Establishing a High-Throughput Locomotion Tracking Method for Multiple Biological Assessments in Tetrahymena

Abstract

Protozoa are eukaryotic, unicellular microorganisms that have an important ecological role, are easy to handle, and grow rapidly, which makes them suitable for ecotoxicity assessment. Previous methods for locomotion tracking in protozoa are largely based on software with the drawback of high cost and/or low operation throughput. This study aimed to develop an automated pipeline to measure the locomotion activity of the ciliated protozoan Tetrahymena thermophila using a machine learning-based software, TRex, to conduct tracking. Behavioral endpoints, including the total distance, velocity, burst movement, angular velocity, meandering, and rotation movement, were derived from the coordinates of individual cells. To validate the utility, we measured the locomotor activity in either the knockout mutant of the dynein subunit DYH7 or under starvation. Significant reduction of locomotion and alteration of behavior was detected in either the dynein mutant or in the starvation condition. We also analyzed how Tetrahymena locomotion was affected by the exposure to copper sulfate and showed that our method indeed can be used to conduct a toxicity assessment in a high-throughput manner. Finally, we performed a principal component analysis and hierarchy clustering to demonstrate that our analysis could potentially differentiate altered behaviors affected by different factors. Taken together, this study offers a robust methodology for Tetrahymena locomotion tracking in a high-throughput manner for the first time.

Keywords: TRex; Tetrahymena; complexity reduction; locomotion; protozoa; toxicity.

Figure A1

The footage of Tetrahymena cell movement in different conditions. (A) Trajectory comparison of wild-type (CU428 strain) and mutant (DYH7neo3) Tetrahymena. (B) Trajectory comparison of Tetrahymena cells in high- and low-nutrient media. (C) Trajectory comparison of Tetrahymena locomotion after exposed to CuSO₄ of different concentrations (0–5000 µM).

Figure A2

Swimming inhibition of Tetrahymena after 30 min exposed to copper sulfate. Using different concentrations (0, 0.5, 5, 50, 500, and 5000 µM), the EC₅₀ was determined at 154.0 µM which reduced swimming speed of cells to 50%. The data are expressed as mean ± standard deviation (SD).

Figure 1

Experimental workflow for T. thermophila locomotion tracking and locomotor activity measurement. T. thermophila samples were transferred to six-well depression slides that were able to accommodate 100–200 µL samples. Original video was captured for 1 min for each sample by a high-resolution CCD mounted onto upright microscope. Later, this 1 min video was handled by TGrabs tool for object identification. Finally, each individual cell’s locomotor trajectory and XY coordinates were tracked by the TRex tool. (A) The output result of XY trajectory from one individual T. thermophila cell. (B) Combining all the data from separate individual cells into a single spreadsheet file using VBA. (C) The trajectory footage of all T. thermophila movement captured in the video.

Figure 2

Summary of multiple endpoints of locomotor activity analyzed in T. thermophila. (A) Total distance traveled (mm), (B) average speed (mm/s), (C) total burst movement count, (D) average angular velocity (°/s), (E) meandering (°/µm), and (F) total rotation movement count were calculated based on the 10 s video recording. Median and interquartile range are used to express the data (n = 249).

Figure 3

Comparison of locomotor activity between wild-type (CU428) and mutant (DYH7neo3) T. thermophila for 10 s. Six locomotor endpoints of (A) total distance traveled, (B) average speed, (C) burst movement, (D) average angular velocity, (E) meandering, and (F) rotation movement were statistically analyzed by Mann–Whitney test (n = 158 for each group; * p value < 0.05; **** p value < 0.0001). Median and interquartile range are used to express the data. Cumulative trajectory paths are presented in Figure A1.

Figure 4

Comparison of T. thermophila locomotor activity in high- and low-nutrient media for 10 s. Six locomotor endpoints of (A) total distance traveled, (B) average speed, (C) average angular velocity, (D) meandering movement, (E) burst movement, and (F) rotation movement were statistically analyzed by Mann–Whitney test (n = 167 for each group; **** p value < 0.0001). Median and interquartile range are used to express the data. Cumulative trajectory paths are presented in Figure A1.

Figure 5

Comparison of T. thermophila locomotor activity after 30 min. exposure of copper sulfate. Six locomotor endpoints of (A) total distance traveled, (B) average speed, (C) burst, (D) average angular velocity, (E) meandering, and (F) rotation movement were calculated based on the 10 s video recording. The data were statistically analyzed by Kruskal–Wallis test (n = 129 for each group, except for 5000 μM n = 26; ** p value < 0.01; *** p value < 0.001; **** p value < 0.0001). Median and interquartile range are used to express the data. Cumulative trajectory paths are presented in Figure A1.

Figure 6

Behavior endpoint comparison in T. thermophila with different treatments. (A) Hierarchical heatmap clustering analysis and (B) principal component analysis (PCA). Four different treatment groups with following conditions: starvation (low nutrients) is displayed in purple color, copper exposure (0.5; 5; 50; 500; and 5000 µM) in blue color, mutant (DYH7neo3) in green color, and the untreated group is included as the control (red color).