Using video microscopy to improve quantitative estimates of protozoal motility and cell volume

The objective of this study was to apply digital imaging to improve quantification of rumen protozoal biomass and distinguish treatment differences in cell motility and volume among ruminal protozoa. Observations of protozoa in rumen fluid treated with essential oils (CinnaGar, CIN; Provimi North America, Brookville, OH) or an ionophore (monensin, MON) indicated possible cell shrinkage. We hypothesized that MON would decrease protozoal motility and interact with CIN on cell volume. In addition, we hypothesized that analysis of still frames from video of swimming protozoa would improve volume prediction accuracy. Flocculated rumen fluid was incubated in batch culture dosed with N-free feed only (control), MON, CIN, or a combination of MON+CIN. Samples were taken at 0, 3, or 6 h post-treatment and wet-mounted on a microscope fitted with a high-definition camera. At 3 h post-inoculation, there was a treatment interaction for average speed such that CIN attenuated the effect of MON, with treatment means of 243, 138, 211, and 183 µm/s for control, MON, CIN, and MON+CIN, respectively. At 6 h post-inoculation, MON decreased average speed by 79.2 µm/s compared with the main effect mean without MON. We measured both minimum and maximum diameters (depth and width, respectively) perpendicular to the longitudinal axis of swimming protozoa, yielding a 3-dimensional estimate of protozoal volume. The ellipsoid formula (4/3)πabc, where a = 1/2 length, b = 1/2 width, and c = 1/2 depth, was compared with previously published volume estimations using genera-specific coefficients (genera-specific coefficient × length × width²). Residuals (genera-specific coefficients - ellipsoid) were plotted against predicted (ellipsoid) and centered to the mean (X_i-x¯) to evaluate both mean and slope biases. For Entodinium spp., Y = 0.248 (±0.037) (X_i - 7.98 × 10⁴) + 1.97 × 10⁴ (±1.48 × 10³); n = 100; r² [coefficient of determination (squared correlation coefficient)] = 0.31, with significant slope and mean biases. For family Isotrichidae, Y = -0.124 (±0.068) (X_i - 2.54 × 10⁶) - 1.21 × 10⁴ (±4.86 × 10⁴); n = 32; r² = 0.10, where slope tended to be different from zero but with no mean bias. For Epidinium spp., Y = 0.375 (±0.056) (X_i - 2.45 × 10⁵) + 6.65 × 10⁴ (±0.28 × 10⁴); n = 64; r² = 0.43, with both mean and slope biases. The present regression analyses demonstrate that the genera-specific coefficient-based method more likely overestimates volume for Entodinium and Epidinium than for the teardrop-shaped Isotrichidae. Based on simulations derived from previous literature reporting treatments that depress protozoal populations or among-animal changes in protozoal population structures, our proposed ellipsoid method offers potential to advance the prediction of treatment effects on protozoal volume and to shift focus from the number of cells present to the diversity, function, and biomass of protozoa under various treatment conditions.