Switching of behavioral modes and their modulation by a geometrical cue in the ciliate Stentor coeruleus

Protists ubiquitously live in nature and play key roles in the food web chain. Their habitats consist of various geometrical structures, such as porous media and rigid surfaces, affecting their motilities. A kind of protist, Stentor coeruleus, exhibits free swimming and adhering for feeding. Under environmental and culture conditions, these organisms are often found in sediments with complex geometries. The determination of anchoring location is essential for their lives. However, the factors that induce the behavioral transition from swimming to adhering are still unknown. In this study, we quantitatively characterized the behavioral transitions in S. coeruleus and observed the behavior in a chamber with dead ends made by a simple structure mimicking the environmental structures. As a result, the cell adheres and feeds in narrow spaces between the structure and the chamber wall. It may be reasonable for the organism to hide itself from predators and capture prey in these spaces. The behavioral strategy for the exploration and exploitation of spaces with a wide variety of geometries in their habitats is discussed.

Prior to transferring a cell into an observation chamber, we washed it with pond water again and equilibrated it for over 20 min. (B) The observation chamber was made by a silicon sheet (0.2 mm depth, 5 mm inner diameter) and covered on a cover slip. Cell behavior was observed by an inverted microscope equipped with the objective PLANPO X1.25 under a bright field using 700 nm light.

Supplementary Figure 3. The time series of the cell length (blue line) and swimming speed (orange line) in 14 cells.
We used 14 cells to obtain the frequencies of each cell state. We recorded each cell for 24.7 ± 3.4 min (SD, 14 cells, maximum 32.0 min, minimum 18.6 min, total 345.1 min). Some cells take the steady cone state (Cells 1-4), others change their states frequently (Cells 9-14). There is a variation in the frequency of changing the states among those cells (Cells 5-8).

Supplementary Figure 4. Additional data for quantitative measurements of cell behavior.
To measure the duration time of each shape transition and the diameter of the swimming trajectory in the trumpet shape, we recorded the cell behaviors and analyzed the data. Because of the very low frequency at the droplet state, we recorded further behaviors of the cells in the droplet state by adding an external mechanical stimulus ( First, we divided the rotating trajectories of cell i (A) into each lap (j-th lap) denoted by "trajectory i, j" (B-D). The diameter of "trajectory i, j", namely, dij, was calculated by the maximum distance dijk between the center at the k-th frame and other centers in "trajectory i, j" and averaging dijk with respect to k (C). The histogram of dij is shown in Figure 3F in the main text.

Supplementary Figure 7.
The drawings of the gel chamber containing a disk-like structure (diameter 2.5 mm, 0.3 mm depth) inscribed in the quasi-2D disk chamber (diameter 5 mm, 0.3 mm depth). The observation chambers were obtained by mold casting in two processes: the preparation of prime molds made of polydimethylsiloxane (PDMS) and the formation of a gel chamber by pouring agarose gel into the PDMS prime molds.

Supplementary Figure 8.
The time series of the cell length (blue) and swimming speed (orange) in the chamber without the crescent areas.

Supplementary Figure 9.
The time series of the cell length (blue) and swimming speed (orange) in the chamber with the crescent areas.

Supplementary Figure 10.
(A) In the contraction process, the two-mode exponential function fits well (2.1 ms and 15.6 ms (n = 8, 8 cells)) to the averaged deformation rate rather than the function with a single time scale. The inset of the graph indicates the coefficient of determination R 2 regarding the numbers of modes.
(B, C) In the two individual contraction processes, the two-mode exponential function also fits well to the deformation rate rather than the function with a single time scale. The inset of the graph indicates the coefficient of determination R 2 regarding the numbers of modes.
The results indicate that individual contraction has two characteristic times in the quick contractile process from trumpet to droplet, but there are not two types of samples.
Supplementary Figure 11. The difference in the beating forms of the ciliary band (membranellar band) in S. coeruleus in cone and trumpet states.
We used the cells collected in the Shiribetsu River and cultured them by the methods described in Section 3.1. Before observation, we washed the cells with fresh modified Peters' solution two times. After 1-2 h, we transferred the cells into a chamber made with a silicon sheet (5 mm diameter and 0.2 mm depth) and covered the chamber with a cover slip. The beating forms were observed by an inverted microscope IX73 equipped with the objective UPLSAPO60XW (Olympus, Tokyo, Japan) and recorded by using the FASTCAM Mini AX50 at 2,000 fps and an exposure time of 1/2,000 s under a brightness field using weak 700 nm light.
We manually traced one membranellar band every 4 ms. Due to 3D beating, we could not extract the segmentation lines in the recovery strokes in the cone state. The beating frequencies are approximately 15 Hz in the cone state and 30 Hz in the trumpet state, and the angles of the beat are approximately 130 degrees in the cone state and 70 degrees in the trumpet state. The parameters of the beating form of the membranellar band are different. The differences may contribute to the switching motility from a straightforward swimming trajectory to a rotating one. Spindle speed (rpm) 8000 8000

Supplementary movie 1
The switching behavior from cone to trumpet. The movie corresponds to Figure 3C.