MEMS-based force-clamp analysis of the role of body stiffness in C. elegans touch sensation

Abstract

Touch is enabled by mechanoreceptor neurons in the skin and plays an essential role in our everyday lives, but is among the least understood of our five basic senses. Force applied to the skin deforms these neurons and activates ion channels within them. Despite the importance of the mechanics of the skin in determining mechanoreceptor neuron deformation and ultimately touch sensation, the role of mechanics in touch sensitivity is poorly understood. Here, we use the model organism Caenorhabditis elegans to directly test the hypothesis that body mechanics modulate touch sensitivity. We demonstrate a microelectromechanical system (MEMS)-based force clamp that can apply calibrated forces to freely crawling C. elegans worms and measure touch-evoked avoidance responses. This approach reveals that wild-type animals sense forces <1 μN and indentation depths <1 μm. We use both genetic manipulation of the skin and optogenetic modulation of body wall muscles to alter body mechanics. We find that small changes in body stiffness dramatically affect force sensitivity, while having only modest effects on indentation sensitivity. We investigate the theoretical body deformation predicted under applied force and conclude that local mechanical loads induce inward bending deformation of the skin to drive touch sensation in C. elegans.

Figure 1. C. elegans gentle touch is enabled by six touch receptor neurons (TRNs) embedded in the outer shell of the body

(Top) Each TRN innervates roughly 50% of the body length. Since wild type C. elegans crawls on its side, ALMR/L and PLMR/L are near the apex of the body at all times. (Bottom) A cross-section through the body plan reveals the key elements of the outer shell: the cuticle, the hypodermis, and the body wall muscles (not to scale). The internal anatomy and organs is not shown. The TRNs are embedded in the hypodermis just below the cuticle. Approximate TRN positions are from Altun, et al. Body plan coordinates: a – anterior, p – posterior, v – ventral, d – dorsal, r – right, l – left.

Figure 2. Integrated piezoresistive cantilever force-clamp and behavioral analysis system

(A) Schematic diagram of the system which is composed of a piezoresistive microcantilever mounted on a printed circuit board and moved in the z-direction by a piezoelectric actuator, a video camera for real-time visualization and posthoc behavioral video analysis, and a high-intensity light source for optogenetics. Actuator position is under feedback control to enable delivery of defined forces (see Methods). Elements marked with asterisks (*) are connected to computers for control and data acquisition. (B) Still image from a behavioral experiment. The bead at the tip of the cantilever is positioned over the anterior portion of a forward moving animal. (C) Applied force is determined from the measured cantilever deflection (z_c) and calibrated cantilever stiffness (k_c). Deflection of the animal’s body surface consists of both local (at the bead-cuticle interface) and global deformation. (D) The force clamp system can deliver forces from nNs to tens of μNs. (E) For behavior experiments, a designated force is held for 100 ms, then the cantilever is retracted to hover above the animal and deliver no force for a 2 s observation period. Micrographs are a series of still images showing a wild-type worm responding to a force of 1 μN (top), but not responding to a force of 0.1 μN (bottom). (F) Body curvature matrices record curvature as a function position along the body’s length (y-axis: top is l = 0% of body length from head, bottom is l = 100%) as a function of time. Red color indicates dorsal bending; blue indicates ventral bending. More intense colors indicate larger curvature. For a 1-μN stimulus, the matrix on the right illustrates a change in movement direction of the animal. (G) Head position plotted versus time also depicts a change in movement direction with 1 μN of force.

Figure 3. Genetic manipulation of cuticle proteins alters body mechanics, but not TRN morphology

Deleting dpy-5 or lon-2 alters (A) body morphology, and (B) body length and stiffness. Data are mean ± s.e.m. of at least 146 animals. * indicates p <0.01 (student’s t-test with a Bonferroni correction for 6 comparisons). (C) The coefficient of variation for a 10 touch stiffness measurement is similar for wild type and lon-2 animals, but smaller for dpy-5 animals. Boxes represent 25^th-75^th percentiles. Thin line is median, thick line is mean (n = 5 animals). * indicates p <0.05 (student’s t-test with a Bonferroni correction for 3 comparisons). (D) The position of anterior TRNs is similar in wild type, dpy-5 and lon-2 animals (not significant, student’s t-test with a Bonferroni correction for 6 comparisons). The micrograph shows CFP-labeled anterior touch receptor neurons (ALM, AVM) in a uIs558 animal. Points are mean ± s.e.m. of at least 4 animals. (E) MEC-4 inter-punctum interval is similar in wild type and dpy-5 mutants, but larger in lon-2 mutants. Micrograph shows MEC-4::YFP puncta in anterior TRNs in a uIs558 animal. Points are mean ± s.e.m. of at least 208 IPIs from the TRNs of at least 3 animals. * indicates p <0.05 (student’s t-test with a Bonferroni correction for 3 comparisons). Some error bars are smaller than data points.

Figure 4. Genetic manipulation of cuticle proteins alters touch sensitivity

(A) F_1/2 differs in dpy-5 animals with respect to wild type, but is similar among lon-2 and wild type animals. (B) Δ_1/2 does not vary in dpy-5 or lon-2 animals with respect to wild type. Data are mean ± s.e.m. of the probability of response of a population of at least 7 animals at each force or indentation level. Wild-type data are shown with dashed lines for clear differentiation. Lines are exponential fits with a horizontal offset of the form p(x) = be^{(x–c) / a}, weighted by the inverse variance of each force or indentation sample. Finely dashed vertical lines highlight F_1/2 and Δ_1/2 for each strain; values and their errors are given in Table 1. * indicates p <0.05 for differences between F_1/2 or δ_1/2 (student’s t-test with Bonferroni correction for 3 comparisons).

Figure 5. Optogenetic induction of body stiffening decreases touch sensitivity

(A) Blue light triggers muscle hypercontraction BWM::ChR2, shortening and stiffening the body. Data are mean ± s.e.m. (n = 20). * indicates p <0.05 (student’s t-test with Bonferroni correction for 2 comparisons). (B) Muscle hypercontraction sometimes (top row), but not always (bottom row), reduces sensitivity to defined forces. Body shortening results in a reduction in the height of the curvature matrix in the second column. (C) BWM::ChR2 animals experience an increase in F_1/2 and (D) δ_1/2 when exposed to blue light. Data are mean ± s.e.m. of at least 3 individual animal response probabilities at each force level. Lines are exponential fits with a horizontal offset of the form p(x) = be^{(x–c) / a}, weighted by the inverse variance of each force or indentation sample. Finely dashed vertical lines highlight F_1/2 and δ_1/2 for each condition; values and their errors are given in Table 1. * indicates p <0.05 for differences between F_1/2 or δ_1/2 (student’s t-test). (E) The responses of BWM::ChR2 animals have similar latency in both the rest and muscle-hypercontracted states. Boxes represent 25^th-75^th percentiles. Thin line is median, thick line is mean. +’s indicate outliers.

Figure 6. Body stiffening significantly increases the force required for half-maximal response probability to touch, but only moderately increases the required indentation

F_1/2 increases with body stiffness under all conditions examined in this study (top: line is a linear fit to log(F_1/2) vs stiffness, k. R² = 0.84. Statistical significance, intercept: p = 0.023, slope: p = 0.028, two-sided linear regression t-test). By comparison, half-maximal indentation, δ_1/2, is less dependent on stiffness, k (bottom: R² = 0.63. Statistical significance, intercept: p = 0.09, slope: p = 0.11, two-sided linear regression t-test). Data are mean ± s.e.m. (see Table 1). Some error bars are smaller than data points.