Abstract
Dendrites are crucial for receiving information into neurons. Sensory experience affects the structure of these tree-like neurites, which, it is assumed, modifies neuronal function, yet the evidence is scarce, and the mechanisms are unknown. To study whether sensory experience affects dendritic morphology, we use the Caenorhabditis elegans’ arborized nociceptor PVD neurons, under natural mechanical stimulation induced by physical contacts between individuals. We found that mechanosensory signals induced by conspecifics affect the dendritic structure of the PVD. Moreover, developmentally isolated animals show a decrease in their ability to respond to harsh touch. The structural and behavioral plasticity following sensory deprivation are functionally independent of each other and are mediated by an array of evolutionarily conserved mechanosensory amiloride-sensitive epithelial sodium channels (degenerins). Our genetic results, supported by optogenetic, behavioral, and pharmacological evidence, suggest an activity-dependent homeostatic mechanism for dendritic structural plasticity, that in parallel controls escape response to noxious mechanosensory stimuli.
Significance Statement Little is known about how neuronal activity and sensory experience influence the structure and function of dendritic trees. Furthermore, how dendritic structures affect neuronal functions remains to a large extent a mystery despite their fundamental functions in neuronal sensory and synaptic activities in health and diseases. Here we show that complex dendritic trees of the nociceptive and proprioceptive PVD neuron in C. elegans display a dynamic nature where they undergo pronounced dendritic modifications during adulthood. These modifications are determined by the received sensory signals generated by conspecific worms on the plate. We revealed functions for degenerins/Epithelial sodium channels in this phenomenon, using genetic and optogenetic approaches. We found that the degenerin MEC-10 acts cell autonomously to translate environmental mechanical signals into structural and functional modifications in the PVD.
Introduction
The general structure of the nervous system has been known for over a century. Groundbreaking studies on synaptic plasticity and its underlying mechanisms have shown that before birth and in adult animals, neuronal activity is needed for synaptic remodeling (Fox & Wong, 2005, Goodman & Shatz, 1993, Katz & Shatz, 1996, Wiesner, Bilodeau et al., 2020, Zuo, Yang et al., 2005). In contrast, the molecular mechanisms responsible for structural remodeling of dendritic trees, as a result of different sensory inputs (experience), especially during adulthood, are less understood (Kolb & Whishaw, 1998, Tavosanis, 2012, Wong & Ghosh, 2002).
Mechanistic understanding of experience-dependent structural plasticity is primarily focused on activity sensation by calcium channels and N-methyl-D-aspartate (NMDA) receptors. These are known to induce downstream signaling cascades affecting, among others, the Rho family of small GTPases, calcium metabolism and microtubule stability (Ghiretti, Moore et al., 2014, Sin, Haas et al., 2002, Vaillant, Zanassi et al., 2002, Zhou, Hong et al., 2006). Several neurological conditions including autism, Down syndrome, fragile X syndrome, and schizophrenia are characterized by abnormal dendritic spine structures (Hu, Huang et al., 2020, Huebschman, Corona et al., 2020, Jan & Jan, 2010, Tendilla-Beltran, Antonio Vazquez-Roque et al., 2019). Uncovering the molecular basis of dendritic tree instability during development and adulthood, may shed light on neurological disease mechanisms and elucidate their behavioral phenotypes.
The dendrite morphology of the Caenorhabditis elegans’ PVD bilateral neurons is composed of repetitive, stereotypical and spatially organized structural units that resemble candelabra (Fig. 1A), making it a useful platform to study dendritic morphogenesis (Oren-Suissa, Hall et al., 2010). While some of the genetically-programmed molecular mechanisms responsible for the morphogenesis and regeneration of PVD’s dendritic trees are known (Dong, Liu et al., 2013, Dong, Shen et al., 2015, Inberg, Meledin et al., 2019, Kravtsov, Oren-Suissa et al., 2017, Oren-Suissa, Gattegno et al., 2017, Oren-Suissa et al., 2010, Salzberg, Diaz-Balzac et al., 2013, Salzberg, Ramirez-Suarez et al., 2014, Smith, Watson et al., 2010), the influence of nurture, e.g. sensory experience, on its structure and function during development and in adulthood remain unexplored. The PVD mediates three sensory modalities (reviewed in (Goodman & Sengupta, 2019)): response to harsh mechanical stimuli (nociception) (Chatzigeorgiou, Yoo et al., 2010), response to low temperatures (Chatzigeorgiou et al., 2010) and proprioception (Albeg, Smith et al., 2011, Tao, Porto et al., 2019). While the PVD response to low temperatures is mediated by transient receptor potential (TRP) channels (Chatzigeorgiou et al., 2010), nociception and proprioception are mediated by degenerins/epithelial Na+ channels (DEG/ENaCs) expressed in the PVD (Chatzigeorgiou et al., 2010, Husson, Costa et al., 2012), which form homo- and hetero-trimers and are involved in force sensing. In mammals, some DEG/ENaCs such as ASIC1a participate in synaptic plasticity and cognitive functions such as learning and memory (Baldin, Barth et al., 2020, Bianchi & Driscoll, 2002, Chen, Bharill et al., 2015, Gillespie, 2001, Gobetto, Gonzalez-Inchauspe et al., 2021, Hill & Ben-Shahar, 2018, Mango & Nistico, 2020, Welsh, Price et al., 2002). As sensory and social isolation affect the behavior and fitness of diverse animals (Bailey & Moore, 2018, Kuhlman, O’Connor et al., 2014, Wilbrecht, Holtmaat et al., 2010, Yu & Zuo, 2011), including primates (Harlow, Dodsworth et al., 1965), studying adult nematode somatosensory neurons can reveal possibly-conserved mechanisms of dendritic plasticity, which are induced by sensory stimuli.
(A) The PVD neuron dendritic tree, cell body (CB) and axon. The red arrow corresponds to the contact point with a platinum wire during posterior harsh touch, while the black arrow represents the behavioral escape response of the worm. The red dashed square represents the analyzed region around the CB. One representative candelabrum is colored by branch orders: blue primary (1°), purple secondary (2°), cyan tertiary (3°) and green quaternaries (4°). A- anterior; P- posterior; V- ventral; D-dorsal. (B) Schematic of the isolation protocol followed at 72 h by posterior harsh touch assay or PVD imaging of adult worms. (C) Isolation of embryos reduced the percentage of worms responding to harsh touch at adulthood. Crowded-black bars, Isolated-red bars. Wildtype (WT) N2 worms (Crowded, n=32; Isolated, n=32), mec-4 (Crowded, n=52; Isolated, n=46), mec-3 (Crowded, n=12; Isolated, n=12). mec-4 animals were assayed as adults after 96 h to account for their slower rate of growth. (D) Growth in plates with chemical cues from adult hermaphrodites did not alter the reduced response rate of isolated mec-4 animals to harsh touch (Crowded, n=33; Isolated, n=28). (E) Harsh touch response in crowded and isolated conditions for mutants of different DEG-ENaCs and the TRP channel gtl-1. WT worms (same set of worms as in Fig. 1C. Crowded, n=32; Isolated, n=32), asic-1 (Crowded, n=46; Isolated, n=30), mec-10 (Crowded, n=38; Isolated, n=31), degt-1 (Crowded, n=37; Isolated, n=31), unc-8 (Crowded, n=18; Isolated, n=15), del-1 (Crowded, n=31; Isolated, n=27), gtl-1 (Crowded, n=27; Isolated, n=25). (F) MEC-10 expression in the PVD rescues mec-10 mutants’ crowded-specific reduction in response to harsh touch. All strains contain the ser-2Prom3::Kaede PVD marker construct, and were tested in the crowded conditions. (WT, n=42; mec-4, n=17; him-5, n=34 (him-5 was used as WT background for several strains after cross); him-5;mec-10, n=35; him-5;mec-10;ser2Prom3::MEC-10, n=33). (G) Isolation for 24 h in adulthood did not affect the response to harsh touch. Worms grown under crowded conditions were isolated for 24 h as young adults and compared against their crowded age-matched counterparts in their response to harsh touch (Crowded, n=25; Isolated for 24 h, n=26). The proportion of responding worms (percentage) ± standard error of proportion is shown. Fisher’s exact test, *p<0.05, **p<0.01, ***p<0.001, n.s. not significant.
In gentle touch circuits, which are distinct from nociception (Chalfie & Sulston, 1981, Li, Kang et al., 2011, Oren-Suissa et al., 2010), Rose and colleagues (Rose, Sangha et al., 2005) found that deprivation of mechanosensory stimulation generated by colliding conspecifics in the growing plate, resulted in modified glutamatergic signaling and reduced response to tap stimulation. Here, we adapted this mechanosensory stimulation paradigm (Rose et al., 2005), where the crowded worms are thought to represent the natural ‘default’ enriched mechanosensory state, to look into nociceptive circuits and identify the mechanism that couples mechanosensory experience to structural and functional dendritic plasticity. We focused on how mechanosensory experience, perceived through DEG/ENaCs, affects structural plasticity of the PVD dendritic trees in adult C. elegans, and whether this entails behavioral consequences. We find that mechanosensory experience not only alters the PVD’s dendritic structure in the adults, but also affects associated behavioral outputs. However, and in contrast to the prevalent hypothesis, these structural and behavioral properties are not correlated.
Results
Sensory isolation induces behavioral plasticity
Rose and colleagues (Rose et al., 2005) has shown that isolation of wildtype (WT) C. elegans causes a decreased response to gentle touch circuits when comparing to worms raised in a crowded setting. We utilized a similar paradigm to study nociception, a behavior associated with PVD neuron activation (Oren-Suissa et al., 2010, Way & Chalfie, 1988), using a behavioral assay that registers escape following prodding with a platinum wire (harsh touch assay; Fig. 1A) (Chalfie & Sulston, 1981, Li et al., 2011, Oren-Suissa et al., 2010). To study whether mechanosensory deprivation affects the nociceptive functions of the PVD, we isolated embryos into single plates where they grew 72 h to adulthood and compared their behavioral response to harsh touch against same-aged adults that were grown on crowded plates (Fig. 1B). We found that ~40% of isolated WT animals responded to harsh touch, compared with ~80% of animals grown in crowded plates (Fig. 1C). To test whether gentle touch neurons are involved in this behavioral difference, we studied mec-4(e1611) mutants, in which gentle touch neurons are degraded (Caneo, Julian et al., 2019, Driscoll & Chalfie, 1991, Hedgecock, Sulston et al., 1983, Li et al., 2011, Suzuki, Kerr et al., 2003), and obtained similar results (Fig. 1C). Given that mec-4(e1611) exhibits the same experience-dependent behavioral plasticity as WT, we conclude that this phenomenon is independent from the mec-4 gentle-touch neurons. As a negative control, we used mec-3(e1338) mutants, that are harsh-touch-insensitive (Way & Chalfie, 1988) and found no responses for both groups (Fig. 1C). Thus, isolation reduces the response to noxious stimuli, in a process that is independent of gentle touch circuits.
To determine whether the effect of isolation on nociception is related to the mechanosensory deficit itself (like friction (Shi, Luke et al., 2016) and collisions (Chatzigeorgiou et al., 2010), in contrast to chemosensory stimuli, we used the mec-4 strain to compare responses of isolated worms to responses of isolated worms grown in the presence of glass beads (resembling a method used by (Sawin, Ranganathan et al., 2000)). Worms grown in isolation with beads had a similar response compared to animals grown in crowded plates (Fig. S1A). While the beads are sufficient to increase isolated animals’ responses, they do not fully recapitulate the crowded state (Fig. S1A). To study whether the effect was chemosensory-mediated, we tested plates prestimulated with pheromones as well as osm-6 mutants (defective in olfactory sensory cilia (Collet, Spike et al., 1998)). We found that isolation induced a reduction in the response to harsh touch, regardless of plate ‘odor’ (Fig. 1D) or chemosensory function (Fig. S1B). We note that the chemosensory defective osm-6 mutants also have a defective ciliated mechanosensory PDE neuron (Collet et al., 1998) (which mediates posterior harsh touch responses (Li et al., 2011)). These results suggest that adult worms display a behavioral plasticity in PVD-related harsh touch circuits, which is dependent on mechanosensory experience, and is independent of the PDE neuron, olfactory and gentle touch neurons.
MEC-10 and other mechanosensory channels mediate isolation-induced behavioral plasticity before adulthood
To study the genetic mechanisms behind the nociceptive response plasticity, we performed a candidate gene screen for degenerins (DEG/ENaC) and transient receptor potential (TRP) channels that are expressed in the PVD, where degenerins mediate mechanosensation and proprioception, while TRPs sense low temperatures. In all these signal transduction pathways some of the outcomes are behavioral responses (Chatzigeorgiou et al., 2010, Huang & Chalfie, 1994). We found that the WT-like isolation-induced reduction in harsh touch response was also present in degt-1 DEG/ENaC and gtl-1 TRP channel mutants, suggesting that these channels are not directly involved in behavioral plasticity following isolation (Fig. 1E). In contrast, for del-1, asic-1 and mec-10 DEG/ENaC mutants the difference between isolated and crowded conditions was undetectable, indicating that they are required for such behavioral plasticity. Interestingly, while the harsh touch response of asic-1 mutants was consistently high, and similar to crowded WT worms, the response for mec-10 mutants was consistently low, similar to isolated WT animals, regardless of sensory experience (Fig. 1E). To test whether the response to harsh touch is dependent on DEG/ENaC activity, we used the DEG/ENaC blocker amiloride (Ben-Shahar, 2011, Bianchi & Driscoll, 2002) on WT worms that were grown in crowded plates. We found that the response to harsh touch was not affected by continuous growth in the presence of amiloride (Fig. S1C). This result possibly supports the idea that different amiloride-sensitive epithelial sodium channels may have positive and negative effects on the response to harsh touch, which are masked by global inhibition (Fig. 1E). This is further evident by the plasticity consequences of some double and triple DEG/ENaC mutant combinations, which are difficult to align into an epistatic genetic model (Fig. S8).
The DEG/ENaC MEC-10 is expressed in the PVD and responds to mechanosensory signals (Chatzigeorgiou et al., 2010) including shear forces (Shi et al., 2016, Shi, Mutchler et al., 2018). Since crowded mec-10 mutants seem isolated-like in their behavioral response (Fig. 1E), we asked how mec-10 mediates the behavioral plasticity in crowded conditions. To test whether the activity of MEC-10 is required autonomously in the PVD, we expressed MEC-10 under a PVD-specific promoter in a mec-10 mutant background (Oren-Suissa et al., 2010, Tsalik, Niacaris et al., 2003). We found that expression of MEC-10 in the PVD rescues the low response to harsh touch in crowded mec-10 mutants, indicating that it acts cell autonomously to modulate behavioral plasticity (Fig. 1F).
To determine whether isolation affects the response to harsh touch during development or in adults, we isolated crowded-raised young adults for 24 h and found no difference in their response to harsh touch (Fig. 1G). Thus, isolation-induced reduction in mechanosensation is determined before young adults, is PVD cell-autonomous, and mec-10-dependent.
Experience affects PVD morphology via mec-10 in adulthood
Sensory experience (such as stimuli deprivation) is known to drive synaptic plasticity in the nervous system (Alvarez & Sabatini, 2007, Fox & Wong, 2005, Zuo et al., 2005). However, little is known about the effects of mechanosensory deprivation on the architecture of sensory neurons. To study whether reduced mechanosensory experience can alter the dendritic structure of the PVD we followed the isolated growth paradigm (Fig. 1B) and examined two morphological features of the PVD (Fig. 2A): The fraction of ectopic (excessive, disordered) branches out of non-ectopic branches (those that form the ‘ideal’ candelabrum), and the percentage of straight 4th order (quaternary) branches, which form the terminal processes of the candelabrum.
(A) Schematic representation of the PVD dendritic structure, marking the morphological characteristics of interest: ectopic branches and quaternary branching geometry. Dashed lines represent ectopic branching at each order. Colors correspond to Fig. 1A. (B-G) Both isolation and mec-10 affect the structure of the PVD: (B) Isolation increases the fraction of ectopic branching, mec-10 mutation increases ectopic branching in the crowded state. (C) Isolation decreases the percentage of straight quaternary branches, mec-10 mutation decreases the percentage of straight quaternary branches in the crowded state. Crowded-black dots, isolated-red dots; WT (Crowded, n=28; Isolated n=26), him-5 (Crowded, n=27; Isolated, n=25; him-5 was used as WT background for several strains after cross), him-5;mec-10 (Crowded, n=30; Isolated, n=24). (D) Expression of mec-10 in the PVD on the background of him-5;mec-10 in the crowded state reduces ectopic branching and (E) increases the percentage of straight quaternary branches. (F-I) Representative PVD images of WT, mec-10 and mec-10;PVD::MEC-10 in different growth conditions. (Scale bar, 50 μm). (J) Embryo isolation with glass beads did not affect ectopic branching. (K) Embryo isolation with glass beads increased the percentage of straight quaternary branches (Crowded, n=28; Isolated n=26; Isolated with beads, n=22. Crowded and isolated WT are the same set of worms as in Fig. 2B,C). Each dot represents a single worm. The mean ± s.e.m. are shown in blue. Box plot with median and hinges for the first and third quartile. The whiskers represent an estimated calculation of the 95% confidence interval. Kruskal-Wallis test, # # p<0.01, # # # p<0.001, Mann Whitney test with Bonferroni Correction α=0.0167. *p<0.05, ** p<0.01, *** p<0.001, n.s. not significant.
When quantifying the branching pattern of the PVD in adulthood, WT (or WT-like him-5 background) animals which were isolated as embryos showed an increase in ectopic branching compared with crowded age-matched worms (Fig. 2B) and the quaternary branches assumed a more rippled shape (fewer straight 4ry branches, Fig. 2C), regardless of the neuron marker utilized (see Fig. S2).
To determine whether chemical stimulation plays a role in the observed morphological plasticity, animals were isolated onto pheromone-conditioned plates (Maures, Booth et al., 2014). Similarly to the response to harsh touch (Fig. 1D), chemical stimulation of the plates did not rescue the isolation-induced increase in the fraction of ectopic branching or the decrease in percentage of straight quaternary branches (Fig. S3). Additionally, we looked at mec-4 gentle-touch-insensitive mutants (Rose et al., 2005, Suzuki et al., 2003) and found that isolation caused a similar change to PVD structure as in the WT (Fig. S4, compare with Fig. 2B and 2C). In summary, these results suggest that mechanosensory experience controls morphological plasticity of PVD dendritic trees independently of mec-4 and chemical stimulation.
We next sought to determine whether DEG/ENaCs mediate these experience-driven morphological alterations, by examining PVD morphology in crowded and isolated mec-10 mutants. We found that crowded mec-10 animals were different from WT crowded worms, and appear more isolated-like in terms of morphological features (more ectopic branches and fewer straight quaternary branches, compared to crowded him-5 control animals; Fig. 2B and 2C). These results suggest that PVD dendritic morphology is affected by sensory experience in a mec-10-dependent pathway. Importantly, these results confirm a role for the mechanosensory channel MEC-10, but not other degenerins (such as ASIC-1), in the plasticity of both the dendritic structure and the nociceptive function of the PVD in the crowded background.
To study whether mec-10 acts cell autonomously to mediate morphological plasticity, we again utilized PVD-specific MEC-10 expression in a mec-10 mutant background and found it reduced the fraction of ectopic branches and increased the percentage of straight quaternary branches, compared to age-matched crowded mec-10 animals without the rescue array (Fig. 2D-I). Thus, MEC-10 acts in the PVD to mediate morphological plasticity.
Having identified an isolation-induced morphological plasticity in adults (observed at 72 h from embryo), we next sought to establish the temporal dependence of the effect, by comparing animals after only 48 h, as young adults (see Fig. S5A). Isolated worms showed a small but significant difference for the percentage of straight quaternary branches, but not in ectopic branching (Fig. S5B-E) when compared with age-matched crowded worms, suggesting that isolation induces some morphological alterations during morphogenesis, while some relate to adult-stage maintenance of the neuron. We then isolated animals grown for 48 h in crowded conditions, and found that isolation of young crowded adults for 24 h induced changes in both measured parameters (Fig. S5F-I), recapitulating isolation of embryos (Fig. 2B-C). Thus, isolation of adults for 24 h is sufficient to affect PVD dendritic tree architecture, in contrast to the developmentally-established behavioral response (Fig. 1G).
To further determine whether the morphological effect on PVD structure is solely mediated by mechanical cues, we used glass beads under isolated conditions. We found that while the presence of glass beads did not reduce the fraction of ectopic branching, it significantly increased the number of straight quaternary branches (compared with isolated animals without beads, Fig. 2J-K). Thus, the mechanosensory stimuli of inert beads can partially rescue isolation-induced morphological changes to the PVD.
Isolation triggers dynamic plasticity of the dendritic tree
To pinpoint the precise time interval required to mediate dendritic arborization changes in adulthood, we examined the effect of varying isolation times on the PVD morphology of crowded-raised young adults. We found that isolation of adults for 2 h had no significant consequence on the structure of the PVD, while isolation of adults for 5 h or longer, significantly increased the proportion of ectopic branches and reduced the percentage of quaternary straight branches compared to crowded worms (Fig. 3A-B).
(A) Isolation of crowded worms induces a time-dependent increase in the proportion of ectopic branches, and (B) a decrease in the number of straight quaternary branches (2 h: Crowded, n=24, Isolated, n=22; 5 h: Crowded, n=24, Isolated, n=27; 15 h: Crowded, n=3, Isolated, n=3; 24 h: Crowded, n=20, Isolated, n=21; animals imaged using polystyrene beads for immobilization, see Materials and Methods). (C) Isolation for 4 h induces more growth of new branches than retraction of existing ones. Ectopic branch dynamics were compared for crowded animals against their state after 4 h isolation. Imaging used 1% tricaine, see Materials and Methods. (D) Isolation of crowded worms for 4 h increases the ratio between ectopic/non-ectopic branches in individual animals and (E) reduces the percentage of straight quaternary branches (Crowded, n=6; Isolated n=6). (F) Crowded, but not isolated worms, show dynamic reduction in the percentage of straight quaternary branches during a 3 h time lapse movie in 1% tricaine; see Movies S1 and S2. (G) Representative time lapse frames of a crowded worm in 1% tricaine. Thick and thin arrows represent straight and non-straight quaternary branches, respectively. For (D, E) the lines connecting two points refer to the same crowded specimen at time 0 and after 4 h isolation. Box plot representation as in Fig. 2. For (A-B) Mann-Whitney test, * p<0.05,.*** p<0.001. For (C-E) Wilcoxon test (for two related samples), * p<0.05. The mean ± s.e.m. are shown for panels A-C, and F.
To further study plasticity at the individual level, we compared the dendritic tree of individual worms grown under crowded conditions before and after 4 h isolation. This revealed isolation induced more events of ectopic branch growth and fewer events of branch retraction (Fig. 3C). Furthermore, this period was sufficient to induce an increase in ectopic branching and a decrease in percentage of straight quaternary branches (Figs. 3D, 3E, and S6). Thus, 4 h of isolation are sufficient to induce changes in the architecture of the adult PVD.
Since the geometry of quaternary branches shows rapid response to the amount of mechanosensory stimulation (Fig. 3A-E), we next asked how dynamic is this experience-dependent morphological change. We performed a 3 h time lapse imaging in crowded and isolated animals anesthetized with 1% tricaine (Movies S1 and S2). We found that while the number of straight quaternary branches in the isolated worms remained low and stable (~8%), for the crowded animals, we observed a gradual reduction in straight quaternary branches, starting after ~2 h of isolation of individual worms on the slide (Fig. 3F,G).
In summary, isolation-induced morphological plasticity is apparent within 2-3 h from the switch from crowded conditions to growth in the absence of mechanical stimuli. This dynamic process involves simultaneous increase in ectopic branching and decrease in straight terminal branches.
Activity affects morphology
It has been shown that activity and sodium influx via the DEG/ENaC UNC-8 promotes synapse elimination in C. elegans (Miller-Fleming, Petersen et al., 2016). To test how global inhibition of DEG/ENaCs affects the morphology of the PVD neuron, we evaluated crowded worms that were grown on plates with amiloride compared to control worms. We found that blocking DEG/ENaCs by amiloride increased the fraction of ectopic branching and decreased the percentage of straight quaternary branches (Fig. S7A-D). Thus, unlike harsh touch response, which is independent of sensory-level global inhibition (Fig. S1C), global inhibition at the sensory level alters dendritic morphology. These results suggest that activity via degenerins modulates the structure of the PVD.
To test whether manipulations of other neuromodulators can affect the branching dynamics of the PVD, we utilized two anesthetics: tetramisole, which activates the nematode nicotinic acetylcholine receptors inducing muscle contraction (Aceves, Erlij et al., 1970) and tricaine, which modulates neuronal activity by blocking sodium channels (Katz, Chu et al., 2020) and analyzed their effects during 2-3 h of time lapse movies of crowded and isolated animals. We first tested the commonly used mixture of 0.1% tricaine and 0.01% tetramisole (the racemic mixture of the enantiomer levamisole) (Kravtsov et al., 2017) and found that it induces more growth than retraction of ectopic branches in crowded animals (Fig. S7E-G). In contrast, 1% tricaine alone has the opposite effect-more retractions of ectopic branches (Fig. S7H-I; Movies S1 and S2). In addition, 1% tricaine induces more growth for isolated compared to crowded worms, indicating that there are activity-dependent intrinsic factors that are different for the two experience states. These results suggest that modulating neuronal activity triggers structural modifications in the PVD, indicating a possible structural sensitivity to the amount of neuronal activity. In summary, we found that mechanosensory experience, DEG/ENaC protein presence and activity, and pharmacological targeting of activity dynamically affect the structure of the PVD.
PVD structure and function are independent
After establishing that mechanosensory experience induces both a behavioral and a structural plasticity of the PVD, we asked whether there is a causal link between the morphology of the dendritic tree of the PVD and its function as a nociceptor (Hall & Treinin, 2011). We followed the isolation protocol described in Fig. 1B for seven combinations of DEG/ENaC mutants and analyzed their response to harsh touch (Fig. S8) and their PVD structure (Fig. S9). In order to compare the morphological properties of the different DEG/ENaC genotypes under different sensory experience (crowded, isolated; Fig. S9), we used discriminant analysis. This provides a supervised classification method to combine all the analyzed morphological phenotypes, including loss of self-avoidance between adjacent candelabra (Fig S10), into a certain value characterizing the PVD morphological state (Fig. 4A). We then superimposed the behavioral results of the DEG/ENaC mutants’ response to harsh touch (Fig. S8) on the morphological clustering, as a binary-like property (<45% responding for isolated WT versus >65% for crowded WT, as shown in Fig. S8) We found no correlation between the morphology of the PVD and the response to harsh touch when testing the different combinations of genotypes and treatments. For example, isolated mec-10;degt-1 double mutants show crowded-like morphology with isolated-like behavioral response (Fig. 4A and Fig. S8, S9). These findings suggest that while degenerins are required for isolation-induced plasticity of both traits, response to harsh touch is independent of the structural alteration of the PVD.
(A) Discriminant analysis shows independence of harsh touch response from the PVD’s morphological classification. Squares indicate the centroids for morphological characteristics analyzed in Fig. S10A-C. The response to harsh touch (Fig. S9) is illustrated by its magnitude (low, <45% in yellow star; high, >65% in light blue circles). The different genotypes are numbered in the list on the right. Crowded-black, Isolated-red. (B-E) Isolated animals show similar PVD morphology regardless of their touch response: (B) isolated responding and non-responding animals are not different with regard to the fraction of ectopic branching, and (C) the geometry of quaternary branches (Isolated non-responding, n=18; Isolated responding, n=20). Box plot representation as in Fig. 2. Mann-Whitney test. n.s. not significant. (D, E) Representative PVD images of responding and non-responding worms. (Scale bar, 50 μm).
An additional line of evidence supporting the independence of the behavioral and morphological phenotypes was demonstrated by isolation of young adult worms for 24 h (Fig. 3). This isolation affects the structure of the PVD (Fig. S5) but has no effect on the response to harsh touch (Fig. 1G). Amiloride also has no effect on harsh touch response (Fig. S1C) but does alter dendritic morphology (Fig S7 A-D). Finally, to directly demonstrate that these two features are independent, we assayed harsh touch responses of isolated animals and then analyzed each individual animal for its PVD morphology. We then compared the dendritic morphology of responding and non-responding worms, and found that the morphological parameters were similar (Fig. 4B-E). Thus, analysis at the level of individual worms failed to demonstrate a correlation between the morphology and the response to harsh touch. In summary, the morphological and behavioral phenotypes were independently affected by sensory experience via degenerins. We cannot exclude the possibility that other functions of the PVD, like the response to low temperatures (Chatzigeorgiou et al., 2010) and proprioception (Albeg et al., 2011) are more tightly associated to the structure of the PVD, nor that the morphological changes induced by isolation are too minor to constitute a difference in neuronal function.
In addition to the isolation-induced changes in the number of ectopic branches and the percentage of straight quaternary branches, we found that worms raised in isolation are also more likely to lose the self-avoidance between two adjacent menorahs (candelabra). This effect is also mec-10-dependent, but appears to act cell non-autonomously. In addition, it is chemosensory and amiloride independent (Fig. S10).
MEC-10 and DEGT-1 localization is experience-dependent
Differential localization of degenerins can affect both the behavioral response to harsh touch and the structural properties of the neuron. When considering the effect of isolation, we hypothesized that changes in the localization patterns of DEG/ENaC can account for plasticity at both the behavioral (Fig. 1) and the structural level (Figs. 2, 3). Since MEC-10 and DEGT-1 tend to co-localize within the PVD (Chatzigeorgiou et al., 2010)·we analyzed the interaction between these two proteins, under different mechanosensory experiences. We found that MEC-10 localization in the plasma membrane, in intracellular vesicular compartments of the axon and in the quaternary branches was reduced after isolation (Fig. 5A-D, Movies S3 and S4). In contrast to MEC-10 (Figs. 5A-D), DEGT-1 localization is reduced only in the cell body following isolation (Fig. 5E-H). Furthermore, degt-1 mutants reduced the overall amount of MEC-10, and more importantly, abrogated the isolation-induced reduction in MEC-10 localization at the quaternary branches and the axon (Fig. 5A-D). In the reciprocal experiment, DEGT-1 localization was affected in mec-10 mutants, as isolated worms exhibit increased localization to the cell body compared with WT isolated worms (Fig. 5E-H). Thus, mechanosensory experience also induces plasticity in the localization pattern of MEC-10 and DEGT-1. We propose this differential localization may be part of the mechanism that independently and locally modulates dendritic and axonal properties, to affect both the structure and the function of the PVD, respectively.
(A) PVD::MEC-10::mCherry localization (shown and labeled in purple, also in Fig. 7) is reduced in the quaternary branches and the axon following isolation, in a degt-1 -dependent manner (WT: Crowded, n=31; Isolated, n=27; degt-1 mutants: Crowded, n=23; Isolated, n=22). (B-D) Representative images and reconstructions for PVD::MEC-10::mCherry localization for crowded and isolated WT worms, and crowded degt-1 worms. (Scale bar, 50 μm). (E) PVD::DEGT-1::mCherry localization level is reduced at the cell body, but not in the quaternary branches or the axon, following isolation, in a mec-10-dependent-manner (WT: Crowded, n=33; Isolated, n=21; mec-10 mutants: Crowded, n=28; Isolated, n=22). The percentage of expressing worms ± standard error of proportion is shown. (F-H) Representative images and reconstructions for PVD::DEGT-1::mCherry localization for crowded and isolated WT worms and crowded mec-10 worms. (I) Isolation leads to a reduced escape response following optogenetic photoactivation of Channelrhodopsin 2 in the PVD (Crowded, n=30; Isolated, n=24, grown on All Trans Retinal. No response was observed for worms grown without All Trans Retinal). Fisher exact test, *p<0.05, **p<0.01, ***p<0.01 n.s. not significant.
Optogenetic stimulation suggests behavioral plasticity is post-sensory
We have shown that sensory deprivation (isolation) affects the localization of two different mechanoreceptors, the degenerins MEC-10 and DEGT-1 in the dendrites, axon and soma of the PVD. It is conceivable that the differential localization of degenerins in different domains of the PVD may affect the morphology and function of the PVD. Thus, we decided to stimulate the PVD independently of the endogenous mechanoreceptors and study whether the escape behavior following harsh touch is similar in isolated and crowded animals. We hypothesized that if we circumvent the normal degenerin-mediated mechano-stimulation, exciting the PVD downstream to sensory activity, the animals grown in isolation will respond to the same degree as if they were in crowded conditions and independently of the dendritic tree morphology. Optogenetic stimulation is an elegant way to bypass the sensory perception mediated by receptors and transduced by diverse signal transduction pathways (Husson et al., 2012). Thus, to investigate whether the isolation-induced decreased response to harsh touch is a PVD-dependent function, we used optogenetic stimulation with Channelrhodopsin (ChR2) expressed in the PVD (Husson et al., 2012). We found that isolation significantly reduced the percentage of worms responding to optogenetic stimulation of the PVD (Fig. 5I), indicating that the plasticity in the response is acting downstream to PVD activation, mechanosensory channels and signal transduction pathways. This activation is sensitive to isolation probably because it is acting pre- or post-synaptically. To determine whether optogenetic stimulation of isolated animals can reverse the morphological changes induced by the absence of mechanical stimuli in PVD, we used optogenetic stimulation of the PVD and found no significant difference between isolated and isolated-optogenetically-stimulated animals in terms of any measured PVD structural characteristic (Fig. S11). Thus, optogenetic stimulation on isolated animals is not sufficient to convert their dendritic trees to crowded-like. In addition, isolated animals respond to optogenetic stimulation in a reduced way suggesting that the escape response is not dependent on the structure of dendritic trees but on unknown downstream pathways.
In summary, genetic, and pharmacological evidence suggest that dendritic structural plasticity is an autonomous activity-dependent homeostatic mechanism. Combined with optogenetic testing, our results indicate the mechanosensory dendritic tree morphology is independent from pre- or post-synaptic degenerin-mediated processes that affect behavioral escape in response to harsh touch.
Discussion
From the evolutionary point of view, dendritic trees and their structural complexity remain mysterious objects, despite many efforts to understand the contribution of arborization complexity to dendritic physiology and function (Hausser & Mel, 2003).
Previous research has demonstrated both cell autonomous (Aguirre-Chen, Bulow et al., 2011, Oren-Suissa et al., 2010, Salzberg et al., 2014) and cell non-autonomous (Dong et al., 2013, Salzberg et al., 2013) mechanisms that regulate the PVD’s dendritic morphogenesis during development. Some studies have also focused on regeneration and aging effects on the tree structure of the PVD, revealing plasticity in the adult stage (Iosilevskii & Podbilewicz, 2021, Kravtsov et al., 2017, Oren-Suissa et al., 2017), similar to what has been shown for Drosophila sensory neurons (DeVault, Li et al., 2018). We found, by using mutants for degenerins, chemosensory stimulation controls, pharmacology, optogenetics and glass beads, that the dendritic structure of the PVD and the behavioral response to harsh touch are activity- and mechanosensory-dependent, but appear to be chemosensory-independent. It is still conceivable that chemosensation can affect nociception and proprioception, in some sort of a cross modal plasticity mechanism, but at the moment we do not have evidence for this scenario in the PVD, where we show that mechanosensory channels (DEG/ENaCs) affect the architecture of the PVD.
Our results suggest that ‘nurture’, manifested as mechanosensory experience, activates mechanotransduction signaling, via DEG/ENaCs amiloride-sensitive activity, to stabilize the homeostatic structure of the dendritic tree in adults. The sensory experience induced when one body side of the worm is in contact with the agar plate is not sufficient for this purpose (Fig. S12); however, other worms on the plate, or the presence of glass beads, elicit significant structural alterations. Moreover, the cell autonomous activity of MEC-10 in the PVD was found to be both necessary and sufficient to preserve the crowded-experience phenotype in terms of the simplified structure of the PVD. While there are other multiple plausible explanations, our failure in remodeling the isolated-state arbor by repeated optogenetic activation of the PVD may provide further evidence for the necessity of degenerins in this structural plasticity, since neuron activation by ChR2 bypasses the activation of mechanically-gated degenerin channels. Figure 6 depicts our working model, where the amount of mechanosensory stimulation, in crowded or isolated conditions, affects the expression and localization of MEC-10 and DEGT-1 in different compartments of the PVD. Localization of MEC-10, probably by forming higher-order complexes with other DEG/ENaCs, can affect the structure of the PVD at the level of the dendritic tree. The morphological and behavioral phenotypes shown here were rescued by a PVD-specific expression of MEC-10, supporting such a cell autonomous mechanism. The experience-induced structural plasticity seems homeostatic at the individual branch dynamics level in the adult. In contrast, we show that the behavioral response to harsh touch is modulated by mechanosensory experience during development and by the presence of degenerins, and perhaps involves other neurons downstream. Based on optogenetic stimulation, we presume that the behavioral plasticity is a pre- and/or post-synaptic property, mediated by DEG/ENaC (Hill & Ben-Shahar, 2018) and related to neurotransmission modulation, independently of the structure of the PVD dendritic tree. These structural and behavioral plasticity are separated in time (adulthood vs. development) and space (dendrite vs. axon; Fig. 6). We found no correlation or causation between the structure and the nociceptive function of the PVD. It is possible that such a link exists, but our physiological and behavioral outputs are not sensitive enough to detect it.
During adulthood, mechanosensory signals maintain the structure of the PVD in crowded animals, with straight quaternary branches and fewer ectopic branches. Sensory deprivation (in isolated animals) results in ectopic dendrites, wavy branches and defects in candelabra self-avoidance. During development, mechanosensory experience alters the response to harsh touch and the crawling gait of the worm, possibly through changes in MEC-10 localization in the axon and mediated by other DEG/ENaCs. Mechanosensory stimuli are a driving force for changes in the compartment specific localization of MEC-10 and DEGT-1 in the PVD, which may affect the structure of the PVD. MEC-10 is represented in blue, DEGT-1 is represented in green.
Somatosensory activation in vertebrates plays a prominent role in shaping the structural and functional properties of dendritic spines, mainly studied in the central nervous system (Gyorffy, Kun et al., 2018, Holtmaat & Svoboda, 2009, Xu, Yu et al., 2009, Yang, Pan et al., 2009). In contrast to mammalian cortical neurons, much less is known about sensory neurons’ degree of plasticity and the molecular mechanisms utilized during adulthood. Here we suggest that degenerins mediate mechanosensation-induced dendritic growth in sensory dendrites. The dendritic plasticity we described bears resemblance to the activity-dependent effect of glutamatergic signaling and NMDA receptors. Activity via NMDA affects dendritic spines as an upstream mechanism of cell signaling, resulting in structural modifications (Nagerl, Eberhorn et al., 2004, Star, Kwiatkowski et al., 2002, Stein, Park et al., 2021, Zhang, Cudmore et al., 2015). It is possible that degenerins mediate mechanosensory signaling sensation, by activating cationic gradients (Kellenberger & Schild, 2002), leading to activation of downstream intracellular signaling pathways (Ghiretti et al., 2014, Sin et al., 2002, Vaillant et al., 2002, Zhou et al., 2006), which in turn stabilize local, actin-mediated (Halpain, 2018, Luo, 2002, Zuo et al., 2005) structural plasticity in the PVD dendritic branches. In Drosophila, at the epithelium, mechanosensitive ion channels, together with E-cadherin-catenin complexes and calcium sensing mechanisms affect epithelial morphogenesis (Roy Choudhury, Groflhans et al., 2021). In parallel, DEG/ENaCs may also modulate pre- and post-synaptic homeostatic signaling in the harsh touch circuit, as has been shown in neuromuscular junctions (Younger, Muller et al., 2013). While this study focuses on degenerins activation in the PVD with emphasis on dendritic structures, future studies may establish the pre- and -post-synaptic mechanisms which act downstream, on the transcriptional and translational levels. These directions for future studies have the potential to increase our understanding of the mechanisms that couple sensory experience, to structural dendritic plasticity.
In summary, we propose that the combinatorial actions of DEG/ENaCs have mechano-signaling functions mediating plasticity in sensory dendritic trees, and provide mechanistic insights into dendritic structural responses to sensory experience in adulthood and the behavioral consequences of such adaptations during development.
Materials and Methods
Strains
Nematode strains were maintained according to standard protocols (Brenner, 1974). The list of the strains is presented in Supplementary Table 1. Strains of the DEG/ENaCs family obtained from the CGC (JPS282: asic-1(ok415) I; ZB2551: mec-10(tm1552) X; VC2633: degt-1(ok3307) V) were crossed with BP709: hmnIs133[ser-2Prom3::kaede]. The validation of F2 homozygotes for the DEG/ENaCs deletions (including single, double and triple mutants) was performed by PCR amplification of the genomic area containing the deleted region.
Primers for Multiplex PCR
The list of the primers used is presented in Supplementary Table 2.
Spinning disk confocal microscopy
Prior to imaging, the worms were mounted on a 10% agar pad placed on a microscope glass slide, in 1μl polystyrene beads (100 nm diameter; Polysciences, Inc.) for their mechanical restraint, and sealed with a coverslip for complete physical immobilization (Kim, Sun et al., 2013). The PVD neuron was visualized using a Yokagawa CSU-X1 spinning disk, Nikon eclipse Ti inverted microscope and iXon3 camera (ANDOR). Images were captured with MetaMorph, version 7.8.1.0. For each worm, a sequential z-series image stack (step size of 0.35 μm) was obtained with an oil Plan Fluor 40X (NA 1.3) lens around the PVD cell body, encompassing approximately 50-100 μm segments both anteriorly and posteriorly.
Data analysis
The analysis of the PVD structure was performed for the area surrounding the cell body. All images were analysed with ImageJ, version 1.48 (NIH), in TIFF format, by producing a maximal intensity z-series projection and converting it to negative form (invert lookup table) for improved visibility. Ectopic branching was defined as described previously (Hausser & Mel, 2003). Briefly, non-ectopic branches form the ‘ideal’ WT candelabrum of the late L4 stage, whereas excess branches, which create non-quaternary terminal ends, are considered ectopic, as illustrated with dashed lines for ectopic branching in Fig. 2A. The total number of ectopic and non-ectopic branches was quantified for each image, and presented as a fraction (ectopic/non-ectopic branches).
The geometry of each quaternary branch (“candle”) was defined in the following manner: Straight geometry-all the pixels that constitute the branch are positioned on a straight line generated with ImageJ. The width of the line (1 pixel) was constant for the entire sets of experiments. The number of straight quaternary branches were divided by the total quaternary branches in the image for each worm and presented as a percentage. The analysis was done only for worms which did not move through the Z series. Moving worms were excluded from the experiment. Self-avoidance defects-the number of events where two adjacent candelabra overlapped (no gap formation), was divided by the total number of gaps between the candelabra within the frame (Fig. 2A). The self-avoidance values are presented as percentage.
Behavioral procedures
Harsh touch assay
After 72 h, adult worms (both isolated and those from the crowded plate as described below) were transferred using an eyelash each to a new agar plate, freshly seeded with 150 μl OP50 (about 16 h after seeding). This step is required to avoid a thick edge of the bacterial lawn, which might interfere with harsh touch response measurement. After ~45 minutes in the plate, the non-moving worms were prodded with a platinum wire posterior to the vulva, above the interface between worm’s body and the agar plate (Way & Chalfie, 1988), every 10 seconds, and the number of responses to harsh touch was counted. Animals with a functional PVD moved, sometimes backing up. More than one response constitutes a responsive animal. The non-responding worms were defined if two prodding events were observed sequentially without response. The percentage of responding worms was calculated for each genotype and treatment. The experimenter was blind to both the genotype and the treatment-crowded or isolated.
Isolation of embryos
To establish the method for conspecifics based mechanosensory stimulation, we performed several calibration experiments with different population densities (250 embryos, 15 adults, 30 adults) to test conditions for the crowded plate.
We found a gradual increase in the crowding effect and reduced variation in the measured parameter. Following that, we decided to work with the group of 30 crowded adults” as a source of mechanosensory signal.
The worm isolation procedure was based on previous work (Rose et al., 2005) with a few modifications, as indicated. Isolated animals were grown on 6 cm agar plate with 150 μl of OP50 E. coli, while crowded plate worms were grown on 600-700 μl OP50 to prevent starvation. The embryos and adult worms were isolated with platinum wire. The plates were sealed with one layer of Parafilm M and placed into a plastic box, at 20 °C, for the entire experiment.
Three experimental groups were used for the 72 h (96 h of experiment was performed only for mec-4 worms, since they were L4-very young adults at 72 h) isolation experiment: (1) Single isolated embryos (2) Crowded worms - the progeny of 30 young, non-starved, adults (approximately, 7,000-9,000 worms in different developmental stages, without approaching starvation) (3) Crowded adult worms that were isolated for a certain amount of time as adults.
After 48/72/96 h (according to the experiment), age-matched worms from each group were transferred to 10% agar pad slides for imaging, as described above (Rose et al., 2005).
Isolation of adults for 2, 5, 15 or 24 hours
Crowded plates were prepared as described above, with progeny from 30 adult hermaphrodites. Animals were separated into individual plates for the desired time window before 72 h have passed (i.e. after ~70 h for 2 h isolation, ~67 h for 5 h isolation etc.). Worms were isolated using an eyelash into new plates containing 150 μl of OP50. At 72 h, the PVD of isolated animals was imaged and compared with age-matched animals which remained in the crowded plate of origin (see Fig. 1B).
Optogenetics stimulation for harsh touch
Crowded and isolated worms (ZX819: lite-1(ce314) X; xIs12[pF49H12.4::ChR2::mCherry; pF49H12.4::GFP]) were grown on OP50 with 100 μM All Trans Retinal (ATR, Sigma R2500), in order to obtain functional Channelrhodopsin (Husson et al., 2012) (as a control we tested the response of worms that were grown on ethanol alone, no response was detected for these worms). After 72 h the worms were singly mounted with eyelash on a chunk (1cm X 1cm) of agar that was mounted on a microscope glass slide. The agar contained fresh but dry OP50, with 100 μM ATR. About 30 minutes following the transfer the worms were tested for the response to light. Worms were stimulated at 488 nm wavelength, with laser intensity of 40% and exposure time of 100 ms, with 10X Plan Fluor (NA 0.3) for ~1 second and the forward acceleration response was measured. The microscope system is as described above. The experimenter determined the presence or the absence of forward acceleration in response to light activation.
Optogenetics stimulation of isolated worms
Isolated eggs were grown as described at the ‘Isolation of embryos’ section. Isolated worms (ZX819: lite-1(ce314) X; xIs12[pF49H12.4::ChR2::mCherry; pF49H12.4::GFP]) were grown on 150 μM OP50 with 100 μM ATR (Husson et al., 2012) or 0.3% ethanol as control. After 72 h both the crowded and isolated worms were singly mounted with eyelash onto a cube (with surface area of about approx.1cm2, depth of 0.5 cm) of agar layered with fresh, fully dried, OP50 with 100 μM ATR or 0.3% ethanol as a control group. The worms were stimulated with 488 nm wavelength, with laser intensity of 40% and exposure time of 100 ms, with 10X Plan Fluor (NA 0.3) for ~60 seconds, every 5 min during 4 h. Stimulation and recording was performed with the microscope system described above.
As a control, before each experiment, crowded ZX819 worms were tested for forward acceleration as a positive control to the functionality of the ChR2. After the end of 4 h stimulation, the worms were mounted for imaging of the PVD at 40X as described under “Spinning disk confocal microscopy” section.
Isolation with chemical stimulation
The glp-4(bn2) mutants which are sterile at 25 °C were used (~40 worms for each plate), in order to prepare conditioned/chemically stimulated plates (Maures et al., 2014) prior to growing isolated animals. The glp-4 mutants were transferred at early larval stage (L1, L2) to a new agar plate for 96 h at 25 °C. After the removal of the glp-4 mutants, the embryo isolation procedure was used, as described before at 20 °C.
Isolation with glass beads
Single embryos were isolated to agar plates with 150 μl OP50 and 2.5 g of glass beads (1 mm diameter) were placed on the OP50 lawn in the middle of the plate. The worms were isolated for 72 h and tested for response to harsh touch as described above.
Pharmacology
Amiloride hydrochloride hydrate (Sigma, #A7410) 1 M stock solution in DMSO was stored at −20°C. A final concentration of 3 mM amiloride in 0.03 % DMSO was prepared in OP50 bacteria (LB medium) and seeded on NGM plates. 650 μl of the OP50 mixture was seeded on each plate. As a control, 0.03 % DMSO was added to OP50 bacteria. For each plate (control 0.03% DMSO or 3 mM amiloride) 30 non-starved adult worms were added. After 72 h at 20 °C the progeny of the 30 adults were tested as young adults for their PVD morphology and their response to posterior harsh touch, as described in the previous sections.
Anesthetics and long-term imaging
In addition to immobilization of worms with 100 nm polystyrene beads as described in the ‘‘Spinning disk confocal microscopy” section above, two additional methods for long term imaging and pharmacological effects assays were used: 1% tricaine (Sigma, A5040) in M9 buffer or a mixture of 0.01% tetramisole (Sigma, T-1512) and 0.1% tricaine in M9 buffer were utilized.
Worms were exposed to the anesthetics (in a glass well) for ~20 min until paralyzed, then transferred to 3% agar pads with 1 μl of the anesthetics. Sequential z-axis image series (0.6 μm step-size) were taken, as described above, every 5 min for 2-3 h.
Imaging individual worms-crowded to isolated
Worms were grown at crowded conditions, as described above. At ~24 h of adulthood, individual worms were placed in a 1% tricaine solution in M9 buffer for 10 min until paralyzed and transferred to 3% agar slides with 1 μl 1% tricaine. Following a short PVD imaging session, the worms were recovered from the slide with M9 and isolated to a new plate for 4 h. Each 4 h-isolated animal was anesthetized once again with 1% tricaine for 10 min prior to PVD imaging.
Analysis of DEG/ENaCs localization in the PVD
Two DEG/ENaCs protein constructs, pF49H12.4::MEC-10::mCherry and pF49H12.4::DEGT-1::mCherry (plasmids kindly provided by W. Schafer’s lab, (Chatzigeorgiou et al., 2010)) were analysed for their localization in the PVD, by comparing worms expressing the co-injection marker punc-122::GFP raised in crowded or isolated conditions, in a similar behavioral assay as described in Fig. 1B. A z-series (step size of 0.35 μm, immobilization performed with polystyrene beads, as described in the microscopy section) of the area around the cell body of the PVD was obtained with a 60X Apochromat (NA 1.4) lens. The images (after maximal intensity projection of the Z-series) were encoded so the analysis was performed in a blind manner. The presence or absence of fluorescent signal was examined in three compartments: The cell body, the quaternary branches and the axon of the PVD.
Rescue strains
Worms from BP1022 (mec-10(tm1552) X; hmnIs133[ser-2Prom3::Kaede]; him-5(e1490) V) were injected into the gonad with a rescuing plasmid for mec-10, with a PVD specific promoter (pWRS825: ser-2Prom3::mec-10 genomic) kindly provided by W. Schafer’s lab·(Chatzigeorgiou et al., 2010). The injection mix contained myo-2p::GFP (20 ng/μl) as a co-transformation marker and pWRS825 (80 ng/μl). For both behavioral and structural characterization, the him-5;mec-10 strains, with and without rescuing plasmid, shared the same crowded plate and were differentiated by the presence or absence of the co-injection marker.
Effect of plate on PVD in contact with agar and the opposing side
Embryos were singled into plates with 150 μl OP50 bacteria at 20°C, and analyzed as young adults 72 h later. Every animal was transferred using an eyelash into freshly seeded plates (<24 h), and noted for its side orientation. Once on these experiment plates, animals were observed for their side orientation at 30-minute intervals for a minimum of 3 h prior to microscopy. For microscopy of both PVD neurites, while few animals were removed from the slide and repositioned using an eyelash, most animals were imaged using a coverslip ‘sandwich’ method (adapted from (Sulston, Schierenberg et al., 1983)), whereby the animal is pressed on an agar pad between two large coverslips, which are subsequently flipped as a single unit and observed from the other side. Briefly, a thin 3% agar pad was pressed onto a 24×50 mm coverslip (170±5 μm thick), placed on a standard microscope slide for ease of handling. Animals were singly placed in a drop of 0.05% tetramisole (Sigma, T-1512) in M9 on the agar pad using an eyelash, and a second, similar coverslip was immediately placed above. The two coverslips were sealed shut onto the supporting slide using two thin strips of marking tape, and once the animal was anaesthetized the PVD on the side closest to the objective was imaged as described above (see Spinning Disk confocal microscopy). For viewing the opposite neurite, the tape was carefully peeled and the two coverslips flipped as one, taped again, and similarly imaged. Note that this method inevitably makes the second side imaged to be fainter due to the laser penetrating through the agar pad layer itself; laser intensities were typically increased by up to 5% in order to compensate for the decrease in fluorescence.
Statistics and data plotting
At least two independent experiments constitute the data set described in each figure. For the morphological characterization of the PVD the results are expressed as means (blue circle) ± standard error of the mean (s.e.m.). In the boxplot (first, third quartiles) the upper whisker extends from the hinge to the highest value that is within 1.5 * IQR (inter-quartile range), the distance between the first and third quartiles. The lower whisker extends from the hinge to the lowest value within 1.5 * IQR of the hinge.
The statistical analyses were performed with SPSS software (IBM, version 20) and “R package”. Two-tailed tests were performed for the entire data sets.
Since for many experiments the distribution of the data was not normal, nonparametric tests were used: Mann-Whitney test for comparison between two independent groups. Kruskal-Wallis test was used for multiple comparisons for more than two groups. For proportions (percentage worms) ± standard error of proportion was calculated. Fisher’s exact test was used for analysis of differences in proportions. To estimate the variability in proportion we calculated the Standard Error of Proportion: 
The dot plot figures were prepared with “R package”, the bar charts with Microsoft Excel software. Final figures were prepared with Adobe Illustrator CS version 11.0.0.
Discriminant analysis
Eight different strains (WT and seven DEG/ENaCs), with two treatments (crowded, isolated worms) for each strain were analyzed for Linear discriminant analysis for morphological characteristics, to evaluate similarity between different strains and treatments. Each worm in the data set was characterized by the three morphological characteristics (the fraction of ectopic branching, the percentage straight quaternary branches and the percentage of self-avoidance defects). The centroid for morphological characterization was calculated for each condition and represented by a square. Data from independent harsh touch experiments are shown for each group. The analysis was performed using SPSS 20.
Acknowledgments
We thank current and former lab members for their intellectual and technical support. Veronika Kravtsov, Sagi Levy, Anna Meledin, Meital Oren-Suissa, Tom Shemesh, Shay Stern, Yehuda Salzberg and Alon Zaslaver for critically reading and commenting on the manuscript. Ehud Ahissar, Dan Cassel, Michel Labouesse and Kang Shen for fruitful discussions. William Schafer, Max Heiman, Hannes Bulow, Yehuda Salzberg and Alexander Gottschalk for plasmids and strains. Some strains were provided by the Caenorhabditis Genetics Center (CGC), which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). This work was supported by grants from the Israel Science Foundation (442/12 and 257/17), Adelis Fund (2023479), and the Ministry of Science and Technology (3-13022).
Footnotes
We describe how we calibrated the experiments to different population densities that give different levels of mechanosensory stimulation. We also removed the gait and proprioception experiments summarized in the original Supplementary Fig. S2 and Movies S1-S4. We added a new Supplementary Fig. S12 showing the data supporting that the contact with the plate does not significantly affect PVD structure between the contacting and opposing bilateral PVDs. We also highlight that the focus of this study is the homeostatic changes in dendritic structure and not the potential changes in homeostatic synaptic plasticity (HSP). The HSP may affect the synaptic strength in response to sensory activities and how isolation may affect downstream circuits and the synapse physiology at the level of downstream transcriptional and translational responses. We now summarize the results that support the phenomenon of Homeostatic Dendritic Structural Plasticity (HDSP) that affect the mechanosensory dendritic tree morphology independently from pre- or post-synaptic degenerin-mediated processes that affect behavioral escape in response to strong mechanical stimuli.
References
