Reactive oxygen species drive foraging decisions in Caenorhabditis elegans

Abstract

Environmental surveillance-mediated behavior integrates multiple cues through complex signaling mechanisms. In Caenorhabditis elegans, neurons coordinate perception and response through evolutionarily conserved molecular signaling cascades to mediate attraction and avoidance behaviors. However, despite lacking eyes, C. elegans was recently reported to perceive and react to the color blue. Here, we provide an explanation for this apparent color perception. We show that internally-generated reactive oxygen species (ROS) occurring in response to light are additive to exogenous sources of ROS, such as bacterial toxins or photosensitizers. Multiple sub-threshold sources of ROS are integrated to coordinate behavioral responses to the environment with internal physiologic cues, independent of color. We further demonstrate that avoidance behavior can be blocked by antioxidants, while ROS is both sufficient and scalable to phenocopy the avoidance response. Moreover, avoidance behavior in response to ROS is plastic and reversible, suggesting it may occur through a post-translation redox modification. Blue light affects C. elegans behavior through ROS generation by endogenous flavins in a process requiring the neuronal gustatory photoreceptor like protein, LITE-1. Our results demonstrate that LITE-1 is also required for ROS-mediated avoidance of pyocyanin and light-activated photosensitizers and this role is mediated through the modification of Cys44. Overall, these findings demonstrate that ROS and LITE-1 are central mediators of C. elegans foraging behavior through integration of multiple inputs, including light.

Graphical abstract

Highlights

• •C. elegans avoidance of pyocyanin and photosensitizers is potentiated by light.
• •ROS generated internally causes changes in C. elegans behavior.
• •LITE-1 plays a role in avoidance of ROS independent of blue light.
• •Redox modification of gustatory receptor LITE-1 is necessary for avoidance behavior.

1. Introduction

Animals monitor their environment to search for food and avoid danger as a means of protection [1,2]. Microorganisms, such as bacteria and yeast found in the soil, make up the food source of Caenorhabditis elegans [3]. Soil-borne microorganisms produce toxins to protect themselves, which causes C. elegans to leave the food source, a response known as avoidance behavior [[4], [5], [6]]. Antimycin A, a toxin that causes production of mitochondrial reactive oxygen species (ROS), can cause avoidance behavior [7,8]. Pseudomonas aeruginosa is a soil-borne bacterium that elicits avoidance behavior in C. elegans due to a wide range of toxins it produces, including those that target mitochondria [[9], [10], [11]]. In addition to toxins, C. elegans avoids blue light [[12], [13], [14], [15]]. Blue light can have adverse effects through the generation of ROS by activating endogenous photosensitizers, such as flavins [[16], [17], [18]].

Pyocyanin is a blue-pigmented toxin produced by P. aeruginosa, which redox cycles with coenzyme Q to generate superoxide (O₂^•-) [[19], [20], [21], [22], [23], [24]]. Recent work showed that C. elegans avoids pyocyanin in a light-dependent manner [25]. Despite lacking a visual system, C. elegans avoids toxins in the presence of blue light, suggesting that color guides foraging decisions. This avoidance behavior required the protein LITE-1 [15,25], a C. elegans taste-receptor homologue that is activated by UV and blue light [[12], [13], [14], [15]]. Light can induce ROS generation through excitation of photosensitive compounds [26]. Flavins are endogenous photosensitizers that absorb blue light and generate ROS [27]. C. elegans recognizes and avoids blue and UV light through the LITE-1 protein, which is regulated by the generation of hydrogen peroxide (H₂O₂) [[13], [14], [15],28]. However, how blue light and H₂O₂ interact to regulate LITE-1 activity at the molecular level remains unclear. Recent work has shown LITE-1 has oxidizable cysteine residues required for function [29]. LITE-1 plays an important role in mediating avoidance of pyocyanin in the presence of white light, as well as the non-pigmented redox cycler paraquat when exposed to blue light [25]. Both pyocyanin and paraquat generate O₂^•-; however, the role of ROS in mediating avoidance was unclear [30,31]. ROS, often considered damaging, can also play a role in beneficial signaling [[32], [33], [34], [35]]. Whether ROS acts in a damaging or protective role depends on a variety of factors, including quantity and timing [36,37]. However, the molecular targets of ROS in signaling cascades are yet to be fully elucidated.

We sought to determine if ROS was mediating the perception of the color in C. elegans. We found that antioxidants prevent the light-dependent avoidance of pyocyanin. C. elegans exhibits avoidance behavior towards photosensitizers in response to wavelength-specific light, which correlates with the amount of superoxide generated. Avoidance behavior was reversible, with avoidance ending when the light is removed. Combined, these findings confirmed that ROS is necessary and sufficient for mediating avoidance of pyocyanin and the photosensitizers tested. LITE-1 plays a key role in mediating avoidance of blue light and ROS, independently, with a cysteine residue playing a necessary role in signaling. In sum, our results show that ROS is a central mediator of avoidance behavior and can act through LITE-1.

2. Material and methods

2.1. Worm strains and maintenance

C. elegans were maintained at 20 °C on nematode growth medium (NGM) plates seeded with E. coli strain OP50 as a bacterial lawn [38]. Wild-type C. elegans (N2 Bristol) and lite-1(ce314) were sourced from the Caenorhabditis Genetics Center. APW364 (lite-1(jbm51 [lite-1 C44S]) X) and APW387 (lite-1(jbm51 [lite-1 C44D]) X) were generated as described in the following section.

2.2. Transgenic C. elegans strains

CRISPR/Cas9 was used to generate point mutations in lite-1. A lite-1 crRNA (CCTGTCGAGTCGTAACAAAT) (0.8 μg/μL) and single-stranded oligo nucleotide (ssODN) repair template (100 ng/μL) were injected into adult C. elegans gonads along with a mix containing 25 mM KCl, 7.5 mM HEPES, 4 μg/μL tracrRNA, 0.8 μg/μL lite-1 crRNA, 0.8 μg/μL dpy-10 crRNA, 50 ng/μL dpy-10 ssODN, and 2.5 μg/μL purified Cas9. Three days after injections, progeny were screened for the dpy phenotype. Single-worm PCR followed by sequencing of the PCR product was performed on dpy worms to confirm presence of the mutation. Strains with the mutation were outcrossed with wild-type (N2) worms to remove the dpy phenotype. Sequences for repair templates and primers are as follows:

LITE-1 C44S Repair Template:

TAGCAAGAAGACCATGATCGCAAAAATCCTGTCGTCGCGAAATAAGTGGGCGATTTCAGATCGAACGCTCTACCCAATTTACTATCTTCTGCT.

LITE-1 C44D Repair Template:

TAGCCAAGAAGACCATGATCGCAAAAATCCTGTCGTCGCGAAAtAAGTGGGCGATTGATGATCGAACGCTCTACCCAATTTACTATCTTCTGCT

LITE-1 Forward Primer:

TGCAGTCTATCGGTGCACAA.

LITE-1 Reverse Primer:

GGAGACAAAAGACCAACCAGC.

2.3. Avoidance assays

Plates had a concentration gradient of drug radiating from the center of the plate. Compounds (30 μL) were combined with OP50 E. coli (10 μL) and pipetted directly to the center of a 35 mm petri dish containing 4 mL of NGM. The drugs used in this study include pyocyanin (Cayman Chemical, 10009594), rose bengal (Sigma, 632-69-9), methylene blue (Sigma, 457250), neutral red (Alfa Aesar, J65643.18), and antimycin A (Sigma, A8674). Concentrations of drugs for corresponding experiments are provided in figure legends. The drug was allowed to diffuse for 3 h, after which 25–35 day one adult worms were placed directly onto the lawn of the plate. Plates were then placed on a white LED light source (Metaphase, 83873) (Fig. S1), emitting 2.46 μW/mm,² or in the dark. After 2 h (0.018 J/mm²), worms on and off the bacterial lawn were counted to determine the fraction of worms avoiding the lawn. In experiments using specific wavelengths of light, Kodak Wratten Filters were used (Red, No. 29 (640–700 nm, 1.54 μW/mm², 0.011 J/mm²); green, No. 58 (500–560 nm, 1.14 μW/mm², 0.008 J/mm²); blue, No. 47 (420–480 nm, 1.96 μW/mm², 0.014 J/mm²)).

2.4. Body bend assays

Day 1 adult worms were moved to a 60 mm agar plate without food. An initial speed was taken by counting the total number of body bends over a 15 s time period (0.836 μW/mm², 0.006 J/mm²) using a MVX10 Fluorescence MacroZoom dissecting microscope (Olympus), which was then multiplied by 4 to get body bends/minute. Worms were then exposed to green (540–580 nm, 435 μW/mm², 0.003 J/mm²) or blue (440–470 nm, 218 μW/mm², 0.0016 J/mm²) light and body bends were again counted for 15 s. For recovery assays, after blue light exposure, worms were allowed to recover in the dark for 30 s prior to counting body bends for 15 s.

2.5. C. elegans drug treatments

L4 worms were grown on NGM plates containing 100 μM EUK-134 (Millipore Sigma, SML0743), 10 mM 4-hydroxy-TEMPO (Millipore Sigma, 176141), 10 μM mitoTEMPO (Cayman, 16621), or 650 μM rose bengal. Worms were cultured for 24 h and then subjected to avoidance or body bend assays, as described above.

2.6. Avoidance reversibility assays

Worms placed on a lawn of OP50 with 6.4 μM rose bengal were exposed to white light (2.46 μW/mm², 0.018 J/mm²) for 2 h. After 2 h, worms that were avoiding the lawn were moved to a fresh plate containing 6.4 μM rose bengal and OP50. Plates were then either exposed to light or placed in the dark. Worms off the lawn were counted every hour for 2 h following the transfer to determine the fraction avoiding under each condition.

2.7. Quantification of superoxide

Dihydroethidium (DHE, ThermoFisher Scientific, D11347) reacts with superoxide to produce the fluorescent molecule 2-hydroxyethidium (2-OHE⁺), which was separated from other DHE oxidation products using HPLC, as previously described [18,[39], [40], [41]]. Briefly, photosensitizers at the indicated concentrations were diluted in D-PBS (phosphate buffer saline containing 0.1 mM diethylenetriaminepentaacetic acid) in the presence of 100 μM DHE and illuminated with white light (2.46 μW/mm², 0.0015 J/mm²) for 10 min. After 10 min, 100 μL were removed and diluted in 100 μL of 200 mM HClO₄/MeOH and placed at 4 °C for 30 min. The samples were then centrifuged at 17,000×g for 20 min and the supernatants were transferred to an equal volume of 1 M K ⁺PO₄^- (pH 2.6). The samples were separated via HPLC on a Polar-RP column (Phenomenex, 150 x 2 mm; 4 μm) using a Shimadzu HPLC and fluorescence detection (RF-20A). A flow rate of 0.1 mL/min was held constant using a gradient of mobile phase A (10% ACN, 0.1% TFA, 89.9% H₂O) and mobile phase B (60% ACN, 0.1% TFA, 39.9% H₂O). The gradient was as follows: 0 min, 40% B; 5 min, 40% B; 25 min, 100% B; 30 min, 100% B; 35 min, 40% B; 40min, 40% B. Standard curves were generated using known concentrations of 2-OHE⁺ and peaks were quantified using Shimadzu Lab Solutions [39,42].

2.8. Statistics

All the data were analyzed by an unpaired, two-tailed t-test or one- or two-way ANOVA with post hoc multiple comparison correction as indicated in the figure legends. P values < 0.05 were considered significant (GraphPad Prism 9).

3. Results

3.1. C. elegans avoids pyocyanin in response to ROS generation

Pyocyanin, a blue pigmented toxin produced by Pseudomonas aeruginosa [19,43,44], can cause mitochondrial damage and induce apoptosis [23,[45], [46], [47]]. Previous work suggested that because C. elegans avoids pyocyanin in response to light, and that blue light enhances avoidance of the ROS-producing toxin paraquat, avoidance was due to discrimination between colors [25]. However, blue light can activate endogenous photosensitizers to produce ROS [27,48]. Pyocyanin and paraquat are redox cyclers that generate O₂^•- independent of light [31,49,50]. We hypothesized that ROS generated under these conditions acted as a mediator of avoidance behavior. To test this hypothesis, we combined pyocyanin with E. coli OP50 in the middle of the plate. Worms were placed directly on this mixture and either exposed to light or placed in the dark for 2 h, after which the number of worms on and off the E. coli lawn was counted to quantify avoidance behavior (Fig. 1A). As previously reported [25], we found that light increased avoidance of the lawn containing pyocyanin (Fig. 1), while there was no effect in the presence of a non-toxic blue dye (Fig. S2). To test if avoidance is dependent on the combination of light and ROS, we treated C. elegans with antioxidants and employed an avoidance behavior assay. The antioxidants EUK-134, a catalase and superoxide dismutase mimetic [51], the O₂^•- scavenger TEMPO, and mitochondrial-targeted TEMPO, mitoTEMPO [52,53], were used. Superoxide, which can be generated by photosensitizers and redox cyclers such as pyocyanin [54,55], is rapidly converted to H₂O₂ [56]. H₂O₂ is considered a second messenger that can modify protein thiol residues to induce redox signaling [[57], [58], [59]]. Treating worms with antioxidants prior to avoidance assays eliminated the light-dependent behavior (Fig. 1B–D). The ability of antioxidants to mitigate avoidance behavior indicates that internally generated ROS is necessary for the light-dependent response to pyocyanin.

Fig. 1.

ROS generation is required for avoidance of pyocyanin. (A) Schematic of the avoidance assay. A mixture containing the drug and E. coli OP50 were mixed and pipetted directly onto the middle of a 35 mm plate. The plates were allowed to dry for 3 h, followed by the addition of day 1 adult wild-type (N2) C. elegans. Plates were placed in the dark or on a light box (2.46 μW/mm²). After 2 h, worms on and off the plate were enumerated to determine the fraction avoiding. (B–D) Pre-treating worms with antioxidants decreases avoidance of pyocyanin (15.6 mM). Worms were grown on plates containing (B) 100 μm EUK-134, (C) 10 mM TEMPO, or (D) 10 μM mitoTEMPO for 24 h prior to avoidance assay. Black circles indicate no light and white circles indicate white light (2.46 μW/mm²). Dots represent an independent trial of 25–35 worms. Bars are represented as means ± standard deviation. ****p < 0.0001 (Two-way ANOVA, Sidak’s multiple comparison).

3.2. C. elegans avoidance behavior is independent of perceived color

Since the avoidance of pyocyanin depended on light, we tested whether pyocyanin can act as a photosensitizer and generate ROS in response to light. We measured the ability of pyocyanin to generate O₂^•- using HPLC separation of dihydroethidium (DHE) products. DHE interacts with O₂^•- to selectively generate the fluorescent molecule 2-hydroxyethidium (2-OHE⁺) or nonspecifically interacts with electron donors to generate ethidium (E⁺) [41,60,61]. We used HPLC to separate the DHE oxidation products, 2-OHE⁺ and E⁺, to quantify O₂^•-. Pyocyanin immediately converted the majority of DHE to E⁺ (Fig. S3) independent of light, likely due to its redox cycling abilities, and precluded measuring 2-OHE⁺ production in response to light.

To examine the wavelength dependence of avoidance behavior, we measured the absorbance spectrum of pyocyanin and found that pyocyanin has a broad absorption spectrum (Fig. 2A). When pyocyanin was exposed to blue or green light there was no increase in avoidance behavior. Hence, the avoidance of pyocyanin could not be potentiated by light filtered by specific colors, but required white light. It was surprising that blue light, which is generally considered to result in ROS generation [16,62], did not potentiate avoidance of pyocyanin. It is possible that the blue filter substantially reduces the number of photons absorbed by the endogenous photosensitizers, decreasing the amount of ROS generated by the photosensitizers, and overall, in C. elegans. We then wanted to determine if this avoidance behavior would extend to additional pigmented molecules that could generate ROS or if this response was exclusive to pyocyanin.

Fig. 2.

Avoidance of photosensitizers requires wavelengths of light that can be absorbed. A-D) A wavelength scan was taken of each chemical pyocyanin (A), methylene blue (B), rose bengal (C), and neutral red (D) to determine its absorbance spectrum. x-axis indicates absorbance (400–700 nm) and y-axis indicates absorption. Avoidance assays were performed by placing plates in the absence (black circle) or presence of light (white circle, 2.46 μW/mm²). Wavelengths were filtered and worms were exposed to red (red circle, 1.54 μW/mm²), blue (blue circle, 1.96 μW/mm²), or green light (green circle, 1.14 μW/mm²) to determine avoidance of each chemical in response to specific wavelengths of light. Pyocyanin (15.6 mM), methylene blue (387.5 μM), rose bengal (25.6 μM), and neutral red (50 mM). Dots represent an independent trial of 25–35 worms. Bars are represented as means ± standard deviation. *p < 0.0332, **p < 0.0021, ***p < 0.0002, ****p < 0.0001 (One-way ANOVA).

The photosensitizer methylene blue was chosen because, while blue in color, it makes ROS selectively in response to red light [63,64], allowing deconvolution of their respective contributions to avoidance behavior. C. elegans did not avoid methylene blue in the dark, but when exposed to white light, the number of worms avoiding the lawn increased (Fig. 2B). To determine if C. elegans was perceiving color or responding to ROS, we used a variety of a range of wavelengths of light in the visible spectrum (red, green, and blue). Avoidance of methylene blue only occurred when worms were exposed to red light, corresponding to the absorption spectrum of methylene blue (Fig. 2B). However, blue light had no effect on avoidance behavior. This finding highlights the requirement of ROS generation for avoidance behavior, irrespective of color.

Further support for this idea came from testing several other photosensitizers of differing colors and absorbance spectra [65]. Rose bengal, a well-known and efficient photosensitizer, is pink and absorbs in the blue/green spectrum [[66], [67], [68], [69]], while neutral red is a red dye with a wide absorption spectra [70]. Similar to methylene blue, excitation light within the absorption spectrum of rose bengal or neutral red facilitated avoidance behavior (Fig. 2C–D), despite neither of these photosensitizers being blue. These results with multiple photosensitizers demonstrate that blue color in a molecule is neither necessary nor sufficient to impact avoidance behavior, but using light within an absorbance spectrum that results in ROS production is sufficient.

3.3. Avoidance of photosensitizers correlates with superoxide generation

We hypothesized that if ROS was playing a role in avoidance behavior, higher levels of ROS would cause an increase in the incidence of behavioral change. To test this hypothesis, we correlated the amount of O₂^•- generated by photosensitizers to the number of worms avoiding the lawn. Four different concentrations were used for each photosensitizer in the avoidance and the O₂^•--measurement assays. Rose bengal and methylene blue avoidance correlated with O₂^•- generation, while neutral red had a weaker correlation (Fig. 3). While rose bengal and methylene blue required high amounts of O₂^•- production for avoidance to begin, neutral red avoidance began at much lower levels of O₂^•- production (Fig. 3), suggesting differences in uptake amongst the photosensitizers. These results show a dose-dependent response between O₂^•- generation and avoidance behavior.

Fig. 3.

Superoxide generation correlates with C. elegans avoidance. (A) 2-hydroxyethidium production (x-axis, 2-OHE⁺) by (A) rose bengal, (B) methylene blue, or (C) neutral red, was correlated with the number of worms avoiding that concentration on an agar plate (y-axis). The concentrations of rose bengal (A, from left to right) are 6.4 μM, 12.8 μM, μM, 25.65 and 51.3 μM. R² = 0.8933. The concentrations of methylene blue (B, from left to right) are 38.75 μM, 77.5 μM, 387.5 μM, and 775 μM. R² = 0.8271. The concentrations of neutral red (C, from bottom to top) are 12.5 mM, 25 mM, 50 mM, and 75 mM. The R² = 0.5036.

3.4. Internally generated ROS alters behavior of C. elegans

Antioxidants mitigated the light-dependent avoidance of pyocyanin (Fig. 1B–D). We hypothesized that O₂^•- plays a role in mediating this avoidance behavior. To test this hypothesis, we treated C. elegans with EUK-134 or TEMPO and exposed them to rose bengal to determine if the avoidance behavior was due to internally-generated ROS [53,71]. Worms were again grown on plates containing EUK-134 for 24 h prior to placing them on a plate with rose bengal mixed with E. coli. Pre-treating worms with EUK-134 decreased the fraction of worms avoiding the lawn, showing a role for O₂^•- or H₂O₂ in mediating avoidance of rose bengal (Fig. 4A). However, light-dependent avoidance behavior was observed in worms treated with EUK-134. This is likely due to the large amount of ROS generated by rose bengal (Fig. 3A), which the EUK-134 could not completely overcome. A similar phenomenon, a reduction in avoidance of rose bengal, was observed when pre-treating with TEMPO (Fig. 4B). Since the antioxidants were only inside the worms, rather than on the plate, all ROS reduction would occur internally. Antioxidants mitigated the avoidance of rose bengal, indicating that internal ROS plays a role in mediating this avoidance behavior.

Fig. 4.

Internal ROS generation drives avoidance of the photosensitizer rose bengal. A-B) Worms were grown on plates containing 100 μM EUK-134 (A) or 10 mM TEMPO (B) for 24 h prior to use in avoidance assays. Worms were then placed on lawns of OP50 containing rose bengal (6.4 μM) and exposed to light (white circle, 2.46 μW/mm²) or in the dark (black circle). After 2 h, worms on and off the lawn were counted to determine the fraction avoiding. Each square represents an independent trial of 25–35 day 1 adult worms. Bars are represented as means ± standard deviation. **p < 0.0021 ****p < 0.0001 (Two way ANOVA, Sidak’s multiple comparison) (C) Exposing C. elegans pre-treated with rose bengal to green light increases movement speed. Worms were grown on 650 μM rose bengal for 24 h. Single worms were placed on an unseeded plate and initial speed was taken at low light exposure (black circle, 0.836 μW/mm²). Speed was again taken during exposure to green light (green circle, 435 μW/mm²). Each dot represents one worm. ****p < 0.0001 (Two way ANOVA, Sidak’s multiple comparison) D) C. elegans pre-treated with rose bengal avoids bacteria in the presence of light. Worms were grown on rose bengal (650 μM) for 24 h and placed on a plate with a lawn of OP50 bacteria. Plates were exposed to light (2.46 μW/mm²) for 10 min before worms on and off the lawn were counted to determine the fraction avoiding. Each circle represents an independent trial of 25–35 day 1 adult worms. Bars are represented as means ± standard deviation. ****p < 0.0001 (Student’s t-test).

Internally-generated ROS was necessary for avoidance behavior (Fig. 1B–D), and we next tested if ROS was sufficient by pre-treating C. elegans with rose bengal. A body bend assay was used to determine changes in speed of C. elegans in response to green light. C. elegans move in a sinusoidal pattern and when exposed to unfavorable conditions, they increase speed to escape. We quantified speed as body bends per minute for a more acute measurement of avoidance behavior. Treating with rose bengal alone did not alter the baseline speed of C. elegans (Fig. 4C). Generation of ROS, through green light exposure, immediately changed movement speed. Green light had no effect on worms that were not pre-treated, showing that the ROS generated by the internal rose bengal caused the behavioral changes. Using green light also separates the effects of blue light, which impacts behavior independently [13]. We subjected rose bengal-treated C. elegans to the avoidance assays, finding that worms avoided the lawn after light exposure. Taken together, these results show that ROS is sufficient to alter movement speed and induce avoidance behavior.

ROS can induce signaling through reversible modifications. We next tested if the behavioral effect is reversible. Rapid clearance of ROS by cellular antioxidants can remove the ROS signal and reverse potential redox modifications [72]. C. elegans are attracted to bacteria, remaining on a lawn of bacteria in a behavior known as dwelling [73]. It is possible that high levels of ROS could damage neurons and permanently alter food sensing, preventing dwelling behavior. To test if avoidance behavior was reversible, we examined how worms avoiding ROS would respond when the ROS source was removed. C. elegans was initially placed on a plate containing a lawn of rose bengal and bacteria (OP50). Worms avoiding the rose bengal (off the bacterial lawn) were transferred to new plates with rose bengal and OP50 and either exposed to light or placed in the dark. In the dark, C. elegans no longer avoided the food, showing that removing the ROS source is sufficient to reverse avoidance behavior (Fig. 5). Worms that continued to be exposed to light avoided the lawn at a similar rate as the original plate. These results show that the avoidance behavior is reversible contingent on the elimination of ROS generation.

Fig. 5.

Avoidance of rose bengal is reversible. Day 1 adult worms are placed on a lawn of OP50 containing 6.4 μM rose bengal and exposed to light for 2 h (first white circle, 2.46 μW/mm²). After 2 h, worms that were off the lawn were removed and placed on a new plate with a lawn containing OP50 and 6.4 μM rose bengal. The plates with these worms were then placed in the dark (black circle) or in the light (second white circle, 2.46 μW/mm²) and the number of worms avoiding the lawn were counted every hour for 2 h. Circles (first bar) represent an independent trial of 45–50 day 1 adult worms. Circles in second and third bars represent independent trials of 15–25 worms. Bars are represented as means ± standard deviation. ***p < 0.0002, ****p < 0.0001 (Two way ANOVA, Sidak’s multiple comparisons).

3.5. Low doses of toxins and light cumulatively cause avoidance behavior

We next investigated how light enhanced avoidance of pyocyanin. Since pyocyanin is not a photosensitizer, we hypothesized that light and pyocyanin were acting via separate additive pathways. As each can generate ROS alone, their combination results in an additive effect to alter behavior. To test whether there was an additive effect occurring, we combined doses of antimycin A, light, and pyocyanin that, independently, have no effect on C. elegans behavior. Antimycin A is a mitochondrial toxin that prevents oxidation of ubiquinol to ubiquinone and interferes with the Q cycle, generating ROS [74,75]. C. elegans avoids bacterial lawns containing antimycin A in a manner that does not require light [7,8]. On their own, the concentrations of antimycin A, pyocyanin, or light treatment did not alter behavior (Fig. 6, Fig. S4), but we hypothesized that the combination of all three agents would have an additive effect to cause light-dependent avoidance. When plates containing pyocyanin and antimycin A were placed in the dark, there was no avoidance of the combined toxins. However, when exposed to light, the combination of all three compounds led to an increase in avoidance of the lawn (Fig. 6). These data demonstrate that an additive effect from different ROS sources can alter C. elegans behavior.

Fig. 6.

Low concentrations of antimycin A, pyocyanin, and light cumulatively cause avoidance behavior. Worms were placed on plates with combinations of low doses antimycin A (20 μM), pyocyanin (0.25 mM), and light (white circle, 2.46 μW/mm²). Black circle indicates no light. Worms on and off the bacterial lawn were counted after 2 h to determine the fraction avoiding. Each dot represents an independent trial of 25–35 day 1 adult worms. Bars are represented as means ± standard deviation. ****p < 0.0001 (Two way ANOVA, Sidak’s multiple comparisons).

3.6. LITE-1 is required for ROS-mediated avoidance behavior

Light-mediated avoidance of pyocyanin was previously shown to require lite-1 [25] (Fig. 7A), and our results demonstrate a ROS-dependent mechanism. Additionally, structural models of LITE-1 revealed a peroxiredoxin-2 (PRDX-2) binding site, which suggests that LITE-1 can be modified by H₂O₂ [29]. Therefore, we tested the relationship between ROS and LITE-1 activation independent of blue light. Using green-light photosensitization of rose bengal, we found that avoidance behavior required lite-1 (Fig. 7B). These results indicate a non-canonical role for LITE-1 in responding to ROS that is not related to blue light.

Fig. 7.

Loss of lite-1 decreases light-mediated avoidance of pyocyanin and rose bengal. Day 1 adult N2 or lite-1 worms were placed on plates with lawns of OP50 E. coli mixed with 15.6 mM pyocyanin (A) or 6.4 μM rose bengal (B) in the center of the plate. Plates were placed either in the dark or exposed to light (2.46 μW/mm²) for 2 h. Worms on and off the plate were counted to determine the fraction avoiding. Each dot represents an independent trial of 25–35 day 1 adult worms. Bars are represented as means ± standard deviation. ***p < 0.0002 (Two way ANOVA).

LITE-1 activity is regulated by H₂O₂, with H₂O₂ inactivating LITE-1 to reset its activity (Fig. 8A) [28]. Many proteins regulated by H₂O₂ have cysteine residues that undergo reversible oxidation, however, not all cysteine residues are reactive [57,76]. Given the putative location of C44, exposed to the cytosol, and the characteristics of a reactive cysteine [77], we hypothesized that C44 in LITE-1 may be modified by ROS signaling. To test if H₂O₂ modification of C44 was necessary and sufficient for LITE-1 signaling, we mutated C44 to serine (C44S) or aspartic acid (C44D) to prevent or mimic H₂O₂ modification, respectively [57]. Since LITE-1 mediates avoidance of blue light, we tested the speed of C. elegans when exposed to blue light, finding that C. elegans increase speed after exposure to blue light (Fig. 8B). Both the lite-1 null [13] and C44D mutant did not change speed in response to blue light, while the C44S mutation responded similarly to wild-type. These results suggested that oxidation had no effect on LITE-1 activation but it may be required for inactivation.

Fig. 8.

LITE-1 cysteine 44 (C44) is required for regulation of LITE-1 activity. (A) Schematic illustration of LITE-1 activity. Blue light both activates LITE-1 and generates ROS. ROS inactivates LITE-1 via oxidation. After exposure to antioxidants, LITE-1 is reduced and can be activated by blue light again. (B–C) Mutation of LITE-1 alters response to blue light. Movement speed was taken prior to blue light exposure (black circle, 0.836 μW/mm²) and after blue light exposure (blue circle, 218 μW/mm²). In the recovery assay (C), speed was determined 30 s after the end of the blue light exposure (second black circle, 0.836 μW/mm²). Worms pre-treated with EUK-134 were grown on 100 μM EUK-134 for 24 h. Each dot represents an independent trial of 1 day 1 adult worm. ****p < 0.0001 (Two way ANOVA, Sidak’s multiple comparisons) (D–E) Day 1 adult worms were placed on plates with lawns of OP50 E. coli mixed with 15.6 mM pyocyanin (D) or 6.4 μM rose bengal (E) in the center of the plate. Plates were placed in the dark (black circle) or exposed to light (white circle, 2.46 μW/mm²) for 2 h. Worms on and off the plate were counted to determine the fraction avoiding. Each dot represents an independent trial of 25–35 day 1 adult worms. Bars are represented as means ± standard deviation. ***p < 0.0002 (Two way ANOVA, Sidak’s multiple comparisons).

We next tested if C44 oxidation would alter recovery from blue light exposure. Wild type C. elegans returns to baseline speed after cessation of blue light exposure in 30 s (Fig. 8C). We hypothesized that the return to normal foraging behavior after blue light exposure required the oxidation of C44, which would be prevented in the C44S mutant. A modified body bend assay was used to determine the recovery rate following blue light exposure. While the wild type worms returned to their baseline speed after 30 s, the C44S mutant remained at an elevated speed after 30 s (Fig. 8C), suggesting that oxidation of C44 was required for LITE-1 inactivation. To determine if ROS is required for LITE-1 inactivation, wild type worms were treated with EUK-134 and subjected to blue light. Similar to the C44S mutation, EUK-134 prevented the recovery to baseline speed after blue light exposure. EUK-134 had no effect on the response to blue light or recovery to baseline in the C44S and C44D mutants (Fig. S5). Together, these results showed that ROS modification is necessary for the regulation of LITE-1 activity.

Our results suggest that LITE-1 is a central mediator in the avoidance response elicited by ROS and blue light. We next tested whether the C44S and C44D mutations would affect avoidance behavior of pyocyanin or rose bengal. The C44S mutant mimicked wild-type worms in that it avoided pyocyanin in a light-dependent manner, while the C44D mutant, similar to lite-1, did not respond to pyocyanin (Fig. 8D). To separate the avoidance behavior from the color blue, we repeated the avoidance assay with rose bengal. Avoidance of rose bengal in the presence of light in the C44S mutant was similar to wild type, while the lite-1 null and C44D mutant avoided the lawn at a lower frequency (Fig. 8E). Taken together, these results highlight the importance of LITE-1 for ROS sensing and avoidance behavior in C. elegans.

4. Discussion

Recent work has shown that mitochondrial ROS alters C. elegans behavior [32,57,[78], [79], [80]]. Collectively, these reports have motivated questions regarding selectivity in ROS signaling cascades – in terms of origin and biological targets. Our results here describe a behavioral output that integrates multiple ROS sources, including photosensitizers, mitochondrial toxins, and antioxidants, to elicit avoidance behavior, and show that ROS is both necessary and sufficient to drive changes in animal behavior. Accumulation of ROS from multiple, sub-threshold sources can have additive (or perhaps even synergistic) effects, resulting in a synthetic phenotype that is readily reversible. Moreover, we elucidated a critical role of LITE-1 in mediating avoidance of ROS independent of blue light through selective redox modification of C44.

The results of this study expand on the role of ROS as a signaling molecule. H₂O₂ generated by light exposure leads to the cessation of feeding, which requires LITE-1 [12]. Our results expand on the role of LITE-1 at the junction of light and ROS, activated by light, and inactivated by H₂O₂, in a cycle that allows C. elegans to avoid light. H₂O₂ can also promote sensory behavior at low doses, increasing the function of sensory neurons [78]. The combination of all of these findings shows a role for H₂O₂ in fine-tuning C. elegans survival behavior.

The avoidance behavior detailed in these studies is driven by ROS that is generated internally. While this could be observed with the photosensitizers using light and direct measurement of ROS generation, the connection between light and pyocyanin was more difficult to understand. Measurement of superoxide generated by pyocyanin using DHE was not achievable due to its redox cycling capabilities. However, combining very low concentrations of pyocyanin and antimycin A with a low dose of light was able to alter behavior (Fig. 6). This experiment shows the effects that different ROS sources can have in combination, and in this case causes C. elegans to avoid the bacterial lawn. The inability of ROS generated by blue light to potentiate avoidance of pyocyanin (Fig. 2A) can likely be explained by the decreased number of photons of blue light being transmitted through the filter. Additionally, the endogenous photosensitizers in C. elegans are likely to absorb some light outside of the blue spectrum, further decreasing the amount of ROS that would be generated when only exposed to the blue light.

Our work shows that internal ROS is necessary and sufficient for altering behavior. Blocking ROS generation by pre-treating worms with antioxidants prevented or diminished avoidance behavior, while pre-treating with rose bengal and exposing to light caused immediate changes in behavior. These results further the idea of the duality of ROS. Large amounts of ROS generated by pyocyanin or photosensitizers will be toxic [81], while lower amounts are important for signaling. In this case, the signaling causes changes in C. elegans behavior that leads to avoidance of the toxin and injury. While ROS generation by methylene blue and rose bengal strongly correlated with avoidance behavior, this was not the case with neutral red. Neutral red can cause avoidance behavior while generating ROS at levels lower than rose bengal and methylene blue (Fig. 3). We speculated that the photosensitizers localize to a different location or C. elegans have different uptake/efflux rates for the compounds. C. elegans has a range of transporters that could be involved [[82], [83], [84]]. LITE-1 is hypothesized to be an ion channel but there is currently no evidence that that it can transport molecules such as the tested photosensitizers [29].

The molecular pathways that cause these behavioral changes in C. elegans are not well elucidated. Behavioral changes can be initiated by blue light itself [13,25]. Blue light can activate endogenous photosensitizers [27], and is detrimental to the health of a wide range of organisms [11,85,86]. C. elegans has multiple light-sensitive proteins, including LITE-1 and GUR-3, which are expressed in neurons [[12], [13], [14], [15],80]. Our results confirm the importance of LITE-1 for light-dependent avoidance of pyocyanin, as well as rose bengal (Fig. 7). Using a combination of rose bengal and green light, we showed that LITE-1 plays a role in the avoidance of ROS. Modeling of LITE-1 activity showed a photo-oxidation relay whereby H₂O₂ oxidizes cysteine residues, resulting in a conformational change that allows an influx of ions [29]. Confirming our results on the importance of C44, the loss of C44 was found to decrease LITE-1 activity by ∼50%. LITE-1 activity had previously been tied to blue and UV light [13,14]. These conditions are often associated with the generation of O₂^•- and H₂O₂, making it difficult to isolate the role LITE-1 plays in C. elegans responding to blue light and ROS. Using only green light, ROS was generated by rose bengal. As neither C. elegans nor LITE-1 respond to green light (Fig. 4C), we elucidated a role for LITE-1 in responding to ROS independent of light. ROS plays an important role in regulating the activity of LITE-1, and the results of this study further the understanding of its role through modification of LITE-1. LITE-1 shifts between an oxidized and reduced state based on environmental conditions, causing C. elegans to increase speed in attempt to leave the location. Recent work has located a possible PRDX-2 binding site on LITE-1 [29]. Since not all cysteine residues are readily oxidized by H₂O₂, the localization of PRDX-2 to LITE-1 suggests a redox relay model to oxidize/reduce C44 (Fig. 8A).

In conclusion, we have determined the mechanism by which light can guide behavior in C. elegans. The avoidance behavior observed does not require perception of specific colors, but the generation of ROS internally. This work expands on the role ROS plays as a signaling molecule, driving behavioral changes in C. elegans to avoid danger and seek favorable environments. The amount of ROS being generated can have an impact on the frequency of the behavioral change, with more ROS leading to a higher incidence of avoidance behavior. LITE-1 plays a central role in avoidance behavior, is required for photosensation, and is involved in sensing internally generated ROS. The ROS generated under conditions that activate LITE-1 (such as blue light exposure) regulates its activity, inactivating LITE-1 through C44. Our results highlight the importance of ROS not only in modifying behavior, but also as a signaling molecule in sensory perception.

Declaration of competing interest

None.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.