Capsaspora exhibits chemotaxis and chemokinesis in response to schistosome sporocysts

The initial report of Capsaspora, isolated from B. glabrata snails, revealed that Capsaspora migrated toward schistosome sporocysts. The sporocysts only attracted Capsaspora after they were initially attacked by Capsaspora (by random contact) and presumably began to ‘leak’ their contents into the media surrounding them [15].We repeated this analysis and indeed repeatedly observed migration of Capsaspora toward schistosomes that already had other Capsaspora cells attached (Figs 1B and S1 and S1–S6 Videos available at https://doi.org/10.5967/sh2f-xt93).

Apart from the direction of movement, we asked if the velocity of Capsaspora movement was influenced by the presence of schistosome sporocysts. We discovered that Capsaspora showed increased overall cell movement (chemokinesis) when co-cultured with schistosome sporocysts (Fig 1C). Therefore, Capsaspora exhibits both chemotaxis and chemokinesis in response to leaking prey. Since chemokinesis is both less studied and easier to monitor than chemotaxis, we aimed to further investigate the chemokinesis behavior in this work.

Capsaspora exhibits chemokinesis in response to cellular components from diverse organisms

We hypothesized that chemical components that leaked from the sporocyst were inducing Capsaspora chemokinesis. To test this hypothesis, we assessed if lysate from mechanically lysed prey could increase Capsaspora motility. We were unable to obtain enough schistosome sporocysts to test their lysate. However, Capsaspora has been shown to adhere and kill Biomphalaria glabrata embryonic (Bge) cells [15]—which are readily culturable in large quantities. Therefore, we tested if Bge cell lysate could induce Capsaspora chemokinesis. Indeed, we observed increased cell movement in the presence of Bge cell lysate in a dose-dependent manner (Figs 1D and S2A). We further explored if other lysed cells induced chemokinesis. We tested lysates from both a mammalian cell line (HEK293T) and an insect cell line (JW18). Both induced chemokinesis (Figs 1E, S2B and S2C). These data together suggested that Capsaspora chemokinesis is induced by general cellular factor(s)—not a highly specific factor produced by only Capsaspora’s putative symbionts. To further probe the promiscuity of chemokinesis induction, we asked if even prokaryotic cell lysates could induce Capsaspora chemokinesis. We tested lysates of both a gram-negative bacterium (Escherichia coli) and a gram-positive bacterium (Bacillus subtilis). Both induced Capsaspora chemokinesis in a dose-dependent manner (Figs 1F, 1G, S2D and S2E). In sum, Capsaspora cell motility increased in response to soluble cell lysates from a remarkably wide range of cells across different domains of life. Therefore, Capsaspora chemokinesis is induced by common and/or diverse cellular component(s).

Bovine serum albumin is sufficient to induce chemokinesis in Capsaspora

Given that Capsaspora motility increases in response to such a range of lysates, we asked if it responded to an extracellular rich biological substance: fetal bovine serum (FBS). Remarkably, Capsaspora increased its movement in the presence of FBS in a dose-dependent manner, as well (Fig 2A). This observation further underscored the breadth of chemokinesis-inducing substances, and it afforded an abundant and stable chemokinesis-inducing material for further mechanistic studies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Bovine serum albumin is sufficient to induce Capsaspora chemokinesis.

(A) Capsaspora motility upon addition of different concentrations of FBS. The final concentration of FBS in the assay is displayed on the x axis. Error bars represent standard error of the mean of a biological triplicate of three individual wells, each with dozens of cells (n = 3). The dashed line indicates baseline motility upon addition of negative control (water). (B) Capsaspora motility upon addition of small molecule (< 30 kDa) or macromolecule (> 30 kDa) components of FBS, relative to full FBS and control (water). Error bars represent standard error of the mean of a biological triplicate of three individual wells of cells (n = 3), except for the FBS positive control, which only had two wells. Individual biological replicates are displayed with white circles. (C) Capsaspora motility upon addition of different concentrations of FBS (dark grey) or lipoprotein-deficient FBS (light grey). Error bars represent the range of a biological duplicate of three individual wells, each with dozens of cells (n = 2). Individual biological replicates are displayed with white circles. The dashed line indicates baseline motility upon addition of negative control (Chernin’s Balanced Salt Solution, CBSS [20]). (D) Capsaspora motility upon addition of heat-treated FBS, compared to untreated FBS and water (control). Error bars represent standard error of the mean of a biological triplicate of three individual wells (n = 3), each containing dozens of cells. Individual biological replicates are displayed with white circles. (E) Bioassay-guided fractionation scheme to purify chemokinesis-inducing substance(s) from FBS. Boxes represent tested fractions from each separation. Dark blue boxes are the most active fractions, which were carried forward. Light blue boxes were less active fractions. Chromatograms, chemokinesis data for individual fractions, and SDS-PAGE image of final active fractions are reported in S3 Fig. LC-MS/MS identification of major proteins is reported in S1 Dataset. (F) Capsaspora motility upon addition of different concentrations of BSA. The final concentration of BSA in the assay is displayed on x axis. Error bars represent standard error of the mean of a biological triplicate of three individual wells, each with dozens of cells (n = 3). The dashed line indicates baseline motility upon addition of negative control (water). Chemokinesis assay method A was used for panels A–C, E. Method B was used for panels D and F.

https://doi.org/10.1371/journal.ppat.1013440.g002

First, we aimed to identify a single pure component from FBS that was sufficient to induce chemokinesis. We investigated if the inducer was a small molecule or a macromolecule by separating FBS with a 30 kilodalton (kDa) molecular weight cutoff filter. We identified that only the > 30 kDa fraction induced chemokinesis—indicating that the inducer is macromolecular (Fig 2B). We previously found that serum lipoproteins trigger cellular aggregation in Capsaspora [8,9]. To determine if these lipids were also important for chemokinesis, we tested lipoprotein-deficient FBS, which contains only ~5% of the natural level of lipoproteins. We found that the depleted FBS induced chemokinesis with the same potency as full FBS, indicating that lipoproteins are not necessary for chemokinesis induction by FBS (Fig 2C). We further investigated if the active component was heat-labile. It was, suggesting the active component is a protein or heat-labile glycan (Fig 2D).

We fractionated the macromolecular material from FBS using a strong anion exchange (SAX) column (Figs 2E and S3). We collected one of the most active fractions and subjected it to further fractionation by size exclusion column chromatography (SEC) (Figs 2E and S3). Chemokinesis-inducing activity appeared to correlate with three proteins that migrated at ~ 70 kDa, ~ 60 kDa, and ~40 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Figs 2E and S3I). The primary proteins present in these gel bands were bovine serum albumin (BSA), alpha-fetoprotein (AFP), and alpha-2-HS-glycoprotein (fetuin A) (S1 Dataset). These are all major components of fetal serum. BSA is readily available as a purified protein. Therefore, we investigated its ability to induce chemokinesis as a single pure component. We found that BSA could induce chemokinesis in a dose-dependent fashion (Fig 2F). Since FBS contains approximately 25 mg/ml of BSA [19], the EC₅₀ value of FBS (~1%, Fig 2A) should contain ~0.25 mg of FBS—which is close to our observed EC₅₀ of pure BSA (~1 mg/ml, Fig 2F). Therefore, it is plausible that BSA is the primary (or sole) chemokinesis-inducing agent in FBS. It is also possible that AFP, fetuin A, and/or other proteins may also contribute in proportion to their concentration in FBS, as well.

Capsaspora chemokinesis response is not specific to bovine serum albumin

Since diverse cell lysates (that do not contain BSA) increase Capsaspora motility, non-BSA components must also be chemokinesis inducers. Among other readily available proteins, human serum albumin (HSA) and oval albumin also induced migration in Capsaspora in a dose-dependent fashion (Fig 3A and 3B). Oval albumin is especially notable, since it is not homologous to BSA. Gamma globulins also appeared to slightly induce motility at only a single concentration (Fig 3C). However, not all proteins increased Capsaspora motility—myoglobin failed to induce a significant increase in motility (Fig 3D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Other proteins increase Capsaspora motility.

Capsaspora motility upon addition of different concentrations of (A) human serum albumin (HSA), (B) oval albumin, (C) G-globulins, (D) myoglobin, (E) fatty acid-free BSA, (F) peptone, (G) glutathione, and (H) mixtures of ‘essential amino acids’ and ‘non-essential amino acids’. In all cases, the final concentration of the inducer in the cell culture is displayed on x axis. Error bars represent standard error of the mean of a biological triplicate of three individual wells, each with dozens of cells (n = 3). The dashed line indicates baseline motility upon addition of negative control (water). Chemokinesis assay method B was used for panels A–E, G. Method A was used for panels F and H.

https://doi.org/10.1371/journal.ppat.1013440.g003

We next considered which moieties of proteins were necessary to promote Capsaspora motility. First, the most active proteins are known to be carriers of lipids in the serum. To test if the protein-bound lipids were important for induction, we examined chemokinesis by fatty-acid free BSA. This BSA induced motility with a similar potency as native BSA, suggesting that the fatty acids bound to BSA are not important for activity (Fig 3E). Next, we tested if whole protein was necessary or if individual peptides (or even amino acids) could induce chemokinesis. Often simple peptides are ligands for receptors that drive cellular responses. For example, human immune cells respond to N-terminal peptides of bacterial proteins, which contain N-formyl methionine [21].Therefore, we tested if individual peptone peptides could induce chemokinesis. They slightly increased motility at a single concentration (Fig 3F). However, another simple peptide (glutathione) failed to induce any aggregation (Fig 3G). Amino acids were incapable of inducing chemokinesis (Fig 3H). These data suggest that some peptides may be sufficient to induce chemokinesis.

Chemokinesis inducer does not substantially decrease adhesion of cells to surface

To investigate how these inducers increase Capsaspora motility, we first asked if BSA influenced the adhesion of Capsaspora to surfaces. Surface adhesion is intimately associated with motility. If cells stick too tightly to a surface, they cannot move [22]. Alternatively, if they adhere too weakly, then mesenchymal-type motility is also impossible [23]. Prior work has shown that extensive coating with BSA (for 1 hour) can decrease the adhesion of Capsaspora to untreated plastic surfaces [14]. Therefore, we examined if BSA decreased Capsaspora adhesion in our chemokinesis experimental conditions (immediate analysis ~2 minutes after addition of BSA to cells adhered to tissue culture-treated surfaces). We found no significant decrease in adherence upon addition of a chemokinesis-inducing concentration of BSA, followed by washing the cells with a level of force that removes approximately half of the surface-attached cells (Fig 4A). Although we cannot rule out a very slight inhibition of attachment speeding up Capsaspora movement, our data do not support this hypothesis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Insights into chemokinesis induction mechanism.

(A) Capsaspora adhesion to tissue culture-treated plastic microplate wells upon addition of 3 mg/ml BSA or control (water) and agitation by pipetting. Error bars represent standard error of the mean of the means of four individual experiments (n = 4). In each experiment, two or three wells were quantified for each condition. The means from each experiment are displayed as white circles. Data from the individual wells from each experiment are in S4 Fig. (B) Capsaspora motility at different time points after addition of BSA (3 mg/ml). Error bars represent standard error of the mean of a biological triplicate of three individual wells, each with dozens of cells (n = 3). The dashed line indicates baseline motility before addition of BSA. (C) Quantity of fluorescently labeled BSA remaining in the Capsaspora motility assay after 1.5 hours and 3 hours. Error bars represent standard error of the mean of a biological triplicate of three individual wells (n = 3). Individual biological replicates are displayed with white circles. See raw gel image in S5B Fig. (D) Capsaspora motility at different time points after addition of BSA (3 mg/ml). Error bars represent standard error of the mean of a biological triplicate of three individual wells, each with dozens of cells (n = 3). The black filled circles represent wells that received BSA (3 mg/ml) at the start and again after ~6 hours. The open grey circles represent wells that received no BSA. The red “X” symbols represent wells that received no initial BSA treatment, but did receive BSA (3 mg/ml) after ~6 hours. The dashed line indicates the time when the ~ 6 hour BSA dose was added to the black and red conditions. (E) Capsaspora motility with different densities of cells upon addition of BSA (3 mg/ml) or control (water). Error bars represent standard error of the mean of two to six replicates of individual wells, each with dozens of cells (n = 3–9). Individual biological replicates are displayed with white circles. (F) Motility of individual Capsaspora cells from five replicate wells, highlighting cells that are distant from others (blue, incapable of cell-cell contact) exhibit increased motility on par with the average of their well. Circles represent each tracked cell, line and error bar represent mean and standard error of the mean for that well. Dashed line indicates the mean negative control migration in wells with water added instead of BSA. In all panels, the chemokinesis assays were performed using method B. (G) Capsaspora intracellular Ca²⁺ induction (measured via fluorescence of Fluo4-AM dye) with full FBS (n = 1), CBSS negative control (n = 1), and > 30 kDa and <30 kDa FBS components (n = 3 each). Shaded regions indicate standard error of the mean. Vertical dashed line indicates when samples were added to Fluo4-labeled cells.

https://doi.org/10.1371/journal.ppat.1013440.g004

Incubation with chemokinesis inducer desensitizes Capsaspora to further chemokinesis induction

We monitored the kinetics of the increased motility phenotype, and we found that it dissipated over the course of hours (Fig 4B). Initially we suspected that BSA might be depleted from the media, causing the decreased chemokinesis. To test this hypothesis, we monitored the depletion of fluorescently labeled BSA (fl-BSA) from the media. After confirming that the fl-BSA induces chemokinesis (S5A Fig), we collected aliquots of the cell culture media during the chemokinesis experiment conditions. We found that the fl-BSA was not significantly depleted on the timescale of decreased motility (Figs 4C and S5B).

Alternatively, we hypothesized that Capsaspora becomes desensitized to its chemokinesis inducer over time. To test this hypothesis, we examined if re-addition of BSA would cause a renewed increase in motility. Even though this re-addition doubled the concentration of BSA, we found no increased motility relative to the untreated control (Fig 4D, black circles vs grey open circles). This desensitization is almost certainly due to the initial BSA treatment, since cells that were not pre-treated with BSA were still capable of a large motility increase upon BSA addition at the later time point (Fig 4D, red “X” marks). In this case the rate of motility decreased more sharply over time—possibly due to the cells running out of nutrients. Ultimately, the chemokinesis induction was temporary and could not be renewed by re-addition of similar concentrations of BSA. This result may indicate an active adaptation of Capsaspora to the presence of the chemokinesis inducer, as many organisms adapt to the current concentration of a molecular cue [17,18].

Chemokinesis response is cell-density-dependent but not contact-dependent

Cell density influences the speed and migration behaviors of different cell types. For example, in multiple cancer cell lines, higher cell density causes faster cell migration compared to lower densities [24,25]. In contrast, some cells in developing animals stop moving when they contact each other at high density [26]. Therefore, we asked if Capsaspora cell density regulated its motility in response to BSA. We seeded cells at various cell densities and tested for chemokinesis response to BSA. All previous experiments in this manuscript employed cells at 5,000–10,000 cells per 100 µl volume in a microtiter well. Here we found that Capsaspora failed to increase motility at the lowest tested density (500 cells per 100 µl volume in a microtiter well, Fig 4E). Therefore, the chemokinesis response may depend not just on the presence of inducing environmental cues, but also on a Capsaspora-derived ‘signal’.

This ‘signal’ may be a soluble factor that is released from cells, or it may be mediated through contacts between adjacent cells [27]. To assess the importance of cell-cell contacts, we examined if cells that were too distant from others for direct contact still exhibited chemokinesis. Prior work has shown that Capsaspora filopodia can be up to 24 µm long [14]. We also have not observed filopodia longer than this. Therefore, to conservatively bin cells that should not be contacting other cells, we monitored cells that were > 60 µm away from any other cell in the well. We compared the chemokinesis response of these ‘distant’ cells to the average of the cells in their well and found that the distant cells migrated as fast as the average (Fig 4F). Therefore, we conclude that the cell-derived chemokinesis signal is not dependent on cell-cell contacts. It may be a secreted soluble molecule.

In sum, it may be that the ultimate chemokinesis ‘signal’ is a soluble factor, which is produced by Capsaspora and secreted in response to the addition of the chemokinesis cue. Alternatively, the density-dependent signal may be constantly secreted by Capsaspora and synergize with the environmental cue to essentially provide an AND logic gate where both the Capsaspora signal and the environmental cue must be present for chemokinesis to proceed.

Chemokinesis is not induced via Ca²⁺ signaling

Chemokinesis and chemotaxis responses are regulated by the Ca²⁺ secondary messengers in cell types including neutrophils and fibroblasts [28,29]. In the protist Dictyostelium discoideum, cellular Ca²⁺ levels spike upon addition of the chemoattractant cAMP, but this response does not appear to be necessary for chemotaxis itself [30]. Cellular aggregation in Capsaspora also requires the presence of extracellular Ca²⁺, although it is unclear if it is related to signaling [8,12]. To investigate the potential involvement of changes in intracellular Ca²⁺ concentration in Capsaspora chemokinesis, we used the cell-permeable Ca²⁺-sensitive dye Fluo4-AM. Although we found that whole FBS induced a stark Ca²⁺-dependent increase in fluorescence, this effect was due to the < 3 kDa small molecules in FBS, not the > 3 kDa macromolecules that induce chemokinesis (Fig 4G). This finding suggests that chemokinesis in Capsaspora is not regulated by the Ca²⁺ secondary messenger. Since aggregation is also activated by > 3 kDa lipoprotein components of FBS, this finding also suggests that Ca²⁺ signaling is not involved in inducing Capsaspora aggregation. Ongoing work is necessary to further elucidate the non-Ca²⁺-related intracellular signaling mechanisms that regulate both chemokinesis and aggregation in Capsaspora.

Chemokinesis alone can improve predation

Our experiments have revealed that Capsaspora cells exhibit chemokinesis in response to various lysates and proteins. While chemotaxis can direct protist predators towards prey [31], the role of chemokinesis in predation is unclear. Previous experiments and mathematical models of predation by bacteria and leukocytes have shown that chemokinesis can improve the efficiency of chemotaxis, but no benefit has been reported for chemokinesis in the absence of chemotaxis [5,6,32,33]. We asked if chemokinesis alone could benefit Capsaspora. To explore this, we developed a mathematical model of Capsaspora cells responding to a lysed (i.e., leaking) schistosome that exudes a chemokinesis-inducing chemical, see S1 Text.

The theoretical expectation is that organisms displaying chemokinesis in the absence of chemotaxis will accumulate away from the prey (a detrimental behavior for predation) [34]. This expectation stems from the facts that (a) the predator cells will move in random directions in the absence of chemotaxis, and (b) the concentration of the chemo-effector is highest near the source (prey). Therefore, cells near the prey will quickly diffuse randomly away from the prey and then slow down and accumulate far from the prey. However, one feature of Capsaspora predation that may mitigate this problem is its irreversible attachment to large prey. Once Capsaspora gets close enough to its prey, it will attach to the prey and not diffuse away until the prey is fully devoured—which can take hours (or possibly longer, depending on the size of the prey and number of Capsaspora cells attached). Since chemokinesis does not change the predator cells’ trajectories, we do not expect it to significantly impact the number of predators that ultimately encounter prey over an infinitely long timescale. However, chemokinesis should increase the rate at which each predator cell randomly encounters the prey and attaches to it. For the cells that do encounter the prey, the decreased time to encounter would increase the amount of time attached to the prey and feeding (i.e., “residence time”). Therefore by this logic, chemokinesis could be advantageous for predation.

To explore this hypothesis, we simulated a population of protists interacting with a schistosome, assumed spherical for simplicity. The schistosome was also assumed stationary, based on the very small movement of these organisms in our in vitro experiments (S1–S6 Videos, available at https://doi.org/10.5967/sh2f-xt93). Further, the schistosome was assumed to exude chemicals with the same diffusivities, and eliciting the same chemokinetic responses for Capsaspora, as bovine serum albumin (BSA, Fig 2F). Details of the simulations can be found in S1 Text. Three replicate simulations were run for both a condition with chemokinesis enabled and one without, each over the course of 2 hours (in line with the length of experiments). As shown in Fig 5A, the simulations predict a higher probability of attachment with chemokinesis enabled during the course of 2 hours. Furthermore, the cells that attach to the prey in the chemokinesis simulation attach earlier than the cells that attach in the non-chemokinesis case (Fig 5B), which gives them an even longer residence time. When considering both the higher number of attached cells and the longer amount of time attached, the ‘cumulative residence time’ of all cells attached to the prey was ~ 1.25 × higher when chemokinesis occurred (Figs 5C and S6). Overall, our model predicts that chemokinesis alone can be an advantageous predatory behavior when the predator irreversibly attaches to its prey to feed for a long time.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Numerical simulations of predator-prey interaction reveal benefit for chemokinesis.

(A) Attachment rates (i.e., the fraction of Capsaspora cells that attached to the schistosome prey) during a 120-minute simulation of Capsaspora cells predating a schistosome with or without chemokinesis to a chemo-effector (with the properties of BSA) exuding from the schistosome. When the cells reach the schistosome, they irreversibly attach to it. (B) Probability of different residence times over the 120 minute simulation period for cells that ultimately attached to the schistosome. “Conditional probability” is the probability of a cell attaching to the schistosome at that moment, given the condition that the cell attaches at some point during the 120 minutes. Cells that never attached are not included in the plot. Three replicate simulations for each condition are shown as individual curves. Curves were generated from kernel density estimates. (C) Cumulative residence times (i.e., the total sum of time that the 2000 Capsaspora cells are attached to the schistosome over the 120 minute simulation) with and without chemokinesis. Additionally, the diagonally shaded bar represents the case where chemokinesis is enabled and four nearby schistosomes were depleted and gave a final burst of chemo-effector at the start of the simulation—no significant difference was observed compared to the standard chemokinesis condition. See S6 Fig for the full kinetic analysis. In all bar plots for this figure, error bars represent standard error of the mean of a triplicate of three individual simulations, each with 2000 cells (n = 3). Individual simulation means are represented with dots.

https://doi.org/10.1371/journal.ppat.1013440.g005

Another way that chemokinesis may aid predation is by increasing dispersion away from recently devoured prey. If the prey releases a large amount of chemokinesis-inducing chemicals, this would increase cell speed, and thus diffusion, encouraging higher dispersal away from the region where there is no longer prey. Thus, randomly moving protist cells need to search for less time before encountering another not-yet-devoured schistosome, which benefits the predator by decreasing time “searching” and increasing the time “residing” attached to prey. To explore this hypothesis, we repeated our simulations, this time using an initial chemical profile that models several recently devoured schistosomes in the vicinity of an intact one. Overall, we did not observe a significant change in the cumulative residence time compared to the standard chemokinesis simulation (Figs 5C and S6). The high diffusivity of the chemo-effector may explain this observation. Since the consumed cells stop secreting the chemo-effector, its concentration dissipates quickly relative to the rate of cell movement.

In sum, our model predicts that chemokinesis alone (without chemotaxis) can be beneficial for predation when the predator irreversibly attaches to its prey. Chemokinesis significantly increased attachment rates of predators to prey over a 2 hour simulation. Our simulation did not include chemotaxis (i.e., directional bias in motility toward prey), and we observed some variation in cumulative residence times in all conditions across repeated simulations (Figs Ib and Ic in S1 Text). Therefore, since random motion has a large effect in predation success, it seems likely that chemotaxis, acting in concert with chemokinesis, can provide an even greater advantage over chemokinesis alone. Specifically, chemotaxis would bias motion of predators towards beneficial trajectories, reducing randomness and increasing success in predation, as mathematical models have predicted in the case of bacteria [5]. To show this, the chemoattractants to which Capsaspora responds would need to be identified and quantified, which is beyond the scope of this study. This would allow our mathematical model to be extended, parameterized, and compared with experimental results.

Materials used

Reagents used in this study are summarized in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Reagents used in this study.

https://doi.org/10.1371/journal.ppat.1013440.t001

Cell strain and growth conditions

Capsaspora owczarzaki cell cultures (strain American Type Culture Collection ATCC 30864) were grown axenically in 25-cm² culture flasks with 6 ml ATCC media 1034 (modified PYNFH medium) containing 10% v/v heat-inactivated Fetal Bovine Serum (FBS, Corning 35–011-CV), hereafter growth media, in a 23 °C incubator. Cells were obtained during exponential growth phase by passaging ~0.9 – 1.2 ml of adherent cells scraped into suspension within the specified density range, into 5ml of growth media. The culture was incubated at 23 °C for 24 h to achieve desired density range of 3 × 10⁶ to 20 × 10⁶ cells/ml.

S. mansoni miracidia isolation and transformation into primary sporocysts

S. mansoni miracidia were isolated according to published protocols [47,48]. Briefly, 3 infected mice livers were obtained from the Biomedical Research Institute (Rockville, MD) (BRI) Schistosomiasis Resource Center. They were shipped overnight with cold packs and added to ice-cold 1.2% sterile NaCl solution immediately after arrival using a sterile 50 ml centrifuge tube. The tubes with the livers were shaken vigorously, and any floating debris and the saline solution was aspirated. The livers were then poured into the ice-cold cup of a Waring blender with another 30 ml of ice-cold sterile 1.2% NaCl solution and blended for 15–25 seconds non-stop making sure the livers were well-blended. The blended livers were centrifuged at 800 × g for 10 min at 4 °C. The supernatant was discarded, and the pellets were washed with ice-cold sterile 1.25% NaCl solution. The pellets were then resuspended with 40 ml room temperature sterile artificial pond water (recipe from BRI, 0.46 μM FeCl₃, 220 μM CaCl₂, 100 μM MgSO₄, 310 μM KH₂PO₄, 14 μM (NH₄)₂SO₄ in water adjusted to pH 7.2 with NaOH) and transferred into the aluminum-covered sterile volumetric flask. The flask was carefully filled with sterile pond water to approximately 3 cm above the foil covering the neck. The flask was covered with a small petri dish and a bright light shone across the top of the flask. The eggs started to hatch at room temperature. When miracidia had amassed at the surface after 10 min, around 10 ml of the pondwater with miracidia was transferred to a sterile petri dish. The collected miracidia were centrifuged at 290 × g at 4 °C for 2 minutes. Only 1 ml of pondwater was left with the miracidia pellet and another 1ml of Chernin’s Balanced Salt Solution (CBSS+) [20] was added to the petri dish. The miracidia were then transferred to a 24-well plate and incubated at 26 °C overnight to stimulate miracidial transformation into primary sporocysts.

Chemotaxis/chemokinesis near schistosomes

Capsaspora cells were counted and resuspended such that the final cell density was 60 × 10³ cells/ml. 200 µl of the Capsaspora cells were seeded in each well of a TC-treated 96-well plate (Corning #353219) and incubated overnight at 26 °C. In the experiment run on fibronectin-coated plated (S1B and S1C Fig), a TC-treated 96-well plate (Corning #353219) was pre-coated with 100 µl of 70 µg/ml fibronectin (Sigma #F1141) in 25% FBS-free media before adding Capsaspora. The next day, schistosome sporocysts were centrifuged for 30 seconds at 300 × g at room temperature and the supernatant was aspirated immediately. The pellet was then resuspended with 650 µl 25% FBS-free media. Media was aspirated out of the Capsaspora wells and replaced with 200 µl of the sporocyst cell culture to afford a few schistosomes per well. Negative control wells received media without schistosomes. For imaging each well, a field of view was set such that there is at least one sporocyst surrounded by many Capsaspora cells. Images were collected every one minute for 4 hours using an A1 Nikon inverted microscope. Cells were then analyzed using Imaris image analysis software (Bitplane, version 10.0.1). ImageProcess was used to invert the image, the schistosome was set as the origin reference, and spots were created. Tracking was enabled. Estimated cell diameter was set to 6.43 µm. Background was subtracted, and the automatic threshold was used to quality filter the spots. Cells 70–240 µm from the origin reference frame were selected because cells closer than 70 µm might attach to the schistosome via filopodia, and cells further than 240 µm were near the boundaries of the field of view and may be near other schistosomes. The autoregressive motion algorithm was used for tracking. To remove spurious readings by floating cells (which appear to migrate abnormally quickly and/or appear/disappear quickly), we excluded cells that migrated > 11.2 µm in adjacent frames or lost tracking for > 3 frames. The ‘fill gap’ function was disabled, and only tracks ≥3 frames were analyzed. For chemotaxis, the distance moved relative to the reference (schistosome) was reported for each cell track. To observe maximal movement, only tracks that had ≥ 2 hours (120 frames) were included for plotting. For chemokinesis, the distance moved frame-to-frame for each cell in each frame was reported. Only data for the first 15 minutes of the video were plotted, and only for cells 70–90 µm from the schistosome, since this was the timeframe and distance with the most notable chemokinesis.

Chemokinesis assay

This assay was always performed with the strain ATCC 30864 unless mentioned otherwise. The standard chemokinesis assay was performed on TC-treated 96-well plate (Corning #CLS3997). Either 5 × 10³ cells or 10 × 10³ cells were seeded in 90 µl of FBS-free growth media per well in the TC-treated 96-well plate and allowed to settle for 2 h. This low cell density was chosen to avoid confounding results from cellular aggregation (which is known to occur with Capsaspora in the presence of FBS at higher cell densities [e.g., 8 × 10⁵ cells per well] in low-attachment [non TC-treated plates]) [8–10].

Chemokinesis quantification

When the Cytation 10 was used to record cell movement, analysis was performed using BioTek Gen 5 software. Briefly, images were pre-processed using a dark background with background flattening using a rolling ball diameter of 9 µm and image smoothing strength of 2 cycles. Kinetic frame alignment was employed based on the previous valid image with a maximum offset of 200 µm. The cellular analysis module was employed to track cells. A primary mask was applied using a threshold value of 5000 with a dark background, split touching objects, track objects, and fill holes in masks. Object size was filtered to be 1–100 µm. To remove spurious readings by floating cells (which appear to migrate abnormally quickly and/or appear/disappear quickly), we made a subpopulation of cells that migrated ≤60 µm and were tracked for ≥ 18 frames. The ‘average velocity’ of this subpopulation of cells was reported as the ‘mean velocity’.

Cell lysis

Biomphalaria glabrata embryonic (Bge) cells [49,50] were grown in Bge medium with 10% FBS at 26 °C. The culture was resuspended to 20 × 10³ cells/ml and snap-frozen in 20 µl aliquots until further use. The frozen aliquot was freeze-thawed 5 times with a 20 second thawing period between each cycle. An additional 80 µl of FBS-free media was added to make the final volume 100 µl. The lysed culture was centrifuged at 15,000 × g for 5 minutes and the supernatant was collected for further testing.

Both animal cell cultures were freshly received and resuspended to appropriate densities using FBS-free media. The cultures were then centrifuged at 15,000 × g for 5 min and the cell pellets were lysed using freeze-thaw method for 5 times with 20 second thawing period between each cycle. The lysed cultures were centrifuged at 15,000 × g for 5 min to collect the lysate supernatant.

E. coli MG1655 culture was grown overnight in lysogeny broth (LB) at 37 °C with shaking and sub-cultured to 50 ml with 1:50 dilution of the overnight culture into fresh media. The 50 ml culture was incubated an additional 18 hours with shaking at 37 °C. The culture was centrifuged at 10,000 × g for 15 min. The cell pellet was resuspended in 1 ml of FBS-free media. The culture was lysed using a probe sonication for 10 seconds followed by a 20-second paused, repeated for a total of 5 min. The lysed culture was centrifuged at 10,000 × g for another 15 min and the lysate supernatant was collected for further analysis.

B. subtilis 168 culture was grown overnight in lysogeny broth (LB) at 37 °C with shaking and sub-cultured to 50 ml with 1:50 dilution of the overnight culture into fresh media. The 50 ml culture was incubated an additional 18 hours with shaking at 37 °C. The 50 ml culture was centrifuged at 10,000 × g for 15 min. The cell pellet was resuspended in 1 ml of 4 mg/ml lysozyme and incubated for 15 min. The culture was centrifuged again at 15,000 × g for 15 min and the cell pellet was resuspended in 1 ml of FBS-free media. The culture was lysed using sonication for 10 seconds followed by a 20-second paused, repeated for a total of 5 min. The lysed culture was centrifuged at 10,000 × g for another 15 min and the lysate supernatant was collected for further analysis.

FBS heat-treatment

100 µl of FBS in an Eppendorf tube was heated on dry heat block at 75 °C for 30 minutes. The treated and non-treated FBS was tested for chemokinesis assay. The FBS-free growth media was used as negative control.

Bioassay-guided fractionation

FBS was initially fractionated using either Amicon Ultra 30 kDa or 100 kDa cutoff filters (Sigma, #UFC5030, #UFC5100) according to the manufacturer’s direction. Briefly, the filters were first washed with 500 µl of CBSS by centrifugation at 14,000 × g for 15 min. Then 500 µl of FBS was added to the cutoff filter and centrifuged at 14,000 × g for another 15 min. The small fractions (< 30 kDa, < 100 kDa) were collected for testing, and 20 µl of each large fraction (> 30 kDa or > 100 kDa) was added to 180 µl of CBSS for testing. 10 µl of each stored fraction was used for testing chemokinesis.

Isolation using strong anion exchange column (SAX)

Isolation using gel filtration (size exclusion chromatography [SEC])

Different cell density seeding

After measuring the cell density in a culture flask, the culture was appropriately diluted such that seeding with 90 µl of each diluted culture achieves the desired density.

Desensitization experiments

Chemokinesis assay was performed using the method B and Gen5 for imaging analysis. 10 µl of 30 mg/ml BSA was added to induce chemokinesis as mentioned at t = 0 and 6h.

Adhesion assay

5 × 10³ cells were seeded in 90 µl of FBS-free growth media per well in the TC-treated 96-well plate and allowed to settle for 2 h. After seeding cells, 85–90% of FBS-free media was removed by aspiration and replaced with Chernin’s balanced salt solution (CBSS+). A Cytation 10 imaging reader was set to capture ~80% of the well’s area to count this initial cell number. After imaging, 10 µl of either BSA or water were added to each well and allowed to sit for 2 min (to mimic the chemokinesis conditions). 85–90% of FBS-free media was removed by aspiration and replaced with 85 µl CBSS+ solution. The Cytation 10 imaging reader was again used to capture ~80% of the well’s area to count the remaining cell number. The remaining cell number was divided by the initial cell number to determine the % cells remaining for each well.

Calcium signaling assay

7 × 10⁴ cells were seeded in 200 µl of FBS-free growth media per well in the TC-treated black-walled 96-well plate (Corning 353219) and allowed to settle for 2 h at room temperature. After seeding cells, 85–90% of FBS-free media was removed by aspiration and replaced with 100 µl Chernin’s balanced salt solution (CBSS+) containing 5 µM Fluo-4 AM dye with 0.1% DMSO and 0.02% pluruonic F-127 surfactant. The cells were incubated in the dark for 45–60 minutes before aspirating the solution and replacing with 100 µl of fresh CBSS + . Cells were imaged using a Olympus IX83 P2ZF inverted microscope fitted with a Yokogawa CSU-W1 spinning disk confocal system, a Hamamatsu ORCA-Flash4.0 camera, and a UPlanSApo 20x/0.75 objective. Fluorescence excitation was performed with a 488 nm laser, and emission was detected at 519 nm with 2x2 binning. Images were taken every 5 seconds for 3 minutes. After 25 seconds of baseline readings, 10 µl of test sample was added (i.e., FBS, CBSS, > 30 kDa FBS, < 30 kDa FBS). Images were analyzed using Imaris image analysis software (Bitplane, version 10.0.1). ImageProcess was used to invert the image, and spots were created. Tracking was enabled. Estimated cell diameter was set to 3.00 µm. Background was subtracted, and the automatic threshold was used to quality filter the spots. The Brownian motion algorithm was used for tracking. To remove spurious readings by floating cells (which appear to migrate abnormally quickly and/or appear/disappear quickly), we excluded cells that migrated > 2.30 µm in adjacent frames or lost tracking for > 3 frames. The ‘fill gap’ function was disabled. The center intensity for each spot (i.e., each cell) was exported for each frame along the entire track. In Microsoft Excel, cell tracks without all frames included were removed. For each cell track, the initial spot intensity was subtracted from all spot intensity readings, thereby normalizing the initial intensity for every cell to zero. For each well at each time point, the normalized spot intensities for each cell were averaged (n ~ 200 cells). Three replicate wells were analyzed, and their mean ± standard error were plotted (n = 3).

Numerical modelling of predator prey interaction with chemokinesis

Simulation of a population of Capsaspora cells was completed in MATLAB; details can be found in S1 Text. Capsaspora cells were modelled as individual agents, using Monte-Carlo methods to account for stochasticity. Rules of motion were inferred using a data driven approach. We analyzed the motility statistics of WT Capsaspora cells on a uniform substrate from Video 1 in reference [14]. We found that Capsaspora cells display strong directional persistence following a truncated normal distribution for the correlation between direction of currently extended filopodia (Fig C in S1 Text) and the direction of the next extended filopodia. This implies that cells tend to carry on travelling in the same direction with only minor adjustments to the right or left, making larger adjustments rarely, so they would take a long time to reverse direction. We then measured distance travelled after extension of filopodia and found it follows an exponential distribution (Fig B in S1 Text), so that distances travelled following filopodia extensions are fairly uniform. To model chemokinesis, we parameterized the curves found for BSA as well as HSA, which were found to have the same shape with the only difference being BSA saturates at a higher locomotion speed, details can be found in S1 Text.

The schistosome was then modelled as a stationary circle, that would cause cells to become irreversibly attached if they made their way to the schistosome surface. The schistosome would then exude the chemokinesis-inducing chemical at a rate dependent on the local attachment rate of Capsaspora cells to the surface. Dispersion of chemical was modelled in a continuum as standard isotropic diffusion with a constant diffusivity. These simulations were run with and without chemokinesis to determine if chemokinesis was beneficial to predation by increasing residency time of Capsaspora at the schistosome surface. The simulations were then run again for the case of a chemical profile reflecting several recently lysed schistosomes, to determine if they would act as a ‘dispersal agent’ and encourage cells to find new prey.