Modulation of C. elegans touch sensitivity is integrated at multiple levels

Abstract

Sensory systems can adapt to different environmental signals. Here we identify four conditions that modulate anterior touch sensitivity in Caenorhabditis elegans after several hours and demonstrate that such sensory modulation is integrated at multiple levels to produce a single output. Prolonged vibration involving integrin signaling directly sensitizes the touch receptor neurons (TRNs). In contrast, hypoxia, the dauer state, and high salt reduce touch sensitivity by preventing the release of long-range neuroregulators, including two insulin-like proteins. Integration of these latter inputs occurs at upstream neurohormonal cells and at the insulin signaling cascade within the TRNs. These signals and those from integrin signaling converge to modulate touch sensitivity by regulating AKT kinases and DAF-16/FOXO. Thus, activation of either the integrin or insulin pathways can compensate for defects in the other pathway. This modulatory system integrates conflicting signals from different modalities, and adapts touch sensitivity to both mechanical and non-mechanical conditions.

Keywords: insulin signaling; integrin signaling; long-term sensitization; mechanosensation; sensory modulation.

Figure 1.

Sensitization to touch by vibration. A, Anterior (red) and posterior (blue) response (mean ± SEM) of wild-type to touch after sustained vibration for the indicated time. P < 0.005 comparing responses with 0 h (*) or 1.5 h (**). N ≥ 3 for each time point. For all figures, N represents the number of independent sets of animals tested, each with at least 10 animals, and n represents the number of animals. B–D, Anterior touch sensitivity (mean ± SEM) in animals vibrated at various frequencies, strengths, and times. The enhancement of touch sensitivity by prolonged vibration was greatest when animals were (B) vibrated for more than 2 h at (C) a frequency of 50 Hz with (D) an average acceleration >1 × g. Values for anterior sensitivity are given as mean ± SEM. The data for no vibration (−) and 2 h, 50 Hz, 1.5 × g vibration from all the experiments were pooled and used for all three figures. Optima of the other parameters were used when testing a particular parameter; *p < 0.02, **p < 0.005, ***p < 0.001, N ≥ 3 for all conditions tested. E, Normalized calcium responses of control (black), vibrated (red) animals, and animals recovered from vibration (blue) and their corresponding Boltzmann fits; n ≥ 6 for all strains. Error bars represent SEM of responses at each displacement. F, Mean ± SEM of the D₅₀ and k values of the indicated cells in wild-type animals. G, Sample calcium responses (blue) from control (top) and vibrated (bottom) animals. The GCaMP3 fluorescence levels are shown with arbitrary units (au). The displacement of each stimulus (black cross, in micrometers) is marked at each peak. Arrows indicate calcium peaks without stimuli. H–K, Statistics of calcium responses in AVA and AVD neurons with (white, V) or without (black, C) sustained vibration. Maximum fluorescence changes (H) were shown instead of ΔF/F₀ because baseline GCaMP3 fluorescence in the AVA neurons was reduced after vibration (J), complicating the interpretation of ΔF/F₀. This reduction of baseline fluorescence was likely due to a change in the baseline calcium level because antibody staining against GCaMP3 in vibrated animals showed no change in the amount of GCaMP3 expressed (K). D₁₀ was shown (I) instead of D₅₀ because estimation of the D₁₀ values are less sensitive to the fast habituation in these cells.

Figure 2.

Modulation of touch sensitivity by stress and vibration. A, Anterior touch sensitivity (mean ± SEM) in control animals (−), dauer larvae (dauer), or animals subjected to sustained vibration (vib), tapping (tap), or sustained channelrhodopsin-2 activation (ChR2). Additionally animals were treated with 230 mm NaCl (NaCl), 230 mm NaCl with 2 mm amiloride (NaCl + Amil), or 1% O₂ (hyp). The amiloride-treated animal contained a bus-17 mutation; **p < 0.005 compared with control, and *p < 0.05, ***p < 0.005 compared with the respective control under each condition, N ≥ 3 for all strains. B, Anterior touch response (mean ± SEM) of wild-type animals at the noted time points after they were transferred from 50 mm NaCl to 230 mm NaCl (blue) or from 230 mm NaCl to 50 mm NaCl (red). N ≥ 4 for all time points; *p < 0.01, **p < 0.005. C, Anterior touch response (mean ± SEM) of wild-type animals grown on NGM plates with the specified concentration of NaCl or with 50 mm NaCl and 380 mm sucrose; N = 3, *p < 0.05. D, Anterior touch response (mean ± SEM) of wild-type animals grown on NGM plates supplemented with 180 mm of the specified salts or 380 mm sucrose; N ≥ 6, *p < 0.05, **p < 0.005.

Figure 3.

Effect of focal adhesion mutants on TRN function. A, The expression of pat-2p::gfp or unc-112p::unc-112::gfp in ALM, AVM, and PLM TRNs of animals fed bacteria containing dsRNA against gfp. Because of the inefficient systemic RNAi in the nervous system, gfp expression in the body-wall muscle, but not in the neurons, is reduced, allowing visualization of GFP in the TRNs. Other focal adhesion proteins have been shown to be expressed in the TRNs and the vertebrate hair cells (Gettner et al., 1995; Hobert et al., 1999; Littlewood Evans and Muller, 2000; Mackinnon et al., 2002; Lin et al., 2003). B, Anterior (A, black) and posterior (P, white) responses (mean ± SEM of responses of individual animals) of mosaic animals with (+) or without (−) the rescuing arrays of the indicated genes in the TRNs. For the focal adhesion genes, n > 20 for anterior responses, and n > 15 for posterior responses. For mec-4, n > 10. For all anterior responses, p < 0.005 between (+) and (−) animals. C, Response (mean ± SEM of responses of individual animals) to three light pulses from focal adhesion mosaic animals lacking the rescuing arrays of the indicated genes in the TRNs but expressing channelrhodopsin-2 (ChR2) in the TRNs and from egl-19 animals expressing channelrhodopsin-2 in the TRNs; n ≥ 20 for all strains tested, *p < 0.05 compared with the wild-type. D, ALM processes (green) and the body-wall muscle (red) in unc-112 mosaic animals with or without unc-112 in the ALM cells. The ALM process is normally ensheathed by the hypodermis, which separates the ALM process from the body-wall muscle in adults. ALM processes that are not ensheathed appear adjacent to the body-wall muscle. The ALM cell body would also be pressed against the body-wall muscle, assuming a half-circle shape instead of the normal raindrop shape. The PLM processes had similar defects, yet retained the touch sensitivity, suggesting that the ensheathment defect alone did not cause touch insensitivity. E, Fractions of TRNs showing ensheathment defects (black) or migration defects (white); n > 15 for ensheathment data and n ≥ 20 for migration data; *p < 0.002 compared with wild-type. The migration defect seen in pat-3 animals could not account for the reduced touch sensitivity either, because animals lacking the second C. elegans α-integrin gene, ina-1, were touch sensitive despite having similar migration defects (Baum and Garriga, 1997; data not shown). F, Fractions of ensheathment-defective ALM cells in TU3595 animals fed with neuron-enhanced RNAi against gfp, mec-4, unc-112, pat-6, or mec-1. See Materials and Methods for detailed scoring standard. N ≥ 3, *p < 0.05 compared with gfp RNAi control. The RNAi-treated animals, except for the gfp control, had reduced touch sensitivity, suggesting that the ensheathment defect cannot solely account for the reduced touch sensitivity.

Figure 4.

The calcium response to touch in focal adhesion mutants. A, Normalized calcium response (mean ± SEM at each given displacement) of wild-type, pat-2 mosaic animals, pat-2 mosaic animals subjected to sustained vibration, unc-112 mosaic animals, and egl-19 animals and their corresponding Boltzmann fits using all data points for a particular strain; n ≥ 6 for all strains. Wild-type data is reused from Figure 1B. B, Maximum calcium response (Max ΔF/F₀) of ALM neurons to saturated stimulation from wild-type, pat-2 mosaic, unc-112 mosaic, and egl-19 animals; *p < 0.05 compared with wild-type, n ≥ 6 for all strains. C, Calcium responses (mean ΔF/F₀ ± SEM) of cultured ALM cells from wild-type, unc-112, pat-2, and egl-19 animals to potassium depolarization; *p < 0.05 compared with wild-type response, n ≥ 9 for all groups. ALM neurons from both pat-2 and egl-19 mutants, either in vivo or cultured, gave a decreased maximum calcium signal without a decrease in sensitivity (D_50(egl-19) = 1.0 ± 0.3 μm, n = 5, k = 2.0 ± 0.6), indicating that the maximum calcium response and sensitivity do not depend on each other. These results suggest that pat-2 and egl-19 mutations changed the calcium response independently of mechanotransduction and of UNC-112. Therefore, PAT-2 additionally modulates the calcium response.

Figure 5.

Insulin signaling affects touch sensitivity. A, Anterior (A, black) and posterior (P, white) response (mean ± SEM) of animals with the indicated phenotypes; *p < 0.005 compared with wild-type, N ≥ 3 for all strains. B, smFISH of akt-1 (left) or akt-2 (right; red), and mec-18 (green) in wild-type (+) or the respective null (null) mutants. The positions of the mec-18 dots delineate the shape of the ALM cell body. C, Anterior (black) and posterior (white) response (mean ± SEM) of daf-2(e1370), daf-2(e979), and daf-2(m41) animals; N = 3, *p < 0.001 compared with wild-type. D, Anterior touch response (mean ± SEM) of daf-2(e1370), daf-16(m26); daf-2(e1370), and daf-16(mgDf50); daf-2(e1370) animals; *p < 0.05, **p < 0.0005 compared with daf-2 alone, N ≥ 3 for all strains.

Figure 6.

Integrin and insulin signaling converge to affect touch sensitivity. A, Anterior response of daf-2(m65), pdk-1, and akt-1 animals with (white) or without (black) added focal adhesion (FA) proteins UNC-112 and PAT-6 in the TRNs; *p < 0.005 with or without UNC-112/PAT-6, N ≥ 3. B, Effect of akt-1(gf), pdk-1(gf), and daf-16 on the anterior touch response of unc-112 or pat-2 mosaic animals. Values are the means ± SEM of responses of individual animals; n > 10, *p < 0.005 compared with unc-112 or pat-2 alone. C, Anterior touch sensitivity of animals with the indicated phenotypes grown with (V, white) or without (C, black) vibration; *p < 0.05, **p < 0.002 with or without vibration, N ≥ 3 for all strains. D, Anterior (A, black) and posterior (P, white) touch response of wild-type, him-4 animals, and him-4 animals carrying mec-3p::pdk-1(gf); *p < 0.005 compared with wild-type anterior response and p < 0.05 compared with him-4 + mec-3p::pdk-1(gf); N ≥ 3.

Figure 7.

Regulation of touch sensitivity through insulin-like peptides. A, The anterior touch sensitivity (mean ± SEM) of ins-10(i) animals, ins-10(i) animals carrying either wild-type copies of ins-10 with different codons or an rde-1 mutation, ins-10(tm3498) animals with or without wild-type copies of ins-10, and ins-22(tm4990) animals with or without wild-type copies of ins-22. N > 3 for all strains tested; *p < 0.01 between the indicated strains, and comparing ins-10(i), ins-10, and ins-22 to wild-type. ins-10(i) had no effect on touch sensitivity in an rde-1 background, which abolishes RNAi. Loss of INS-33, which is needed for larva development(Hristova et al., 2005), also resulted in a weak reduction of touch sensitivity in some abnormally developed animals (data not shown). INS-33, however, probably acts generally, since ins-33 larvae that had normal growth were touch sensitive (data not shown). In addition, the difference between ins-33 and wild-type animals was statistically insignificant after Bonferroni correction. B, ins-10p::gfp expression in well fed and starved animals. The green image (GFP), the red image (myo-2p::mCherry), and the DIC image were merged. Strong GFP expression was seen throughout larval development in M4 and I5 and weakly and sporadically in the MC pharyngeal interneurons, RIS interneurons, and an additional pair of nerve ring interneurons tentatively identified as either the RMF cells or RMH cells. C, Anterior touch response (mean ± SEM of individual animals) of animals with no ablation (−), mock ablation (mock), or with I5 and/or M4 ablated; n ≥ 12 for all, *p < 0.001. D, Expression of ins-10p::gfp or ins-10p::rfp in well fed L3 animals, starved L3 animals, dauer larvae, or (E) adult animals in normoxic or hypoxic conditions. Images of myo-2p::mcherry and ceh-22p::gfp expression in the pharyngeal muscles were merged with the green ins-10p::gfp (D) and red ins-10p::rfp (E). F, G, The intensities (mean ± SEM of individual animals) of ins-10p::gfp or ins-10p::rfp in the M4 and I5 neurons, ceh-22p:gfp or myo-2p::mCherry in the pharynx, and mec-3p::rfp in the TRNs in (F) starved L3 or dauer larvae and in (G) adult animals under normoxic or hypoxic conditions; n ≥ 5 for all conditions.

Figure 8.

INS-10 regulates touch sensitivity through insulin signaling. A, Normalized calcium responses (mean ± SEM) of wild-type (black) and ins-10(i) (red) animals with different probe displacements, and their corresponding Boltzmann fits; n ≥ 16 for both strains. The wild-type data were reused from Figure 1B. B, Anterior touch response (mean ± SEM) of ins-10(i) animals with or without akt-1(gf), pdk-1(gf), mec-3p::pdk-1(gf), or mec-17p::unc-112::gfp and mec-17p::pat-6::gfp; p < 0.005 for all other strains compared with ins-10(i) alone, N ≥ 3. C, Anterior touch sensitivity (mean ± SEM) of wild-type, daf-16, or akt-1(gf) animals, or animals carrying mec-3p::pdk-1(gf) grown under normoxic (N, black) or hypoxic (H, white) conditions; N ≥ 4 for all strains and conditions tested, p < 0.005 comparing hypoxic wild-type to other hypoxic strains. D, Anterior touch sensitivity (mean ± SEM) of wild-type starved L3 larvae, wild-type dauer larvae, or dauer larvae expressing mec-3p::pdk-1(gf); N ≥ 3, *p < 0.001, **p < 0.0005 compared with wild-type dauer larvae.

Figure 9.

The ASE neurons and INS-22 regulates touch sensitivity. A, Anterior touch response (mean ± SEM) of ins-22 animals, ins-22 animals expressing mec-3p::pdk-1(gf), and daf-16; ins-22 animals; *p < 0.05 compared with ins-22, N ≥ 3. B, Expression of ins-22p::gfp in animals grown on high sucrose or high salt (NaCl). Images of myo-3p::mcherry were merged with green ins-22p::gfp images. C, Anterior touch response (mean ± SEM) of wild-type and che-1 animals with or without mec-3p::pdk-1(gf); N ≥ 3, p < 0.01 between che-1 and the other two strains. D, Quantification of ins-22p::gfp in animals grown on control NGM plates and NGM plates supplemented with 180 mm NaCl or 380 mm sucrose; *p < 0.05 compared with control and p < 0.005 compared with sucrose, n > 15 for all samples. E, Anterior touch response (mean ± SEM) of the indicated animals grown with 50 mm NaCl (black, C) or with 230 mm NaCl (white, S); *p < 0.05 and **p < 0.01 comparing the respective strain on high salt to wild-type grown on high salt, ***p < 0.01, N ≥ 3 for all strains tested. AU, arbitrary units.

Figure 10.

Behavioral consequences of mechanosensory modulation. A, The number of animals responding to a pulse vibration by backward movement (black) or nonbackward movement (forward or no movement, white) in wild-type and ins-10(i) animals. The data are pooled from four independent trials; p < 0.0001 comparing ins-10(i) to wild-type. B, The ratio of chemotaxis efficiency index (mean ± SEM; CEI, the fractions of animals that have reached the diacetyl spot) with or without tapping every 30 s during a 12 min chemotaxis assay; *p < 0.05 comparing animals with or without TRN::pdk-1(gf), or comparing ins-10(i) or wild-type dauer with wild-type adults, N ≥ 3 for all strains. C, The ratio of CEI (the fractions of animals that have reached the diacetyl spot; mean ± SEM) with or without tapping. The data used are the same as in B. D, Fraction of animals moving backward in response to a 0.5 s vibratory pulse without sustained vibration (no vib) or with background sustained vibration for the indicated time. The total number of animals tested is noted at each point; *p < 0.01 compared with no vib and 2 h using Fisher's exact test. E, F, Anterior (E) and posterior (F) response (mean ± SEM) of wild-type (blue) or akt-1 (red) animals to touch before habituation (pre) or after vibration for more than 2 h and rested for the indicated amount of time, normalized to the response before habituation; *p < 0.005 comparing wild-type and akt-1 responses, N > 10 for all time points for each strain.

Figure 11.

Regulation of TRN mechanosensation. Sustained vibration (green pathway) sensed by integrins, and hypoxia, dauer formation, and high salt (red pathway) sensed by insulin peptides, converge on AKT-1and DAF-16 to modulate touch sensitivity in the anterior TRNs. Dashed connectors indicate that the signaling strength is attenuated by the corresponding signals.