Introduction

Adaptive behavioral responses require animals to interpret sensory signals in the context of prior experiences, enabling effective responses to changing environmental conditions. For example, rodents associate contextual cues, such as odors and colors, with rewards or punishments, leading to context-dependent adjustments in approach or avoidance behaviors [1,2]. Disruptions in the biological processes underlying this plasticity, as seen in neurological and psychiatric disorders, often result in impaired behavior modulation based on prior sensory experiences [3–7]. Recognizing the critical role of context-dependent behavioral modulation for survival and well-being, ongoing research focuses on uncovering the neuroplastic mechanisms responsible for these adaptive processes.

Studies across diverse organisms reveal that animals, regardless of nervous system complexity, integrate prior sensory experiences to shape behavior [8–15]. In the nematode C. elegans, which has only 302 neurons, locomotion strategies are modified by prior exposure to stimuli such as odors, touch, light, and starvation, demonstrating robust adaptive behaviors [16–23]. The wide range of adaptive behaviors, combined with the well-mapped C. elegans nervous system and abundant genetic tools, have empowered the discovery of fundamental circuit mechanisms and molecular pathways underlying context-dependent behavioral modulation [24–28].

The C. elegans sensory system comprises 60 ciliated sensory neurons specialized in detecting chemical, thermal, mechanical, and olfactory stimuli [29,30]. Each sensory neuron has a dendritic apparatus that converts specific environmental signals into neuronal activity, which drives signal flow through downstream networks of interneurons and motor neurons, ultimately shaping behavior [31–33]. Among these, AFD, a pair of bilaterally symmetric left-right neurons, serves as the primary thermosensory neurons required for thermotaxis (movement toward the a preferred temperature) [34]. The dendritic endings of AFD, enriched with microvilli, serve as the primary site for thermal detection. Even when detached from the cell body, these dendritic endings retain calcium responsiveness to temperature stimuli [35]. Moreover, genetically disrupting these sensory endings abolishes the Ca²⁺ response in AFD cell bodies to temperature changes and eliminates thermotaxis, underscoring their essential role in thermal sensing [36–38].

The second messenger cyclic guanosine monophosphate (cGMP) plays a central role in modulating AFD thermosensory activity. Temperature changes in AFD neurons trigger intracellular Ca²⁺ dynamics, with Ca²⁺ levels rising during warming and falling during cooling [35,39,40]. This response relies on cGMP synthesis mediated by the guanylyl cyclases (GCs) gcy-8, gcy-18, and gcy-23, which have overlapping but distinct roles. Specifically, gcy-8 is essential for generating thermally evoked Ca²⁺ responses in AFD [41,42], whereas gcy-18 and gcy-23 act synergistically to sustain temperature sensing [41,43]. The cGMP produced by these cyclases activates cyclic nucleotide-gated (CNG) channels, comprised of alpha-subunits TAX-4 and beta-subunits TAX-2, resulting in increases in intracellular Ca²⁺. Mutations in either tax-4 or tax-2 abolish thermally evoked Ca²⁺ responses in AFD and disrupt thermotaxis, underscoring the importance of cGMP signaling for AFD-mediated thermosensation [40].

AFD transmits thermal signals to several downstream interneurons, including AIY, AIZ, and AIB, which process this information to guide temperature-dependent behaviors. Thermotaxis in C. elegans encompasses three categories: movement toward warmer temperatures (thermophilic), movement toward cooler temperatures (cryophilic), and movement within a preferred temperature range (isothermal) [44,45]. Each behavior type depends on distinct AFD synaptic partners: AIY is critical for thermophilic and isothermal behaviors [46,47]; AIZ is specific for cryophilic behavior [48–50]; and AIB also supports thermophilic behavior, though its role is less critical than that of AIY [51]. These findings demonstrate the functional specificity of synaptic partners of AFD neurons in shaping distinct temperature-dependent behaviors.

In this study, we demonstrate that AFD plays a novel role in tactile-mediated, context-dependent locomotion modulation, thereby expanding its function beyond temperature sensing. The mechanisms underlying this role differ markedly from those involved in thermosensation. First, context-dependent modulation requires distinct cGMP signaling components. Second, while the microvilli-rich sensory endings of AFD are essential for thermosensation, they are dispensable for context-dependent modulation. Third, this modulation depends on mechanosensory neurons and on electrical synapses between AFD and the AIB interneuron, which connects to mechanosensory circuits. Finally, disruption of inx-7 and inx-10 genes, encoding gap junction protein innexins in AFD and AIB, impairs context-dependent locomotion modulation. Remarkably, restoring AFD-AIB connectivity via engineered Cx36 electrical synapses rescues this function. Collectively, our findings uncover an unconventional mechanism by which AFD mediates behavior modulation based on tactile experience.

Results

C. elegans display distinct locomotion in identical environments after exploring different physical structures

To test whether prior physical experience influences C. elegans behavior, we allowed worms to explore microfluidic chambers with distinct pillar designs before recording their locomotion in identical assay zones (Fig. 1A). Because the environment within the assay zone is the same across chambers, behavioral differences observed there are likely to reflect an influence of prior tactile experience. To implement this assay, we designed two microfluidic chambers that differ only in the physical structure of the exploration zone: a uniform chamber and a binary chamber (Fig. 1A). Before reaching the assay zone, worms explore surrounding areas with pillar designs that differ in size and spacing (exploration zone, Fig. 1A). This design allowed us to assess how exposure to different physical contexts impacts locomotion in an otherwise identical setting.

Figure 1.

Worms were allowed to explore the chambers for 60 minutes before behavior was recorded in the assay zones. We found that worms in the uniform chamber assay zone exhibited locomotor behavior distinct from those in the binary chamber assay zone (Fig.1 B-F). Specifically, worms in the uniform chamber assay zone moved approximately 40% faster (Fig. 1B, F), spent ∼20% less time idle (Fig. 1C, F), reversed direction ∼20% less frequently (Fig. 1D, F), and turned ∼30% more often (Fig. 1E, F) than worms in the binary chamber assay zone. Among the parameters measured, locomotion speed showed the most robust and reproducible modulation, establishing a stable behavioral metric for dissecting the underlying circuit and molecular mechanisms. These findings demonstrate that locomotion is shaped by past sensory experience rather than solely by immediate environmental cues.

Because worms in the binary chamber are exposed to both pillar types and remain free to move between exploration and assay zones, the behavioral differences described above could reflect exposure to a more complex physical environment rather than prior experience alone. To directly test whether locomotion is modulated by prior physical experience independently of continued access to the exploration zone, we designed microfluidic chambers in which the assay zone could be separated from the exploration zone by a removable barrier (Fig. 1–Supplement 1A). In these chambers, worms were initially allowed to explore the entire device, including exploration zones that either matched or differed from the assay zone. A barrier was then inserted to prevent worms in the assay zone from re-entering the exploration zones.

Under these conditions, locomotion immediately after barrier insertion was higher in worms that had previously explored physical settings matching the assay zone (205 ± 8 µm/s) than in worms that had explored non-matching settings (151 ± 7 µm/s; p = 0.006; Fig. 1–Supplement 1B). This difference persisted when worms were recorded 40 minutes after barrier insertion, with animals in matching chamber retaining their higher locomotion rates (218 ± 11 µm/s) compared to those in non-matching chambers (185 ± 8 µm/s; p = 0.02; Fig. 1–Supplement 1B). These findings demonstrate that prior exploration of distinct physical environments can modulate locomotion even when worms are prevented from returning to those environments, supporting a role for prior physical experience independent of ongoing sensory input.

Guanylyl cyclase gene gcy-18 is required for context-dependent locomotion modulation in chambers

To identify mechanisms underlying context-dependent locomotion adjustments, we employed a previously reported preference assay that uses microfluidic arenas to assess the ability of worms to select a preferred space [52]. Because worms must adjust their locomotion based on their surroundings to remain within a preferred area, we hypothesized that mutations abolishing this preference would reveal key mechanisms underlying context-dependent adjustments in the uniform and binary chambers. Accordingly, we screened for mutant worms that failed to choose their preferred area, aiming to identify the genetic mechanisms mediating locomotion adjustments in response to different physical structures.

Given the significant role of cGMP in neuronal plasticity and behavior modulation, we performed a small-scale genetic screen using loss-of-function mutants for 24 genes encoding guanylyl cyclases (GCs), the enzymes responsible for cGMP synthesis. Using the preference assay, we identified two putative null mutants, gcy-18(gk423024) and gcy-12(gk142661), that displayed significant defects in spatial preference (Fig. 2–Supplement 1A). We then compared the locomotion of gcy-18 and gcy-12 mutants in both uniform and binary chambers. gcy-12 mutant worms exhibited normal speed adjustment compared to wild type (Δspeed: gcy-12: 24 ± 5% vs. N2: 35 ± 7%, p = 0.27, Fig. 2–Supplement 1B), indicating that gcy-12 is not required for locomotion modulation. In contrast, gcy-18(gk423024) mutants failed to execute context-dependent speed adjustment (Δspeed gcy-18: 3 ± 5% vs. N2: 35 ± 7%; p = 0.0013; Fig. 2A). A second allele, gcy-18(nj38), similarly showed significantly reduced Δspeed (11 ± 2%; p =0.011 compared to wild type; Fig. 2A). Together, these results indicate that gcy-18 is required for context-dependent behavioral adjustment.

Figure 2.

To confirm that the defects observed in gcy-18 mutants were not due to background artifacts, we introduced a single-copy gcy-18 transgene driven by the gcy-18 promoter into gcy-18(gk423024) mutants. Expression of this transgene restored locomotion adjustments to near wild type levels (Δspeed: gcy-18p::gcy-18: 28 ± 6% vs. N2: 35 ± 7%, p = 0.8, Fig. 2B), confirming that loss of gcy-18 underlies the observed behavioral defects. The identification of gcy-18 was particularly intriguing, as its expression is restricted to a single pair of neurons, the thermosensory AFDs. Given that our assay chambers lacked thermal gradients, these findings suggest a previously unrecognized role for gcy-18 and AFD neurons in mediating locomotion adjustments based on prior exploration of physical environments.

To further investigate the role of AFD-specific guanylyl cyclases in context-dependent locomotion modulation, we examined gcy-8, which is essential for generating thermally evoked Ca²⁺ responses in AFD [41,42]. In thermal responses, disrupting gcy-8 alone is sufficient to impair AFD function, whereas disruption of gcy-18 alone does not [41,43], indicating the critical role of GCY-8 in AFD thermosensation. If gcy-18 influenced behavioral adjustment in microfluidic chambers solely through its role in thermosensation, we would expect that a loss-of-function mutation in gcy-8 would have a similar or even more disruptive impact on locomotion modulation. However, gcy-8 mutant worms exhibit little change in their ability to adjust locomotion speed, with Δspeed values similar to that observed in wild type worms (Δspeed: gcy-8: 29 ± 7% vs. N2: 35 ± 7%; p = 0.96; Fig. 2C). These results indicate that while gcy-8 is critical for thermosensation, its mutation does not affect context-dependent locomotion modulation, suggesting that independent cGMP signaling mechanisms govern AFD neuronal activity in thermosensation and context-dependent behavior in microfluidic chambers (Fig. 2D).

Distinct CNG channels mediate context-dependent locomotion modulation and thermosensation

The role of guanylyl cyclase in cGMP synthesis prompted us to investigate whether cyclic nucleotide-gated (CNG) channels are involved in context-dependent locomotion modulation. CNG channels, composed of α and β subunits, open upon binding cyclic nucleotides (e.g., cGMP and cAMP), allowing Ca²⁺ influx [53–56] to control signal transduction and neural plasticity across species [57–64] (Fig. 3A). In C. elegans, mutations in either the tax-4 or tax-2 genes, which encode the α and β subunits of CNG channels respectively, impair thermotaxis behavior [47,65] and eliminates AFD responses to temperature changes [40,66]. To determine their role in context-dependent locomotion modulation, we examined tax-4(p678) and tax-2(p694) mutant worms using our microfluidic chamber assay.

Figure 3.

We found that tax-2 is essential for context-dependent locomotion modulation, as tax-2 mutants did not exhibit differences in speed between chamber types (Fig. 3B, C). In contrast, tax-4 mutants displayed an enhanced locomotion adjustment compared to wild type worms (Δspeed: tax-4: 68 ± 6% vs. N2: 41± 6%; p = 0.002; Fig. 3C), a striking difference from the lack of speed changes seen in tax-2 mutants (Δspeed: tax-2: 2 ± 2%; p < 0.0001 compared to wild type). We examined the locomotion speed of mutant worms in the binary chambers, which we refer to as the basal speed because wild type worms consistently move slowest in this environment. We found tax-4 mutants showed significantly higher basal speed than the wild type (tax-4: 193 ± 9 µm/s vs. N2: 104 ± 4 µm/s, p < 0.0001; Fig. 3B), indicating that tax-4 mutants can still adjust locomotion speed based on prior experience even when basal locomotion is accelerated.

To further examine how basal locomotion influences context-dependent modulation, we took advantage of the fact that worm locomotion rates are sensitive to environmental temperature [67,68]. We examined the relationship between temperature-induced changes in basal locomotion and locomotion modulation based on prior experience of physical surroundings. Wild type worms cultivated at either 23°C or 17°C and then transferred to 20°C exhibited significantly changes in basal locomotion speed. Despite changes in basal speed, the extent of locomotion modulation in the chambers was largely unaffected (Fig.3–Supplement 1A). Specifically, locomotion speed was positively correlated with temperature shifts in both chambers (binary chambers: r = 0.52, p < 0.0001; uniform chambers: r = 0.48, p < 0.0001; Fig.3–Supplement 1B), whereas no significant correlation was observed between temperature and Δspeed (r = −0.03, p = 0.22; Fig.3—Supplement 1C), supporting the notion that basal locomotion speed and context-dependent modulation are regulated independently.

Together, these observations indicate that TAX-2 is required for integrating prior experience into locomotion modulation, whereas TAX-4 primarily regulates overall locomotion level. Although both pathways engaging cGMP signaling, their distinct phenotypes suggest separate regulatory mechanisms for basal locomotion and context-dependent adjustments.

Given that four other genes (cng-1, cng-2, cng-3 and cng-4) are predicted to encode CNG channels in C. elegans, we investigated their role in locomotion modulation in microfluidic chambers [69]. Due to abnormal development in cng-1 mutants, we were unable to assess their locomotion; however, we were able to evaluate mutants of the other three genes. We found that cng-3(gk541751) mutant worms showed significantly decreased speed modulation compared to wild type (Δspeed: cng-3: 5 ± 4%, p = 0.0002; Fig. 3D), while cng-2(tm4267) and cng-4(gk195496) mutants exhibited normal modulation comparable to wild type worms (p = 0.54 and p = 4.47 respectively; Fig. 3D). The identification of CNG-3 was particularly intriguing, as recent studies suggest that it mediates changes in AFD neuronal activity based on prior temperature experiences [70,71], indicating that CNG-3 contributes to neural plasticity rather than direct stimulus detection. Our results further support this view by showing that CNG-3 plays a broader role in locomotion modulation in environments lacking thermal gradients, which suggests its involvement in neuronal plasticity beyond the thermosensory modality.

Next, we determined whether AFD is the site where CNG-3 functions in context-dependent locomotion modulation. To do so, we introduced single-copy transgenes expressing cng-3 cDNA under the control of two AFD-specific promoters, srtx-1Bp and gcy-18p, into cng-3(gk541751) mutant worms. Regardless of the promoter used, AFD-specific expression of cng-3 restored the locomotion speed modulation to levels comparable to wild type (Δspeed: srx-1Bp::cng-3: 47 ± 8%; p < 0.0001 compared to cng-3 mutant, p = 0.91 compared to wild type; and gcy-18p::cng-3: 40 ± 6%, p = 0.0007 compared to cng-3 mutant, p = 0.99 compared to wild type; Fig. 3E). These results indicate that AFD neurons are the site of action for CNG-3 in mediating context-dependent locomotion modulation. Together, our findings reveal a role for AFD beyond thermosensation and show that components of a cGMP signaling pathway enable AFD to modulate locomotion in an experience-dependent manner.

AFD sensory microvilli are dispensable for locomotion modulation, but AFD neurons are required

The finding that GCY-18 and CNG-3 in AFD neurons are required for context-dependent locomotion prompted us to test whether the AFD thermosensory apparatus is also necessary for this function. This apparatus consists of sensory microvilli at the tip of the AFD dendrite (Fig. 4A) and is essential for temperature sensing and thermotaxis [36,38]. For example, mutations in the kcc-3 gene, which encodes a K⁺/Cl⁻ cotransporter in the amphid sheath glia, disrupt microvilli formation and significantly impair thermotaxis toward the cultivation temperature [37].

Figure 4.

Despite this disruption, locomotion modulation based on experience of physical settings remained intact in kcc-3(ok228) mutant worms compared to wild type (Δspeed: kcc-3: 33 ± 4% vs. N2: 42 ± 7%, p = 0.28; Fig. 4B, C). kcc-3 mutants also displayed elevated basal locomotion rates (basal speed: kcc-3: 174 ± 8 µm/s vs. N2: 115 ± 7 µm/s, p < 0.0001; Fig. 4B), similar to what we observed in tax-4 mutants. This supports our earlier finding that overall locomotion rate does not necessarily correlate with context-dependent modulation.

To further test whether AFD sensory endings are required for context-dependent locomotion modulation, we examined ttx-1(767) mutant worms, which completely lack AFD microvilli and exhibit more severe thermotaxis defects than kcc-3 mutants [38]. Despite the loss of sensory microvilli, ttx-1 mutants retained the ability to modulate locomotion in microfluidic chambers. The extent of modulation was modestly reduced and more variable compared to wild type animals (Δspeed: ttx-1: 29 ± 10% vs. N2: 42 ± 7%, p = 0.29, Fig. 4D). Because ttx-1 mutations disrupt the expression of multiple AFD genes, including components of the cGMP signaling pathway [72], this increased variability may reflect broader perturbations in AFD function. Nonetheless, these results indicate that AFD sensory microvilli are not strictly required for context-dependent locomotion modulation.

Given that gcy-18 is required for context-dependent locomotion modulation and that GCY-18 localizes to the distal dendrite of AFD, we next examined how disruption of sensory microvilli affects its localization in AFD. We used a split-GFP strategy to visualize endogenous GCY-18 [73]. A tandem array of seven GFP11 β-strands (GFP11x7) was inserted at the C-terminus of GCY-18 using CRISPR-Cas9. When complemented with GFP1-10, GCY-18::GFP11x7 fluorescence was strongly enriched at the AFD sensory microvilli near the nose (Fig. 4–Supplement 1A-A′′), consistent with previous reports [42,74,75]. In addition, weaker but reproducible GCY-18 signal was detected near the AFD soma and axon (Fig. 4–Supplement 1A′′′). Importantly, in kcc-3, which exhibit disrupted sensory microvilli, and ttx-1 mutants, which lack sensory microvilli, GCY-18 remained localized to the distal dendrite and was still detectable near the soma and axon (Fig. 4–Supplement 1B-B′′’ and 1C-C′′′). Although these experiments do not identify the precise subcellular site at which GCY-18 acts, they show that disruption or loss of sensory microvilli does not substantially alter GCY-18 localization within AFD.

We next asked whether AFD neurons themselves are required for this behavior. Using a published strain expressing caspase under the AFD-specific gcy-8 promoter [76], we found that AFD-ablated worms completely lost the ability to adjust locomotion in microfluidic chambers. These animals showed no detectable differences between uniform and binary chambers in locomotion speed, idle time, or turning frequency (Fig. 4E). Together, these results indicate that while AFD sensory microvilli are dispensable for context-dependent locomotion modulation, the AFD neurons themselves are essential.

Building on our finding that locomotion modulation can be driven by prior physical experience even after worms are prevented from re-entering the exploration zones, we next tested whether AFD is required for this modulation using chambers in which the exploration and assay zones were separated by a removable barrier (Fig. 1–Supplement 1A). Under these conditions, locomotion modulation was significantly reduced in AFD-ablated worms (Δspeed: -AFD = 1 ± 6% vs. N2 = 23 ± 7%; p = 0.036; Fig. 4–Supplement 2A). Similarly, gcy-18 mutants showed defective locomotion modulation (Δspeed: gcy-18 = -1 ± 8% vs. N2 = 23 ± 7%; p = 0.034; Fig. 4–Supplement 2A). These results indicate that AFD and gcy-18 are required to generate locomotion modulation in response to recent physical experience.

Finally, to determine whether the modulation defects observed in gcy-18 mutants and AFD-ablated worms could be attributed to developmental abnormalities or gross motor impairments, we measured locomotion speed and body length on standard NGM plates. Both day-1 adult AFD-ablated worms (speed: 281 ± 10 µm/s; p = 0.33; body length: 1.12 ± 0.01 mm; p = 0.76) and gcy-18 mutants (speed: 291 ± 13 µm/s; p = 0.22; body length: 1.15 ± 0.02 mm; p = 0.86) showed locomotion speeds and body lengths comparable to wild type controls (speed: 252 ± 30 µm/s; body length: 1.14 ± 0.02 mm; Fig. 4–Supplement 2B, C). These results indicate that the loss of context-dependent locomotion modulation is not due to developmental defects or gross impairments in locomotion.

Context-dependent locomotion modulation requires the MEC-10 mechanosensory channel subunit

Having established that AFD integrates prior physical experience independent of its thermosensory apparatus, we next asked whether direct mechanosensory input is required for context-dependent locomotion modulation. The mec-10 gene encodes an amiloride-sensitive Na⁺ channel protein required for sensing body touch. Mutations in mec-10 disrupt mechanotransduction channels in the worms, impairing touch sensitivity [77,78]. We found that mec-10(tm1552) mutants failed to display locomotion adjustment compared to wild type worms (Δspeed: mec-10: 5 ± 4% vs. N2: 33 ± 6%, p = 0.0003; Fig. 5A). These results show that mec-10 is crucial for context-dependent locomotion modulation, supporting the hypothesis that locomotion is modulated depending on the worm’s tactile experience.

Figure 5.

The mec-10 gene is expressed in several mechanosensory neurons, including the six touch receptor neurons (TRNs) and the polymodal nociceptors FLP and PVD [77,79]. To determine which neurons are required for tactile-dependent locomotion modulation, we expressed mec-10 cDNA under cell-specific promoters: mec-18p (TRNs) [80], egl-44p (FLP) [81], or mec-10p (TRNs, FLP, and PVD) [79]. Expression in either FLP or TRNs alone did not restore modulation, as worms carrying egl-44p::mec-10 (Δspeed: −11 ± 4%) or mec-18p::mec-10 (Δspeed: −13 ± 4%) transgenes showed significantly reduced Δspeed compared to wild type (Δspeed: N2: 33 ± 6%; p < 0.0001 for both; Fig. 5A). By contrast, mec-10 co-expression in both FLP and TRNs (Δspeed: 16 ± 4%), or expression from the mec-10 promoter (Δspeed: 23 ± 4%), restored Δspeed to wild type levels (p = 0.20 and p = 0.57, respectively; Fig. 5A). These findings indicate that mec-10 expression across multiple mechanosensory neuron types is required for context-dependent locomotion modulation. It is also worth noting that, while both tactile-dependent locomotion modulation and previously reported spatial preference require FLP, only the former depends on TRNs. Together, these findings suggest that distinct subsets of mechanosensory neurons differentially contribute to behaviors shaped by tactile experience.

AIB interneurons are crucial for tactile-dependent behavioral plasticity

Touch-sensitive neurons that express mec-10, including TRNs, FLP, and PVD, do not form direct synapses with AFD, suggesting that tactile information is relayed through intermediary neurons. Because the interneuron AIB receives synaptic input from FLP and forms electrical synapses with AFD, we hypothesized that AIB could serve as a conduit for mechanosensory signals to reach AFD. To test whether AIB is required for tactile-dependent modulation, we examined locomotion in worms with genetically ablated AIB neurons using npr-9p::caspase expression [82]. AIB-ablated worms failed to adjust locomotion speed, showing a near-complete loss of modulation (Δspeed: −1 ± 5%) compared to wild type (30 ± 8%, p = 0.001, Fig. 5B). These results demonstrate that AIB is required for AFD-mediated tactile-dependent locomotion modulation. However, because mec-10-expressing TRNs are also required, additional pathways beyond AIB likely contribute to transmitting tactile information to AFD, potentially involving indirect synaptic connections through other interneurons or long-distance signaling via neuropeptides or other modulators (Fig. 5C).

Mutations in innexin genes disrupt tactile-dependent modulation

Electrical synapses between AIB and AFD neurons in C. elegans are formed by gap junction proteins known as innexins [83]. If electrical coupling between AIB and AFD contributes to tactile-dependent locomotion modulation, then disruption of innexin genes should impair locomotion adjustments in microfluidic chambers. To test this, we examined loss-of-function mutants of several innexin genes. C. elegans possesses 25 innexin genes [84,85]. Single-cell RNA-seq data from the CeNGEN database [86] indicate that three genes (inx-4, inx-10, and inx-19) are expressed in AFD, one (inx-1) is expressed in AIB, and one (inx-7) is expressed in both. We found that inx-4, inx-7, and inx-10 mutants exhibited a partial but significant reduction in locomotion modulation compared to wild type worms (Δspeed: N2: 45 ± 7%; inx-4: 23 ± 5%, p = 0.037; inx-7: 21 ± 5%, p = 0.013; inx-10: 16 ± 3%, p = 0.0007; Fig. 6A). Mutations in inx-1 also reduced Δspeed values, although the effect did not reach statistical significance (Δspeed: inx-1: 25 ± 6%, p = 0.076), whereas inx-19 mutations had little impact on speed modulation (inx-19: 32 ± 5%, p = 0.51, vs. wild type; Fig. 6A). Furthermore, inx-7; inx-10 double mutants exhibited a more pronounced reduction in Δspeed compared to single mutants (Δspeed: inx-7; inx-10 double mutant: 6 ± 5%; p = 0.032 compared to inx-7 single mutant; Fig. 6B). Together, these findings suggest that multiple innexin genes act redundantly to support tactile-dependent locomotion modulation.

Figure 6.

Targeted electrical coupling between AFD and AIB restores tactile-dependent modulation

We next tested whether restoring inx-10 specifically in AFD would be sufficient to rescue the behavioral defect. Using the AFD-specific srtx-1b promoter, we expressed inx-10 cDNA in inx-10 mutant worms. These transgenic animals displayed significantly improved locomotion modulation (Δspeed: 42 ± 5%) compared to non-transgenic inx-10 mutants (15 ± 4%; p = 0.018; Fig. 6D), indicating that inx-10 expression in AFD alone is sufficient to restore function.

To directly test whether restoring the electrical connection between AFD and AIB could rescue modulation in innexin mutants, we expressed the mammalian gap junction protein Cx36 [87–90] in both AFD (srtx-1Bp) and AIB (inx-1p) neurons in inx-7; inx-10 double mutants (Fig. 6C). Cx36 expression significantly improved locomotion modulation, restoring Δspeed from 7 ± 4% to 35 ± 8% (p = 0.0035; Fig. 6D). These results demonstrate that although innexin mutations disrupt connectivity across multiple neurons, restoring a single targeted electrical connection between AFD and AIB is sufficient to rescue tactile-dependent locomotion modulation, identifying the AIB-AFD electrical synapse as a critical circuit element for tactile-dependent behavioral control.

Discussion

Our study reveals that the thermosensory AFD neuron plays an unexpected role in coupling tactile experiences to locomotion modulation in C. elegans. We show that: 1) C. elegans exhibit distinct locomotion patterns in identical environments, influenced by prior tactile experiences; 2) tactile-dependent behavior modulation requires the AFD neuron but not its thermosensory dendritic endings; 3) thermotaxis and tactile-dependent locomotion modulation rely on distinct sets of cGMP signaling proteins and CNG channel subunits; 4) the AIB interneuron, which connects AFD to mechanosensory circuits, is essential for tactile modulation; and 5) electrical synapses between AFD and AIB are critical for this process. Below, we discuss the implications of these findings.

A thermosensation-independent role for AFD in tactile-dependent behavior modulation

Our findings highlight the remarkable versatility of sensory neurons in shaping behaviors beyond their primary functions. This supports the emerging view that sensory neurons have multifaceted roles in circuit modulation, extending well beyond canonical sensory activities. Such complexity is encoded at multiple levels within neural circuits. In C. elegans, for instance, the ASH neuron drives avoidance responses to various noxious stimuli, including hyperosmolarity, nose touch, and volatile repellents, demonstrating polymodal functionality [91–94]. Similarly, the olfactory receptor neuron AWC exhibits stochastic responses to temperature, introducing variability that enhances the flexibility and reliability of thermotactic behavior [95–97]. In contrast to these examples, the role of AFD in tactile-dependent plasticity does not rely on its sensory endings but is mediated through synaptic connectivity. This reveals an unexpected mechanism by which AFD influences context-dependent behavioral plasticity via network-level interactions rather than direct sensory activity.

Distinct cGMP signaling pathways in AFD mediate sensation and plasticity

Our data indicates that the CNG channel α subunits TAX-4 and CNG-3 play distinct roles in AFD function. Although TAX-4 is essential for AFD to detect temperature and generate thermally evoked responses [40], it is dispensable for tactile-dependent locomotion modulation. By contrast, CNG-3, which is not required for temperature sensing [70], is crucial for tactile-dependent locomotion modulation. Consistent with a broader role in sensory plasticity, CNG-3 is required in AWC neurons for short-term odor adaptation [98] and in AFD for adjusting temperature thresholds based on prior thermal experiences [70]. Our demonstration that CNG-3 is essential for tactile-dependent locomotion modulation further supports the view that CNG-3 enables context-dependent behavioral plasticity rather than directly mediating stimulus detection.

CNG channels are heteromeric complexes composed of both α and β subunits. Our data, along with previous work, show that TAX-2, the CNG β subunit in C. elegans, is required for both tactile-dependent locomotion modulation and thermotransduction [40]. Interestingly, TAX-4, the canonical α partner of TAX-2, is not required for tactile-dependent modulation. This suggests that TAX-2 may support experience-dependent neural plasticity either on its own or by pairing with alternative α subunits such as CNG-3. This type of functional specialization is reminiscent of photoreceptor neurons, where the composition and ratio of CNG channel subunits determine light sensitivity and enable contrast detection under varying light intensities [99–101]. It is worth noting that the tax-2(p694) mutation affects tax-2 expression in four neurons (AFD, BAG, ASE, and ADE) [102]. Moreover, the strong convergence of functional and circuit-level evidence, including the requirement for GCY-18, the rescue of CNG-3 function in AFD, and role of AIB–AFD coupling, points to AFD as the most likely site of TAX-2 function in tactile-dependent locomotion modulation.

In addition to the diversity of CNG channels, guanylyl cyclases in AFD also exhibit functional specificity. For example, mutations in gcy-18 alone significantly disrupt tactile-dependent locomotion modulation without substantially affecting temperature sensing [41,43], whereas gcy-8 is essential for generating thermally evoked Ca²⁺ responses [41,42] yet appears dispensable for tactile-dependent plasticity. These results show that AFD uses distinct sets of guanylyl cyclases to execute dual roles: precise temperature sensing and tactile-dependent behavioral modulation.

Differential regulation of context-dependent plasticity and basal locomotion rates in AFD

Our results further suggest that the mechanisms controlling basal locomotion and context-dependent behavioral modulation are not only distinct but also operate independently. Temperature fluctuations significantly affect basal locomotion rates, i.e., higher temperatures accelerate speed, while lower temperatures slow it down [67,68]. However, tactile-dependent locomotion modulation in microfluidic chambers remains robust despite these temperature-induced changes in basal movement. This independence is further supported by our observations in tax-4 and kcc-3 mutants, which display elevated basal locomotion rates yet still adjust their locomotion in response to tactile experiences. These findings indicate that the processes governing basal locomotion and context-dependent modulation are controlled by separate, non-interfering mechanisms, even though both pathways engage cGMP signaling.

Electrical synapses link the thermosensory neuron AFD to tactile-dependent behavioral modulation

AIB-AFD electrical synapses are pivotal for tactile-dependent behavioral plasticity. Disrupting this connection impairs tactile-dependent locomotion modulation, whereas restoring an engineered electrical synapse rescues this behavior. The successful replacement of invertebrate innexins with vertebrate connexins, which perform analogous functions despite lacking sequence homology, indicates that the core function of these synapses lies in their connectivity rather than in their specific protein sequences. This observation suggests that the transfer of signaling molecules such as Ca²⁺ and cGMP through electrical synapses may facilitate the coordination of neuronal activity across distinct networks [103–106].

Together, these findings support a model in which AIB functions as a hub neuron that relays mechanosensory input from FLP to AFD to modulate locomotion (Fig. 5C). However, because electrical synapses are often bidirectional, information flow may also occur in the opposite direction, from AFD to AIB. Determining the directionality and dynamics of this signaling remains an important question for future investigation. A similar circuit logic has been proposed for thermal avoidance behavior, in which AIB integrates inputs from FLP via chemical synapses and couples to AFD via electrical synapses, forming an FLP-AIB-AFD circuit. In that case, the directionality of information flow also remains unresolved [107].

Finally, the proposed circuit mechanism for tactile-dependent modulation does not exclude additional routes of modulation. These may include extra-synaptic signaling mechanisms, such as neuropeptide release from other touch-sensitive neurons. While the precise molecular and circuit mechanisms remain to be elucidated, our findings establish that AIB-AFD electrical synapses are essential for incorporating tactile context into locomotor control.

In summary, our findings reveal an unexpected role for AFD in tactile-dependent behavioral plasticity. They show that specialized molecular mechanisms, including distinct cGMP signaling pathways, diverse guanylyl cyclases, and functional electrical synapse connectivity, enable AFD sensory neurons to perform dual functions: detect stimuli and support tactile-dependent plasticity. A limitation of this study is that the directionality and mode of information flow between AFD and AIB remain unresolved, and defining this relationship will be an important goal for future investigation. Nevertheless, our results establish the AFD-AIB connection as a regulatory hub in tactile-dependent locomotor modulation and expand our understanding of sensory neuron versatility in adaptive behavior.

Methods

Nematode Strains and Growth

C. elegans strains were maintained under standard conditions at 20°C on nematode growth medium (NGM) agar plates [108]. These plates were seeded with E. coli OP50 lawns. Strains used in this study are listed in Table S1 and are available from the corresponding author upon request.

Molecular Biology

All DNA constructs were sequence-verified. cDNA sequences encoding mec-10, cng-3, and inx-10 were amplified from a cDNA library made with the SuperScript III Reverse Transcriptase kit (Thermo Fisher Scientific, MA) from total RNA extracted from adult N2 worms. A codon-optimized cDNA sequence for the mouse gap junction protein connexin-36 (Cx36) was previously described [109]. DNA plasmid constructs for C. elegans expression were assembled by LR recombination using the MultiSite Gateway Three-Fragment Vector Construction Kit (Thermo Fisher Scientific, MA). Entry clones were constructed as follows: (i) promoter sequences flanked by attL4 and attR1 sites; (ii) cDNA sequences flanked by attL1 and attL2 sites; and (iii) sl2::mNeonGreen::rab-3 3′utr flanked by attR2 and attL3 sites. Destination vectors used for LR recombination included the pDEST R4-R3 Vector II and a custom-built destination vector BJP-C957, which contains a hygromycin resistance cassette for selection, an intestinal fluorescent marker (vha-6p::mScarlet::tbb-2 3’utr) for visual screening transgenic worms, and flanking mos1 sequences to facilitate homologous recombination for Mos1-mediated single-copy insertion (MosSCI). Promoters used for tissue- or cell-specific expression were: gcy-18p (2 kb) and srtx-1Bp (493 bp) for AFD; mec-18p (390 bp) for TRNs; egl-44p (5 kb) for FLP; inx-1p (1kb) for AIB, and mec-10p (3 kb) for expression in TRNs, FLP, and PVD. Plasmids used to generate transgenic strains in this study are listed in Table S2.

Transgenes, Germline Transformation, and Genome Editing

Strains carrying extrachromosomal arrays were generated by microinjecting various plasmids together with co-injection markers. The co-injection markers included unc-112p::gfp (30 ng/µl), rpl-28p::neoR (20 ng/µl), and Cbr-unc-119 (20 ng/µl). The blank vector pBluescript was used as filler DNA to bring the final total DNA concentration to 100 ng/µl. A detailed list of the plasmids used to create transgenic worms in this study is provided in Table S2.

Mos1-mediated transgene insertion [110,111] was used to generate single-copy transgenic animals. Briefly, worms carrying the ttTi5605 Mos1 site on chromosome II were injected with a plasmid mixture containing the transgene of interest (40 ng/µl), along with BJP-B908, a plasmid expressing Cas9 and a guide RNA targeting the Mos1 site (60 ng/µl). Selection markers included rps-0p::hygR for hygromycin resistance and vha-9p::mScarlet for intestinal red fluorescence to enable visual screening.

The gcy-18::gfp11×7 knock-in strain was generated using a modified CRISPR–Cas9 protocol employing preassembled Cas9 ribonucleoprotein (RNP) complexes, co-injected with a single-stranded DNA repair template to facilitate homologous recombination [112,113]. Cas9 protein and crRNA (targeting sequence: TACAACGACTGAAAGAGGAG) were obtained from Integrated DNA Technologies (IDT, NJ). The single-stranded DNA repair template for generating gcy-18::gfp11×7 was synthesized by Twist Bioscience. Successful editing produced the pek331 allele, which carries an in-frame GFP11×7 insertion at the C-terminus of GCY-18. The insertion was verified by PCR and DNA sequencing. A worm strain, BJH833, carrying both pek331 and muIs253 [eft-3p::sfgfp1-10::unc-54 3’utr], was used to visualize GCY-18 by complemented split-GFP fluorescence [114].

Fabrication of PDMS Microfluidic Chambers

Microfluidic chamber designs were created in AutoCAD and fabricated as 100 µm SU8-silicon molds using photolithography (FlowJem, NO, Canada). PDMS devices were produced by pouring degassed PDMS prepolymer (Sylgard 184, Dow Corning, MI) onto the molds, degassing again under vacuum (−25 inHg, 15 min, room temperature), and curing at 80 °C for 2 hours. The cured PDMS was peeled from the mold, and inlet/outlet ports were punched using a 2 mm biopsy punch (Acuderm Inc., FL). For the chambers with a removable barrier shown in Figure 1–Supplement 1A, cast PDMS chamber structures were bonded onto a thin semi-cured PDMS sheet to create a permanent, watertight seal. To prepare the sheet, ∼30 ml of PDMS was poured into a 6-inch Petri dish, degassed for 15 min, and partially cured at 65 °C for 15-20 min, until firm but still tacky. PDMS chamber structures were gently placed onto the sheet with ∼2 mm spacing to allow the formation of a barrier, then fully cured at 80 °C for 30 min. PDMS devices were bonded to 1 mm glass microscope slides using oxygen plasma treatment (PDC-32G, Harrick Scientific, NY, 1 min).

Behavior Recording and Analysis

Videos were captured for 10 seconds at 7.5 frames per second using a Basler acA2440 camera (Basler, PA) equipped with an AF Micro-NIKKOR 60 mm lens (Nikon, NY). Behavioral data were extracted from the videos using WormLab software (MBF Bioscience, VT) through individual worm tracking. Only those worms that were already within the assay area at the start of the recording were analyzed, and any worms that entered later were excluded. Crawling speed was determined by the total distance traveled by a worm during the 10 second recordings. A turn was defined as a change in crawling direction of 15° or more. Reversals were defined as sudden movement in the opposite direction for at least 5 frames (∼0.6 seconds). Worms were considered idle if their absolute velocity (distance traveled between frames) was below 20 µm/s for more than 10 frames (∼1.3 seconds).

For behavior analysis, worms were age-synchronized by transferring 30-50 young adult worms onto an NGM plate seeded with OP50 and allowing them to lay eggs for 2 hours before removal. The plates were then kept at room temperature (22°C) for 3 days, allowing the eggs to develop into young adults, identified by the presence of 5-10 eggs arranged in single rows within the gonads of all worms on the plate. To wash the synchronized adults off the plate, worms were gently transferred into Eppendorf tubes using glass Pasteur pipettes. After allowing the worms to settle at the bottom of the tube (∼1 minute), the supernatant was removed and replaced with fresh M9 buffer. This washing step was repeated three times to remove residual OP50. Next, approximately 30-70 worms were loaded into an inlet of the chambers shown in Figure 1 using a 20µl pipette and allowed to roam freely in the chamber for 60 minutes before behavior in the assay area was recorded.

To assay worms confined to the assay area by a barrier (Fig. 1–Supplement 1A), PDMS chambers were prepared with a removable gap separating the training and assay zones. M9 buffer was first injected into the exploration zone, after which ∼100 worms were loaded and allowed to explore for 10 minutes, during which they moved only within the reach of the M9 buffer, typically up to the gap between the exploration and assay zones. To remove the barrier, additional M9 buffer was gently introduced into the assay zone until it merged with the buffer, allowing worms to move freely between zones. After 15 minutes of exploration, the barrier was reintroduced by completely drying the space between the zones with a Kimwipe, once again confining worms to areas containing M9 buffer. Behavior in the assay zone was recorded immediately after the barrier was reinserted and again 40 minutes later.

Fluorescence microscopy

Widefield fluorescence microscopy was performed using a Leica DMi8 inverted microscope equipped with 63x/1.4NA and 100x/1.4NA oil-immersion objectives (Leica Biosystems, NJ). Worms were mounted on thin 3% agar pads and anesthetized with 10 mM levamisole diluted in M9 buffer. Images were acquired with an Andor iXon-888 EMCCD camera (Oxford Instruments, England), with an exposure time of 150 ms and an electron-multiplying (EM) gain of 200. Acquisition settings were kept constant across experiments.

Statistical Analysis

Behavior measurements were calculated for each individual chamber, and these chamber-specific values were then used to generate the final mean values shown in the bar plots. The sample size was defined as the number (approximately 25) of chambers analyzed per experiment. An unpaired Student’s t-test with Bonferroni correction was used to compare average speeds between uniform and binary chambers. To quantify the relative difference between groups, we calculated the percent change for each value in the uniform chambers relative to the mean of value from the binary chambers. Specifically, we subtracted the mean of the values from the binary chamber from each individual value in the uniform chamber. The result was then divided by the mean of the binary chamber and multiplied by 100 to express the difference as a percentage. For comparisons among multiple strains, we conducted one-way ANOVA followed by a Tukey–Kramer post hoc test on all possible pairwise comparisons. All statistical analyses were performed using R.

Figure supplements

Figure 1–Supplement 1.

Figure 2–Supplement 1.

Figure 3–Supplement 1.

Fig. 4–Supplement 1.

Figure 4–Supplement 2.

Table S1.

Table S2.

Data availability

The data supporting the findings of this study are available in Dryad at https://doi.org/10.5061/dryad.k3j9kd5pz.

Acknowledgements

This research was supported by NIH grants R01NS109476 and R01NS115974 to JB. MR was supported by T32GM136534 and F31NS129545. The authors thank the Caenorhabditis elegans Genetics Center (funded by the NIH Office of Research Infrastructure Programs P40 OD010440) for providing worm strains. We also thank Lin Zhang and Yan Liu for their support in constructing DNA plasmids and generating transgenic worm strains. Special thanks to Melissa Woo (Mercer Island High School), Tabitha Ngo, and Sahithi Challam (Interlake High School) for their assistance with worm maintenance. We appreciate the members of the Bai lab for their critical reading of the manuscript. Additionally, we acknowledge the multiple mentorship programs at Fred Hutchinson Cancer Center, including the SURP Undergraduate Researchers and Pathways Undergraduate Researchers programs, supported in part by the Cancer Center Support Grant P30 CA015704, for fostering relationships with high school and undergraduate trainees who contributed to this project.

Additional information

Funding

HHS | NIH | National Institute of Neurological Disorders and Stroke (NINDS) (R01NS109476)

• Jihong Bai

HHS | NIH | National Institute of Neurological Disorders and Stroke (NINDS) (R01NS115974)

• Jihong Bai

HHS | NIH | National Institute of General Medical Sciences (NIGMS) (T32GM136534)

• Manuel Rosero

HHS | NIH | National Institute of Neurological Disorders and Stroke (NINDS) (F31NS129545)

• Manuel Rosero

HHS | NIH | National Cancer Institute (NCI) (P30CA015704)

• Jihong Bai