred (nonadherent) aggregates. Significantly more WT aggregates were nonadherent compared to coYki -/- aggregates (Figure 3—figure supplement 1), indicating that coYki negatively regulates aggregate–substrate adhesion. Together, our data suggest that coYki affects the morphology of cell aggregates by affecting cell–substrate adhesion but not cell–cell adhesion.

To characterize properties of coYki -/- cells that may contribute to abnormal aggregate morphology, we used time-lapse microscopy to examine the dynamic morphology of individual WT and coYki -/- cells in adherent culture. Interestingly, the cortices of coYki -/- cells were much more dynamic than those of WT cells. Whereas coYki -/- cells display extensive membrane protrusions and retractions, such dynamic membrane structures were rarely observed in WT cells (Video 2). To characterize the differences in cell membrane dynamics, we quantified the occurrence of cell protrusions in WT cells, coYki -/- cells, and coYki -/- cells expressing a coYki rescue transgene. WT cells rarely showed protrusions. In contrast, coYki -/- cells showed an average of 3.1 protrusions per minute (Figure 4A), a phenotype that was rescued in coYki -/- cells expressing a coYki transgene (Figure 4A, Video 3). These results suggest that coYki affects the cortical dynamics of Capsaspora cells.

In principle, the extensive membrane protrusions made by coYki -/- cells may represent F-actin-rich pseudopodia, which drive amoeboid motility, or, alternatively, F-actin-poor blebs, which are actin-depleted extensions of plasma membrane generated by disassociation of the membrane from the cell cortex or cortical rupture (Charras and Paluch, 2008). To distinguish between these possibilities, we stably expressed a fusion protein of the F-actin-binding LifeAct peptide (Riedl et al., 2008) with mScarlet in WT and coYki -/- cells and examined F-actin dynamics by time-lapse microscopy. At the basal side of WT cells, LifeAct-mScarlet was often enriched at the cell edge corresponding to the direction of cell movement, suggesting that Capsaspora undergoes F-actin-mediated amoeboid locomotion (Video 4). As cells move, some LifeAct-mScarlet signal appears as stationary foci on the basal surface, suggesting the existence of actin-rich structures that mediate cell–substrate interaction in Capsaspora (Figure 4B, Video 4). Optical sections at mid-height of WT cells showed that F-actin was enriched at a circular cell cortex, where transient bursts of small F-actin puncta were often observed (Figure 4B, Video 5). In contrast, cortical shape was less circular in coYki -/- cells, and F-actin concentration along the cortex was less uniform (Figure 4B). During the formation of the membrane protrusions produced by coYki -/- cells, F-actin was initially depleted compared to the rest of the cell (Figure 4C, Video 6). After the membrane protrusions had fully extended, F-actin accumulated first at the distal end and then along the entire cortex of the protrusion, followed by the regression of the protrusion. A similar sequence of cytoskeletal events occurs in blebbing mammalian cells (Charras et al., 2006). These results suggest that the protrusions observed in coYki -/- cells are actin-depleted blebs, and that after bleb formation, F-actin reaccumulates within the bleb, leading to cortical reformation and bleb regression.

To determine whether the abnormal cortical cytoskeleton is specific to coYki -/- cells in adherent culture, we examined coYki -/- cells within aggregates by time-lapse microscopy. LifeAct-mScarlet-expressing cells were mixed with cells stably expressing the green fluorescent protein Venus so that LifeAct-mScarlet-expressing cells were well-spaced and able to be imaged individually. As in adherent cells, coYki -/- cells within aggregates showed dynamic formation of actin-depleted blebs (Video 7). Thus, coYki -/- cells show aberrant cortical cytoskeleton and excessive membrane blebbing in both isolated adherent cells and within aggregates.

Loss of coYki causes both bleb-like protrusions and abnormal aggregate morphology, suggesting that the cellular phenotypes observed in coYki -/- cells (membrane blebbing and related cortical cytoskeletal defects) may underlie the aberrant aggregate morphology. We tested this hypothesis by treating coYki -/- cells with blebbistatin, a myosin II inhibitor named after its bleb-inhibiting activity in mammalian cells (Straight et al., 2003). Strikingly, not only did low concentrations of blebbistatin suppress the formation of blebs in coYki -/- cells (Figure 4D, Video 8), blebbistatin also rescued the abnormal morphology of coYki -/- aggregates (Figure 4E). Whereas coYki -/- cell aggregates were asymmetric and showed low circularity (Figure 4E’’’), coYki -/- aggregates treated with blebbistatin were circular and resembled WT aggregates (Figure 4E’’’’ and F). These results suggest a causal link between the membrane/cytoskeletal defects observed in individual coYki -/- cells and the abnormal morphology of coYki -/- aggregates.

In mammals and in Drosophila, YAP/TAZ/Yorkie is regulated by cytoplasmic sequestration in response to phosphorylation of HXRXXS motifs by the LATS/Warts kinases (Zheng and Pan, 2019). coYki has three HXRXXS motifs that align with HXRXXS motifs on YAP and an additional nonconserved N-terminal HXRXXS motif (Sebé-Pedrós et al., 2012). To test whether these motifs affect coYki localization in Capsaspora, we transiently expressed fusions of mScarlet with coYki or coYki with the four HXRXXS motifs mutated to be non-phosphorylatable (4SA) along with a histone H2B-Venus fusion protein, which marks the nucleus (Parra-Acero et al., 2018). Cells transfected with mScarlet-coYki showed mScarlet enriched in the cytoplasm relative to the nucleus (Figure 5B and E). In contrast, the majority of cells transfected with mScarlet-coYki 4SA showed uniform mScarlet signal throughout the cell or enriched mScarlet signal in the nucleus (Figure 5C and F). To quantitatively characterize the subcellular localization of coYki, we examined the mScarlet fluorescence intensity of line traces across the nuclear–cytoplasmic boundary for cells expressing mScarlet-coYki fusion proteins. Whereas mScarlet fluorescence was reduced in the nucleus relative to the cytoplasm in cells expressing mScarlet-coYki (Figure 5G), fluorescence intensity was similar in the nucleus and cytoplasm in cells expressing mScarlet-coYki 4SA (Figure 5H). These results show that HXRXXS motifs affect the subcellular localization of coYki and suggest that, like in animals, phosphorylation of coYki by Hippo signaling leads to cytoplasmic sequestration. This regulation of coYki nuclear localization, along with the previous finding that coYki can induce the expression of Hippo pathway genes when expressed in Drosophila (Sebé-Pedrós et al., 2012), suggests that the function of coYki as a transcriptional regulator and Hippo pathway effector is conserved between Capsaspora and animals.

To characterize the transcriptional targets of coYki that may underlie the morphological/cytoskeletal defects of coYki -/- cells, we performed RNAseq on WT and coYki -/- cells in adherent culture. This analysis revealed 1205 differentially expressed genes, including 397 downregulated (Supplementary file 1) and 808 upregulated genes (Supplementary file 2) in the coYki mutant. Functional enrichment analysis of these two gene sets revealed distinct enrichment of functional categories (Figure 6A and B). Genes predicted to encode actin-binding proteins were enriched in the set of genes upregulated in the coYki -/- mutant (Supplementary file 3), suggesting that these genes may play a role in the aberrant cytoskeletal dynamics observed in coYki -/- cells. Nevertheless, we cannot formally exclude the possibility that coYki may also regulate cytoskeletal dynamics in a transcription-independent manner, an example of which has been reported in Drosophila (Xu et al., 2018).

To further inquire into the potential function of coYki-regulated genes, we searched the 1205 differentially expressed genes against a previously reported Capsaspora phylome (Sebé-Pedrós et al., 2016) to identify 638 Capsaspora genes with predicted human or mouse orthologs (Supplementary file 4). Ingenuity Pathway Analysis of this mammalian ortholog set showed that the two most significantly enriched functional categories corresponded to cell movement and cell migration (Figure 6C). Notably, regulation of cell migration was reported as the most enriched functional category among YAP target genes in glioblastoma cells (Stein et al., 2015), and cell motility was an enriched category in an integrative analysis of gene regulatory networks downstream of YAP/TAZ utilizing transcriptomic and cistromic data from multiple human tissues (Paczkowska et al., 2020). However, in contrast to functional enrichment studies in metazoan systems examining genes regulated by YAP/TAZ/Yorkie (Ikmi et al., 2014; Stein et al., 2015; Zanconato et al., 2015), no enrichment was detected in functional categories of cell proliferation or the cell cycle (Figure 6C). This finding is consistent with the lack of a proliferative phenotype in coYki -/- cells.

Capsaspora cells were obtained from ATCC (ATCC strain number 30864). Cells were maintained in ATCC Medium 1034 (modified PYNFH medium) supplemented with 25 μg/ml ampicillin (referred to below as ‘growth medium’) in a 23°C incubator. Cells were kept in either 25 cm² cell culture flasks in 8 ml medium or in 75 cm² cell culture flasks in 15 ml medium. To induce cell aggregation, cells were resuspended to 7.5 × 10⁵ cells/ml in growth medium, and 1 ml of this cell suspension was added per well to a 24-well ultra-low attachment plate (Sigma CLS3473). To test the effect of blebbistatin on aggregate morphology, a 20 mM stock solution of blebbistatin (Sigma B0560) in DMSO was made, and cells were resuspended to 7.5 × 10⁵ cells/ml in growth medium with 1 μm blebbistatin or DMSO only as a vehicle control. Aggregates were then generated as described above and imaged with a Nikon Eclipse Ti inverted microscope with NIS-Elements acquisition software.

To characterize cell proliferation during growth on a solid substrate, Capsaspora cells were diluted to 1 × 10⁵ cells/ml in growth medium, and 800 μl of this dilution was pipetted into multiple wells in a 24-well polystyrene cell culture plate. To prevent edge effects, wells at the edge of the plate were not used and were instead filled with 800 μl of water. The plates were then kept in a 23°C incubator. Each day, 1.5 μl of a 500 mM EDTA solution was added to a well, and cells in the well were resuspended by pipetting up and down for 1 min. Resuspended cells were then transferred to a 1.5 ml tube, vortexed, and then the cell density was determined by hemocytometer. To characterize cell proliferation in shaking culture, cells were diluted to 1 × 10⁵ cells/ml in growth medium, and 20 ml of this dilution was added to a 125 ml Erlenmeyer culture flask. Cultures were then incubated in an orbital shaker at 150 rpm and 23°C. Cell density was determined daily by transferring 200 μl of culture to a 1.5 ml tube, adding 1.5 μl of 500 mM EDTA solution to disassociate any aggregated cells, vortexing, and counting by hemocytometer.

To fix and stain aggregates, aggregates were generated as described above, and then aggregates were collected from a 24-well plate by gentle pipetting. 500 μl of aggregate suspension was added dropwise to 9 ml of PM PFA (100 mM PIPES pH 6.9, 0.1 mM MgSO₄, 4% PFA filtered through a 0.45 μm syringe filter) in a 15 ml polystyrene tube, and tubes were left undisturbed for 1 hr to allow fixation. Aggregates were then centrifuged at 1000 × g for 3 min and resuspended in 500 μl of PEM buffer (100 mM PIPES pH 6.9, 1 mM EGTA, 0.1 mM MgSO₄). After transferring aggregates to a microcentrifuge tube, aggregates were centrifuged for 2000 × g for 2 min, the supernatant was removed, and aggregates were resuspended in 300 μl of PEM.

As we found that aggregates were fragile during continued centrifugation-resuspension cycles, we developed a protocol where solution changes could be done without centrifugation using 24-well plate inserts with an 8 µm pore membrane at the insert base (Corning 3422). Before use with aggregates, 100 μl PEM buffer was added to a well in a 24-well plate, 300 μl of PEM was added to a membrane insert, and the insert was placed in the well and allowed to drain (contact between the membrane and the solution within the well is critical for draining). Aggregates in 300 μl of PEM were added to the insert, and the insert was moved to a new well with 100 μl of PEM and allowed to drain. As a wash step, the membrane insert was then moved to a well containing 1200 μl of PEM, allowing the membrane insert to fill with PEM buffer, and after a 5 min incubation, the membrane was moved to a new well with 100 μl of PEM and allowed to drain. The membrane insert was then moved to a well containing 1200 μl of PEM block/permeabilization solution (100 mM PIPES pH 6.9, 1 mM EGTA, 0.1 mM MgSO₄, 1% BSA, 0.3% Triton X-100) and incubated for 30 min at room temperature. The membrane insert was moved to a well with 100 μl of PEM and allowed to drain and then moved into a well containing 1 ml PEM with DAPI (2 μg/ml) and phalloidin (0.1 μM) and incubated for 15 min at room temperature. Aggregates were then washed once more in 1200 μl PEM. To mount aggregates for imaging, the membrane insert was placed in a new well with 100 μl of PEM buffer and was allowed to drain, aggregates were resuspended in the residual buffer (approximately 50 μl) by gentle pipetting, and then aggregates were pipetted onto a microscope slide and allowed to settle for 5 min. Kimwipes were then placed in contact with the edge of the liquid on the slide, allowing for removal of excess buffer by capillary action. 15 μl of Fluoromount G (SouthernBiotech) was pipetted onto the aggregates, which were then covered with a 22 × 40 mm cover glass and sealed with nail polish. Aggregates were subsequently imaged with a Zeiss LSM 880 confocal microscope. Staining of adherent cells with phalloidin was done as previously described (Parra-Acero et al., 2020).

To examine cell proliferation within aggregates by EdU labeling, aggregates were generated in ultra-low adherence 24-well plates as described above. 3 days after aggregation was initiated, 100 μl of growth medium containing EdU was added to aggregates to result in a 150 μM EdU concentration. Aggregates were incubated with EdU for 4 hr and then fixed and resuspended in 300 μl PEM buffer as described above. Aggregates were then processed for imaging following the manufacturer’s protocol (Thermo Fisher C10339) using the 24-well plate membrane insert method described above to transfer aggregates between solutions. For all steps, wells with 1 ml of solution were used for washing or staining cells in membrane inserts, and wells with 100 μl of solution were used to drain wells in preparation for transfer into the next solution.

For the sequences of synthesized DNA fragments used in this study, see Supplementary file 5. For a list of plasmids and oligonucleotides generated for this study, see Key resources table. To generate a vector for mScarlet expression in Capsaspora (pJP71), a DNA construct containing an open-reading frame encoding the mScarlet protein (Bindels et al., 2017) codon-optimized for expression in Capsaspora and under control of the previously described EF1-α promoter and terminator from the pONSY-mCherry vector (Parra-Acero et al., 2018) was constructed by gene synthesis (sJP1) and cloned into the EcoRV site of the pUC57-mini vector (GenScript). A KpnI site and an AflII site were incorporated at the beginning and end of the coding sequence, respectively, allowing for the construction of N or C-terminal protein fusion constructs by Gibson assembly. To generate a plasmid for expression of mScarlet and a Geneticin resistance gene (pJP72), the NeoR coding sequence from the Dictyostelium pDM323 vector (Veltman et al., 2009) was codon-optimized for Capsaspora, the optimized sequence was synthesized with the promoter and terminator from the Capsaspora actin ortholog gene CAOG_06018, and the synthesized fragment (sJP2) was cloned into the SmaI site of pJP71 by Gibson assembly. To generate plasmids expressing mScarlet and either nourseothricin resistance (pJP102) or hygromycin resistance (pJP103), an identical strategy to that used in constructing pJP72 was used, except either a synthesized codon-optimized nourseothricin acetyltransferase gene (sJP3) based on the sequence of pUC18T-mini-Tn7T-nat (Lehman et al., 2016) or a synthesized codon-optimized hygromycin B phosphotransferase gene (sJP4) based on the sequence of Dictyostelium vector pDM358 (Veltman et al., 2009) was used.

To construct a plasmid for expression of a mScarlet-coYki fusion protein in Capsaspora (pJP80), we synthesized a gene (synthesized gene fragment sJP5) encoding the coYki protein (GenBank accession number JN202490.1) and cloned this gene into an AvrII digest of pJP72 by Gibson assembly. To limit recombination of the transgene with the endogenous coYki loci, we recoded the WT coYki reading frame using synonymous codon replacement, resulting in a coYki gene with 73% nucleotide similarity to the WT coYki gene but encoding an identical polypeptide. To generate a plasmid expressing a mutant mScarlet-coYki protein lacking four predicted phosphorylation sites located in HXRXXS motifs (pJP90), pJP80 was mutated by site-directed mutagenesis so that the codons encoding serine residues corresponding to WT coYki residues 90, 152, 184, and 305 were changed to alanine.

To generate a plasmid encoding coYki and hygromycin resistance for rescue of coYki -/- phenotypes (pJP119), a DNA fragment encoding coYki and a c-terminal OLLAS epitope tag (Park et al., 2008) was generated by PCR from pJP80 using primers oJP101 and oJP102. This DNA fragment was then cloned into a KpnI and AflII digest of pJP103 using Gibson assembly.

To generate a plasmid encoding a LifeAct-mScarlet fusion protein and hygromycin resistance (pJP118), a DNA fragment optimized for expression in Capsaspora encoding the LifeAct peptide (Riedl et al., 2008) (sJP6) was synthesized and cloned into the KpnI site of pJP103 by Gibson assembly. To generate a plasmid expressing Venus and Geneticin resistance (pJP114), the Venus reading frame was amplified from p_EF1α -CoH2B:Venus (Addgene #111877, Parra-Acero et al., 2018) using primers oJP103 and oJP104 and cloned into a KpnI and AflII digest of pJP72 by Gibson assembly.

24-well plates were prepared for transfection by inserting one sterile 12 mm circular glass coverslip in each well to aid in cell adhesion during transfection. Capsaspora cells at exponential growth phase were collected by pipetting medium over attached cells, resuspended to 7.5 × 10⁵ cells/ml in growth medium, and then 800 μl of this cell culture was added per well to the prepared 24-well plate. Cells were incubated at 23°C overnight. The following day growth medium was removed from cells and was replaced by 800 μl of transfection medium (Scheider’s Drosophila Medium [Thermo Fisher 21720024] with 10% FBS [Thermo Fisher 26140079] supplemented with 25 μg/ml ampicillin). After a 10 min incubation at room temperature, transfection medium was removed from the cells and replaced with 500 of fresh transfection medium. Transfection mixes were then prepared: 100 μl of Opti-MEM I Reduced Serum Medium (Thermo Fisher) was added to a 1.5 ml tube, and 1 μg of transfecting DNA was added to this medium (for multiple plasmids, an equivalent amount of each plasmid by mass was used). 3 μl of TransIT-X2 transfection reagent (Mirus Bio) was then added to the tube, and the solution was immediately mixed by pipetting up and down. Transfection mixes were incubated at room temperature for 5 min, and then 70 μl of the transfection mix was added dropwise to one well in the 24-well plate. Cells were incubated at 23°C for 24 hr, and then the medium was removed and replaced with 800 μl of growth medium. To image fluorescent transient transfectants, cells were then resuspended by pipetting up and down, and 300 μl of resuspended cells in growth medium were transferred to a well of an 8-well chambered coverglass slide (Nunc). Cells were then imaged 48 hr after addition of growth medium using a Zeiss LSM 880 confocal microscope. Labeling of the nucleus of transient transfectants was done by cotransfection of cells with pONSY-CoH2B:Venus, which encodes the Venus fluorescent protein fused to the Capsaspora histone H2B protein (Addgene #111877, Parra-Acero et al., 2018).

To examine mScarlet-coYki fluorescence intensity in different subcellular regions, Capsaspora cells were transiently transfected with plasmids encoding histone H2B-Venus and either mScarlet-coYki or mScarlet-coYki 4SA and imaged as described above. Using ImageJ, a line 2.03 µm in length, corresponding to 25 pixels in the image, was drawn across the nuclear–cytoplasmic boundary as determined by the H2B-Venus signal. The line was drawn perpendicular to the nuclear boundary, with pixels 1–12 within the nucleus and pixels 14–25 in the cytoplasm. The ‘Plot Profile’ function was used to plot the mScarlet fluorescence intensity along the line. To test whether fluorescence in the nuclear vs. cytoplasmic region was significantly different, mean fluorescence intensity values for each pixel position along the line were determined for each independent experiment. Then, an average fluorescence intensity value for all pixels in the nucleus (pixels 1–12) and all pixels in the cytoplasm (pixels 14–25) was calculated for each independent experiment. Average nuclear and cytoplasmic fluorescence intensity values for thre independent experiments were then analyzed by paired t-test.

Transforming plasmids were linearized by digestion with either ScaI-HF or AseI restriction enzymes (NEB), which cut within the ampicillin resistance gene, and then purified from solution and resuspended in nuclease-free water. Capsaspora cells were transfected with the linearized plasmids using the TransIT-X2 transfection reagent as described above. Two days after growth medium was added to cells following transfection, the growth medium was removed and replaced with 800 μl of growth medium supplemented with selective drugs. Due to observations of variability in cell viability between transfections, three drug concentrations were tested in parallel for each transfection: for Geneticin (Thermo Fisher), cells were treated at 40, 60, or 80 μg/ml; for nourseothricin (GoldBio), cells were treated at 50, 75, or 100 μg/ml; for hygromycin B (Sigma), cells were treated at 150, 200, or 250 μg/ml. Cells were grown in selective medium for 2 weeks, and medium was changed every 3 days. Following selection, clonal cell populations were generated by resuspending cells in growth medium by pipetting, diluting cells to 3 cells/ml in growth medium, and adding 100 μl of this dilution to 200 μl of growth medium per well in a 48-well plate (this procedure generated approximately 10 wells with cell growth for each 48-well plate). To image fluorescence in stably transfected lines, clonal cell populations were transferred into wells in a four-well glass bottom chamber slide (Nunc), incubated at 23°C for 24 hr, and cells were then imaged using a Nikon Eclipse Ti inverted microscope with NIS-Elements acquisition software.

To disrupt coYki (CAOG_07866), we attempted to delete a 228 bp segment of the coYki open-reading frame that encodes the predicted TBD (Sebé-Pedrós et al., 2012) using a PCR-generated gene targeting construct. We designed a pair of primers that amplify the drug resistance markers described above, including the actin (CAOG_06018) promoter and terminator, with 90 bp of homology adjacent to the sequence targeted for deletion at the 5′ end of the primers (oJP105 and oJP106). These primers were used to amplify the Geneticin resistance cassette from pJP72 by PCR, the resulting PCR product was gel purified and resuspended in nuclease-free water, and WT Capsaspora cells were transfected with this DNA as described above. Clonal populations of drug-resistant transformants were generated as described above, each clone was grown to confluency in one well in a 24-well plate, cells were collected by pipetting up and down, and genomic DNA was prepared following the protocol described below. To genotype potential mutants, we performed diagnostic PCRs on genomic DNA: to test for presence of the WT allele, we used a forward primer with homology to the genome 5′ of the coYki sequence targeted for deletion (oJP107) and a reverse primer with homology within the sequence targeted for deletion (oJP108). To detect successful deletion, we used oJP107 as a forward primer and a reverse primer with homology to the geneticin resistance gene (oJP109). 40% of analyzed clones (6 of 15 tested clones) showed a PCR product indicative of disruption of the coYki allele (Figure 1—figure supplement 2D). However, all clones showed a band indicative of an intact WT coYki gene. We therefore reasoned that, at least for the culture conditions used during transfection, Capsaspora cells may be diploid.

After the isolation of clonal lines that were heterozygous for coYki disruption with the Geneticin resistance marker as indicated by diagnostic PCR, we attempted to disrupt the remaining coYki allele using nourseothricin resistance as a marker. pJP102 was used as a template for PCR using the same primers that were used to generate the Geneticin resistance construct (oJP105 and oJP106) to generate a nourseothricin resistance cassette flanked by homology arms targeting coYki for disruption. This construct was used to transfect a Geneticin-resistant heterozygous coYki disruption mutant, and drug selection of transfectants was done as described above using simultaneous selection with 60 μg/ml Geneticin and 75 μg/ml nourseothricin. After generating clonal populations of drug-resistant transformants, diagnostic PCR was done to detect the intact coYki allele as described above or a deletion allele containing the nourseothricin resistance gene using primers oJP107 and oJP110. 27% of analyzed clones (3 of 11 clones) showed a PCR product indicative of disruption of the coYki locus with the nourseothricin resistance marker. Two independent clonal cell lines that showed absence of the WT coYki allele and presence of both disruption alleles by diagnostic PCR were then sequenced by performing PCR across the deleted region of coYki using primers oJP115 and oJP116. For both mutant clones, this PCR generated two bands, which were of the sizes expected for disruption of coYki with the nourseothricin resistance cassette and disruption of coYki with the Geneticin resistance cassette. Sanger sequencing confirmed that the sequence of these bands matched the expected sequences (Figure 1—figure supplement 3), confirming biallelic disruption of coYki. These two coYki -/- clones were used interchangeably for subsequent phenotypic studies. For all reported phenotypes, the phenotype was observed in both of these independent clonal coYki -/- cell lines.

To quantify coYki gene expression in a putative homozygous coYki disruption mutant (coYki-/-) by qPCR, 5 × 10⁶ WT or coYki-/- cells were collected from culture flasks while in growth medium by pipetting up and down, and RNA was isolated using an RNeasy Mini Plus kit (QIAGEN). cDNA was made using the iScript cDNA Synthesis Kit (Bio-Rad), qPCR reactions were made using iQ SRBR Green Supermix (Bio-Rad), and qPCR was performed using a CFX96 Touch Real-time PCR detection system (Bio-Rad) with the following primers: oJP111 and oJP112 to detect GAPDH, oJP113 and oJP114 to detect a region of coYki within the sequence targeted for deletion.

Capsaspora genomic DNA was prepared for PCR analysis following a procedure previously developed for Dictyostelium discoideum (Charette and Cosson, 2004). Cells grown in a 24-well or 48-well plate in growth medium were collected by pipetting up and down, pipetted into a 1.5 ml tube, centrifuged at 4000 × g for 5 min, and resuspended in 20 μl of nuclease-free water. 20 μl of Lysis buffer (50 mM KCl, 10 mM TRIS pH 8.3, 2.5 mM MgCl₂, 0.45% NP40, 0.45% Tween 20, and 800 μg/ml Proteinase K added fresh from a 20 mg/ml stock) was added to cells, which were then incubated at room temperature for 5 min. Tubes were then placed at 95°C for 5 min. After cooling to room temperature, 1 μl of this sample was then used as a template in a 20 μl diagnostic PCR.

ImageJ was used to quantify aggregate size and circularity using aggregate images. Images were converted to 8-bit format, processed twice using the ‘Smooth’ function, and a Threshold was adjusted for each individual image. The Analyze Particles command was then run with a gate for particle size at 400-infinity µm² and the ‘exclude on edges’ option selected. The returned values for aggregate area and circularity were then used for further analysis. Values corresponding to more than one adjacent aggregate interpreted by the algorithm as a single particle were discarded.

mScarlet-expressing aggregates in low-adherence 24-well plates were imaged with a Zeiss LSM 880 confocal microscope using a ×10 objective. Z-stack imaging of aggregates was done using a 7.7 µm interval between sections. This method allowed visualization of the 3D structure of the side of the aggregate closest to the objective. Orthogonal views of the aggregate through the approximate midline of the aggregate were generated using ImageJ. To measure the average curvature of aggregates, an open B-spline curve was fit to the aggregate border in an orthogonal view of the aggregate using Kappa (Mary and Brouhard, 2019). An initial curve segment was manually defined on the side of the aggregate closest to the imaging objective by defining points along the aggregate border, and an open B-spline curve was fit to these points. An average curvature for each aggregate was calculated in Kappa. Curvature is given as 1/R, where R is the radius of curvature for a given point on the curve.

To examine cell–cell adhesion, cells were collected from a culture flask by pipetting growth media over the culture surface, diluted to 1 × 10⁶ cells/ml in growth medium, and vortexed to disrupt cell clusters. 1 ml volumes of cell culture were then transferred to 1.5 ml tubes and incubated on a Labquake rotator (Thermo) for 1 hr at room temperature. Cultures were then examined by hemocytometer, and the number of cells in each cluster of cells was counted for at least 35 cell clusters for each independent experiment.

To examine cell–substrate adhesion, cells were diluted to 5 × 10⁵ cells/ml in growth medium, and 3 ml volumes of culture were transferred per well to six-well tissue culture-treated polystyrene plates (Sigma CLS3506) and incubated at 23°C for 48 hr. For agitation, plates were then placed on an orbital shaker and shaken at 140 RPM for 10 min, and then the medium from each well (‘disassociated fraction’) was transferred to a separate tube and 3 ml of new growth medium was added to each well. 3 μl of 500 mM EDTA was then added to each well to detach cells from the culture surface, medium was pipetted up and down over the culture surface 20 times, and this resuspension (‘adherent fraction’) was transferred to a separate tube. Cell densities of the collected fractions were determined by hemocytometer counts, total cell amounts were calculated by adding the cell numbers for disassociated and adherent fractions for each condition, and the percent of adherent cells for each condition was then calculated. As a control to examine cell–substrate adhesion in the absence of agitation, the above protocol was followed, except the orbital shaker agitation step was omitted.

To image cells by time-lapse transmitted light or fluorescence microscopy, cells were resuspended to 1 × 10⁸ cells/ml in growth medium, and 10 μl of this cell suspension was added as a spot in the center of a well in a four-well chambered coverglass slide (Nunc). After a 1 hr incubation at room temperature to allow the cells to settle, the 10 μl volume of medium was removed, and 1 ml of growth medium was added to the well, resulting in a spot of cells in the center of the well. Cells were incubated at room temperature for 24 hr, a field of view with well-spaced cells was located, and then cells were imaged by time-lapse microscopy using a Nikon Eclipse Ti inverted microscope with NIS-Elements acquisition software. To image the effect of blebbistatin on individual cells, similar imaging of cells was done, except that cells were treated with either DMSO or 1 μm bebbistatin (Sigma B0560) for 1 hr before imaging. To image aggregates by time-lapse microscopy, aggregates were generated in ultra-low attachment plates as described above. During aggregate formation, four-well chambered coverglass slides (Nunc) were coated with UltraPure agarose (Thermo Fisher) by making a 1% agarose in nuclease-free water mixture, microwaving to dissolve the agarose, adding 800 μl of molten agarose per well, and removing the molten agarose after 10 s by aspiration. Coated wells were then allowed to dry at room temperature in a cell culture hood, resulting in a thin coating of agarose. This coating functions to prevent aggregate adhesion to the glass surface. After aggregate formation, aggregates were resuspended in the well by gentle pipetting, and then an 800 μl volume of resuspended aggregates was transferred into one agarose-coated well in a four-well chambered coverglass slide. Aggregates were then imaged using a Zeiss LSM 880 confocal microscope.

To perform RNA-seq on Capsaspora cells, WT or coYki -/- cells were collected from growth flasks by removing growth medium, adding fresh growth medium, and resuspending cells attached to the flask by pipetting the medium over the surface. Cells were then diluted to 2 × 10⁵ cells/ml, and for each genotype two 75 cm² culture flasks were prepared by adding 16 ml of culture dilution to each flask. Culture flasks were incubated at 23°C for 2 days, and then cells were collected by pipetting growth medium over attached cells. Cells were collected by centrifugation at 2100 × g for 5 min, and then all cells for each genotype were combined, resuspended in 500 μl of growth medium, and transferred to a microcentrifuge tube. RNA was then prepared with the RNeasy Plus Mini Kit (QIAGEN) using the QIAshredder spin column (QIAGEN) to homogenize the lysate. RNA samples were run on an Agilent Tapestation 4200 to determine level of degradation, thus ensuring only high-quality RNA was used (RIN Score 8 or higher). A Qubit 4.0 Fluorimeter (Thermo Fisher) was used to determine the concentration prior to staring library prep. 1 µg of total DNAse-treated RNA was then prepared with the TruSeq Stranded mRNA Library Prep Kit (Illumina). Poly-A RNA was purified and fragmented before strand-specific cDNA synthesis. cDNA was then A-tailed and indexed adapters were ligated. After adapter ligation, samples were PCR-amplified and purified with AmpureXP beads, then validated again on the Agilent Tapestation 4200. Before being normalized and pooled, samples were quantified by Qubit then run on an Illumina NextSeq 500 using V2.5 reagents. Three biological replicates were sequenced for each genotype, with 22–31 million reads generated per sample. The FastQ files were checked for quality using FastQC (v0.11.2) (Andrews, 2010) and fastq_screen (v0.4.4) (Wingett, 2011). FastQ files were mapped to GCF_000151315.2 reference assembly using STAR (Dobin et al., 2013). Read counts were then generated using featureCounts (Liao et al., 2014). Trimmed mean of M-values normalization and differential expression analysis were performed using edgeR (Robinson et al., 2010) (false discovery rate [FDR] ≤ 0.05, absolute log2(fold change) ≥ 0.5, log(counts per million) ≥ 0). Phylogenetic trees were downloaded from PhyloDB (Huerta-Cepas et al., 2014) corresponding to Capsaspora (PhyID 101) and pairwise distances for each gene were extracted using an R package ape (Paradis and Schliep, 2019). Closest human and mouse orthologs were then extracted and used to annotate the Capsaspora genes. These annotated gene names were then used with IPA (Krämer et al., 2014) (QIAGEN Inc, https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis) to get significantly enriched pathways in ‘Diseases and Bio Functions’ category. To identify functional enrichment in the sets of genes upregulated or downregulated in coYki -/- cells, DAVID Functional Annotation Tool (Huang et al., 2009) was used to identify functional categories with FDR < 0.05 using all Capsaspora genome genes as the gene population background.

All statistics were done using Prism (GraphPad Software, San Diego, CA). Student’s t-test was done to compare differences between two groups, and one-way analysis of variance (ANOVA) with a post-hoc test was used to compare differences among groups greater than two. For all experiments where multiple cells or aggregates were measured for each biological replicate, statistics were done by calculating the mean value for all cells or aggregates for each biological replicate and performing statistics on these sample-level means, following Lord et al., 2020.

In the article, we use italicized text to refer to a gene (e.g., coYki), unitalicized text to refer to a protein (e.g., coYki), ‘-/-’ to indicate a homozygous deletion mutant (e.g., coYki -/-), and ‘>’ to indicate a promoter driving the expression of a gene (e.g., EF1 >coYki).

RNAseq data generated in this study have been deposited into the NCBI SRA (BioProject PRJNA759885). The following plasmids have been deposited at Addgene: pJP72 (#176479), pJP102 (#176480), pJP103 (#176481), and pJP118 (#176494).