Moderate heat stress-induced sterility is due to motility defects and reduced mating drive in Caenorhabditis elegans males

Sexual reproduction is highly temperature sensitive. Diverse species including Asian rice, nematodes, fruit flies, cattle and humans exhibit a dramatic loss of fertility at sublethal temperatures, which are only 1–2°C higher than temperatures where 100% of the population is fertile (David et al., 2005; Lam and Miron, 1996; Morrell, 2020; Petrella, 2014; Prasad et al., 2006). While conditions of moderate heat stress negatively impact female fertility in some species, they impact male fertility more broadly (David et al., 2005; Iossa, 2019; Petrella, 2014; Sales et al., 2018; Zinn et al., 2010). In ectothermic organisms that cannot regulate their body temperature, such as the nematode Caenorhabditis elegans, temperature effects on fertility are even more pronounced (Kingsolver et al., 2013; Walsh et al., 2019). As climate change-associated temperature variability increases, it is crucial to understand how rising temperatures affect fertility.

Prior research in many species has focused primarily on the negative impact of moderate heat stress on germline-associated processes in males, such as sperm production and quality (Kirk Green et al., 2019; Llamas-Luceño et al., 2020; Rao et al., 2016). However, in C. elegans males, moderate heat stress has only a minor impact on germline-associated processes, which cannot account for the near-total male sterility observed (Nett et al., 2019; Petrella, 2014; Sepulveda and Petrella, 2021). In C. elegans, male sterility under moderate heat stress is primarily attributed to somatic-associated changes in mating organ anatomy and mating behavior, rather than sperm production and quality (Nett et al., 2019). Moderate heat stress-induced changes in male mating behavior have also been observed in Drosophila males (Dolgin et al., 2006; Fasolo and Krebs, 2004; Jørgensen et al., 2006), suggesting that behavioral changes may be widespread and underappreciated. To accurately predict the effects of climate change on species’ survival, the effects of heat stress on behaviors that contribute to species' fertility need to be accounted for.

Male C. elegans behavior is generally dictated by two drives: feeding and mating; and balancing these drives is crucial for successful reproduction (Barr et al., 2018; Barrios, 2014; Muirhead and Srinivasan, 2020). Each drive elicits specific behaviors that are well characterized. The feeding drive predominates when males are hungry, eliciting a series of behaviors that result in food intake necessary for survival. For the feeding drive to succeed, males must: (1) sense they are hungry, (2) initiate the search for food, (3) maintain food-searching behavior until food is located, and (4) ingest food. Upon satiation under optimal conditions, the mating drive predominates, eliciting a series of behaviors that result in successful reproduction. For the mating drive to succeed, males must: (1) initiate mate-searching behavior if mates are not immediately detectable, (2) pause mate-searching behavior when mates are sensed by either touch or chemical signals, (3) initiate copulatory behaviors once a mate is found, and (4) complete copulation and transfer sperm. Males must also have the physical endurance to perform these behaviors to reproduce successfully. Our previous work has shown that under moderate heat stress, males fail to initiate copulatory behaviors when mates are present and therefore do not produce progeny (Petrella, 2014; Nett et al., 2019). However, the effects of moderate heat stress on the balance of the feeding and mating drives, and how moderate heat stress may impact other reproductive behaviors were unknown.

Here, we asked which components of C. elegans male mating behavior are negatively impacted by heat stress. We found that moderate heat stress (27°C) resulted in motility defects that likely limited male mating. We also report that moderate heat stress leaves feeding drive largely intact but reduces mating drive in C. elegans males by interrupting mate-searching behavior and male copulatory behavior. The cumulative result of these changes is that very few heat-stressed males are interested in mating, and/or are able to mate successfully, resulting in poor reproductive outcomes.

Caenorhabditis elegans worms were maintained using standard methods at 20°C on nematode growth media (NGM) plates (Brenner, 1974) spotted with AMA1004 Escherichia coli (Casadaban et al., 1983). Wild strain males were generated by heat shocking L4 hermaphrodites at 30°C for 5 h followed by recovery at 20°C and selection of male progeny. Male populations were subsequently maintained by crossing males and hermaphrodites of the same genotype at 20°C. CX32 [odr-10(ky32) X], CX3410 [odr-10(ky225) X], DR476 [daf-22(m130) II], JU1171, LKC34, N2 and MT7929 [unc-13(e51)] were obtained from the Caenorhabditis Genetics Center (CGC, University of Minnesota, Minneapolis, MN, USA) which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). UR460 [him-5(e1490); kyIs53 (Podr-10::odr-10::gfp)] was obtained from Douglas Portman PhD (University of Rochester Medical Center, Rochester, NY, USA). The wild strains used in this study were isolated from the following localities: JU1171 (Concepción, Chile), LKC34 (Madagascar) and N2 (Bristol, UK).

Two temperature treatments were used in this study: (1) continuous exposure to 20°C, and (2) continuous exposure to 27°C. The 20°C experiments were performed on male worms maintained continuously at 20°C. The 27°C experiments were performed on male worms upshifted to 27°C at the L1 stage and maintained continuously at 27°C until assayed. For mate search experiments, unc-13(e51) hermaphrodites were maintained at 20°C, and only exposed to elevated temperature at the young adult stage for the duration of experiments at 27°C. For both temperature treatments, male celibacy was ensured by isolating males from hermaphrodites at the L4 stage and allowing them to mature to adulthood overnight (18–24 h). Males were isolated at a population density of ≤20 worms per plate.

Pheromone response assays were performed similar to Sammut et al. (2015). NGM plates were poured 2 days before the experiment. One day before the experiment, plates were spotted with 50 µl AMA1004 E. coli grown to OD₆₀₀=1.0. One day before the experiment, males and strain-matched (same genotype) or DR476 (ascaroside defective, N2 genetic background) hermaphrodites were isolated from each other at the L4 stage and allowed to mature to adulthood (18–24 h). On the day of the experiment, adult hermaphrodites were rinsed twice with M9 buffer dispensed in a three-well watch glass to remove residual E. coli. These adult hermaphrodites were used to make ‘hermaphrodite-conditioned buffer’ by placing them in the lid of a 1.7 ml microfuge tube with M9 buffer at a concentration of 1 hermaphrodite µl⁻¹ of M9 for 5 h in a humidity chamber at 20°C. After the conditioning period, the buffer was separated from the associated hermaphrodites and stored at 4°C in a new 1.7 ml microfuge tube until use within 2 h. Test plates were prepared by dispensing 1 µl plain M9 buffer and 1 µl conditioned M9 buffer across from each other, 2 mm from the center of the bacterial food spot. Experiments were performed after the buffer spots had completely dried (3–5 min). Five to eight adult test males were rinsed twice with plain M9 buffer dispensed in a three-well watch glass to remove residual E. coli and placed on the test plate equidistant from the control and the conditioned spots, 3 mm away from the center. Males were recorded at 1× magnification using a Nikon SMZ 1500 stereo microscope equipped with an HR Plan APO 1× objective lens (Nikon Instruments Inc., Melville, NY, USA) for 20 min using a Q imaging Exi Blue camera (Teledyne Photometrics, Tucson, AZ, USA) and Q Capture Pro 7 software (Teledyne Photometrics). Three to five test plates per strain and temperature condition were imaged for each replicate, with a single 20 min video captured per plate. For experiments with strain-matched wild strain males and hermaphrodites, 3–5 biological replicates were performed for each strain and temperature condition, with n=15–91 males scored in total. For experiments with N2 males and daf-22(m130) hermaphrodites, 2–3 biological replicates were performed for each temperature condition, with n=21–26 males scored in total.

Videos were subsequently analyzed using FIJI software (Schindelin et al., 2012). Each time a male entered either the control or conditioned spot, the amount of time the male spent in the spot was scored. Each entrance to a spot was considered an independent event if, after leaving, the male worm had moved at least 2 mm away from the spot before re-entering. If males encountered each other in either spot and engaged in mating behavior for at least 5 s, the corresponding entrance events were discarded from the analysis. Data were analyzed with the Mann–Whitney U-test using the Bonferroni–Dunn method to correct for multiple comparisons in Prism 9 (GraphPad, San Diego, CA, USA).

Leaving assays were performed as in Lipton et al. (2004). NGM plates were prepared 2 days before the experiment. One day before the experiment, plates were spotted with 18 µl AMA1004 E. coli grown to OD₆₀₀=1.0. Plates were used the day after spotting. Two days before the experiment, strain-matched (same genotype) hermaphrodites were isolated from males at the L4 stage and allowed to mature to adulthood. One day before experiments, the adult hermaphrodites were fixed in fresh 4% formaldehyde in PBS solution overnight at 4°C (Barrios et al., 2008). On the day of experiments, formaldehyde-fixed (FA) hermaphrodites were washed 3 times in PBS buffer. Test plates were prepared by placing two FA hermaphrodites equidistant from each other, 2 mm from the center of the food spot. Control plates were prepared by disturbing the food lawn at the location where each hermaphrodite would be placed on an experimental plate. One day before the experiment, males were isolated from hermaphrodites at the L4 stage and allowed to mature to adulthood. On the day of the experiment, a single male was rinsed once in M9 buffer and was placed at the center of the food spot of an experimental or control plate. Plates were examined at intervals (2, 4, 6, 8, 20, 22 and 24 h) to check for ‘leavers’. A male was considered a leaver and left censored if the male or worm tracks were observed on the agar within 2.5 mm of the edge of the plate. For experiments testing wild males, 10 test plates and 10 control plates per strain were set up for each experiment and four biological replicates were performed for a total n=36–47. For experiments testing N2, odr-10(ky32) and odr-10(ky225) males, 3–4 biological replicates were performed for a total n=27–43. We used the previously published model of male leaving behavior (Lipton et al., 2004), which employs a single exponential decay function which remains constant over the course of an experiment: N(t)/N(0)=exp(−λt), where N(t) refers to the number of males retained at a given time and the hazard value (λ) gives an estimate for the probability of leaving (P_L). P_L values were calculated by fitting the data using the R survival package (https://CRAN.R-project.org/package=survival) with an exponential parametric survival model using maximum likelihood (Hilbert and Kim, 2017). For experiments using wild strains or odr-10 mutants, the average P_L was calculated for each replicate and the data were compared with a two-way ANOVA using Tukey's multiple comparison test in Prism 9 (GraphPad).

One day before the experiment, NGM plates were spotted with 50 µl AMA1004 E. coli grown to OD₆₀₀=1.0. The bacteria were spread over the NGM plate using a sterile glass spreader to 2 mm from the edge of the plate. The day before the experiment, males and hermaphrodites were isolated from each other at the L4 stage and allowed to mature to adulthood. The day of the experiment, 10 males or hermaphrodites were washed once in M9 buffer and then placed at the center of the food spot. Subsequently, the males or hermaphrodites were allowed to move about the plate and acclimate for 2 h. After the acclimatization period, the plates were imaged at 1× magnification using a Nikon SMZ 1500 stereo microscope equipped with an HR Plan APO 1× objective lens (Nikon Instruments Inc., Melville, NY, USA) capturing 30 s videos using a Q imaging Exi Blue camera (Teledyne Photometrics) and Q Capture Pro 7 software (Teledyne Photometrics). Each plate was imaged at least once but as many as 6 times to capture each worm on the plate. Care was taken to film each worm once, but occasional reimaging of a worm may have occurred. To assess absolute movement velocity and convert pixels to millimeters, a known millimeter scale was imaged and recorded. Videos were processed using FIJI; and maximum movement velocities for each worm were calculated using the WrmTrck plugin for FIJI (Nussbaum-Krammer et al., 2015; Schindelin et al., 2012). Detailed instructions for utilizing WrmTrck are available at the WrmTrck website (https://www.phage.dk/plugins/wrmtrck.html). When assessing movement velocity, worms that did not move at all during the 30 s time frame or moved for less than 5 s (frames) during the 30 s period were excluded from the analysis. To assess the proportion of worms that were motile during the 30 s test period, worms that did not move at all or moved less than 1 mm from their initial position were considered not moving. Two to five plates of worms were assessed for each strain, sex and temperature condition per replicate, with four replicates being performed for a total n=52–153. Maximum velocity was analyzed with the Mann–Whitney U-test using the Bonferroni–Dunn method to correct for multiple comparisons in Prism 9 (GraphPad); and the number of worms in the test population that moved versus those that did not move was analyzed using a Fisher's exact test, also in Prism 9.

Liquid thrashing assays were performed in M9 buffer as in Chatterjee et al. (2013). A single male or hermaphrodite was allowed to crawl on the agar surface of a NGM plate to remove excess bacteria and then transferred to M9 buffer in a three well watch glass. The number of full body bends (thrashes) was recorded for 60 s. Twenty individuals were assessed for each biological replicate and 3–4 biological replicates were performed for each strain, sex and temperature condition for a total n=60–80. Data were analyzed with the Mann–Whitney U-test using the Bonferroni–Dunn method to correct for multiple comparisons in Prism 9 (GraphPad).

Food search assays were executed similar to Ryan et al. (2014). One day before the experiment, NGM plates were spotted with 7 µl AMA1004 E. coli grown to OD₆₀₀=2.5. One day before the experiment, males were isolated from hermaphrodites at the L4 stage and allowed to mature to adulthood. The day of the experiment, a single male was rinsed once in M9 buffer and was placed 3.2 cm away from the food spot. Plates were examined every 15 min for 2 h to ascertain whether a male had reached the food spot. Ten test plates per strain and temperature condition were set up for each experiment and five biological replicates were performed for a total n=45–70. Survival curves were generated using the Kaplan–Meier method and analyzed using the Log-rank (Mantel–Cox) test in Prism 9 (GraphPad).

These experiments were performed similar to the food search assays described above with one major change: on the day of the experiment, a 20 μl droplet of 1 μm diameter Polybead^® Carboxylate Red Dyed Microspheres (Polysciences Inc., Warrington, PA, USA; catalog number 19119-15) in 1× M9 buffer at a concentration of 2.275×10¹⁰ particles ml⁻¹ (a 1:1 dilution of polystyrene microspheres to buffer) was placed at the site of male worm release (3.2 cm from the food spot) and allowed to dry. Males were released at the center of the beads and the assay was performed. Four to five biological replicates were performed for each strain and temperature treatment (20 and 27°C) for a total n=54–94. Survival curves were generated using the Kaplan–Meier method and analyzed using the Log-rank (Mantel–Cox) test in Prism 9 (GraphPad).

Mate search assays were performed as in Fagan et al. (2018). One day before the experiment, NGM plates were spotted with 10 µl AMA1004 E. coli grown to OD₆₀₀=1.0 which was spread into a 1 cm×2.5 cm rectangular food lawn using a sterile glass spreader. One day before the experiment, males were isolated from strain-matched (same genotype) hermaphrodites at the L4 stage and allowed to mature to adulthood. For experiments using FA hermaphrodites, hermaphrodites were isolated 2 days before experiments, and 1 day before experiments the young adult hermaphrodites were fixed in fresh 4% formaldehyde in PBS solution overnight at 4°C (Barrios et al., 2008). On the day of the experiment, three FA hermaphrodites were rinsed 3 times in PBS buffer and placed equidistant from each other at one end of the food lawn. For experiments with live unc-13(e51) hermaphrodites, 1 day before the experiment, L4 unc-13(e51) hermaphrodites were isolated and allowed to mature to adulthood. On the day of the experiment, three unc-13(e51) hermaphrodites were rinsed once in M9 buffer, placed equidistant from each other at one end of the food lawn and allowed to condition the plate for 2 h. For both experiments, a single male was rinsed once in M9 buffer and placed on the opposite end of the food lawn. Plates were checked every 5 min for 1 h to determine whether males had reached the vicinity of the hermaphrodites and commenced mating behaviors. Ten plates per strain and temperature condition were set up for each experiment. Six to seven biological replicates were performed, for a total n=60–73. Survival curves were generated using the Kaplan–Meier method and analyzed using the Log-rank (Mantel–Cox) test in Prism 9 (GraphPad).

Two days before the experiments, strain-matched (same genotype) L4 hermaphrodites were isolated and allowed to mature to adulthood. One day before the experiment, test plates and control plates were spotted with two 18 µl spots of AMA1004 E. coli grown to OD₆₀₀=1.0, 1.5 cm apart from each other flanking the center of a NGM plate. One day before experiments, the adult hermaphrodites were fixed in fresh 4% formaldehyde in PBS solution overnight at 4°C (Barrios et al., 2008). The day of the experiment, FA hermaphrodites were rinsed 3 times in PBS buffer. Test plates were prepared by placing two FA hermaphrodites equidistant from each other near the center of one of the two food spots, 2 mm from the center. Control plates were prepared by disturbing the food lawn at the location where each hermaphrodite would be placed on an experimental plate. One day before the experiment, males were isolated from hermaphrodites at the L4 stage and allowed to mature to adulthood. The day of the experiment, a single male was rinsed once in M9 buffer and placed 1.5 cm from the center of an experimental plate or control plate. Ten test plates and 10 control plates per strain and temperature condition were set up for each experiment. Plates were examined at intervals (2, 4, 6, 8, 20, 22 and 24 h) to check for the location of the male. For experimental plates, locations were scored as: food spot without hermaphrodites, food spot with hermaphrodites, or agar. For control plates, locations were scored as: left food spot, right food spot, or agar. Each time a male changed location on the plate (e.g. from a food spot without hermaphrodites to a food spot with hermaphrodites), it was scored as a ‘switch’. Five to six biological replicates were performed, for n≥43 total. Data were analyzed with the Mann–Whitney U-test using the Bonferroni–Dunn method to correct for multiple comparisons in Prism 9 (GraphPad). Males on test plates were also scored for mating. Males engaging in scanning, turning or maintaining contact with the hermaphrodite were considered to be mating. Four to six biological replicates were performed, for a total n=37–60. Data were analyzed with the Mann–Whitney U-test using the Bonferroni–Dunn method to correct for multiple comparisons in Prism 9 (GraphPad).

Pharyngeal pumping assays were performed as in Raizen et al. (2012) and Hobson et al. (2006). Two days before the experiment, NGM plates with 100 μg ml⁻¹ streptomycin (NGM+Str) to select for AMA1004 E. coli were poured and allowed to cool and dry overnight (Casadaban and Cohen, 1980; Casadaban et al., 1983). One day before the experiment, NGM+Str plates were spotted with AMA1004 E. coli in LB with 100 μg ml⁻¹ streptomycin grown to OD₆₀₀=1.0 and allowed to dry overnight. Fresh NGM+Str plates were used for experiments and remaining plates were used to maintain worms for subsequent experiments. One day before the experiment, L4 males were selected and allowed to mature overnight into young adults (YA). On the day of the experiment, YA males and L4 males were cloned out individually onto NGM+Str plates and allowed to acclimate for 30 min. Worms were observed using a Nikon SMZ 1500 stereo microscope equipped with an HR Plan APO 2× objective lens (Nikon Instruments Inc.) for 30 s and the number of pumps was counted. Pumps were recorded 3 times for each worm, with each 30 s recording period separated by 10 s. The number of pumps was doubled to determine pumping rate (pumps min⁻¹). The three pumping rate measurements were averaged for each worm. Three to four biological replicates for each strain, temperature (20 and 27°C) and developmental stage (L4 and YA) were performed for a total n=21–40. Data were analyzed with the Mann–Whitney U-test using the Bonferroni–Dunn method to correct for multiple comparisons in Prism 9 (GraphPad). Worms were also scored for activity. A worm that pumped during at least one assessment period was considered active. A worm that failed to pump during all three assessment periods was considered inactive. Data were analyzed using Fisher's exact test in Prism 9 (GraphPad).

Visualization and quantification of ODR-10::GFP was performed as in Wexler et al. (2020). Two days before imaging, UR460 males and hermaphrodites were isolated from each other at the L4 stage and allowed to mature to adulthood (18–24 h). To image well-fed animals, young adults were left on standard NGM plates for an additional 24 h before imaging. To image starved animals, young adults were rinsed 3× in M9 buffer and transferred to fresh unspotted NGM plates with 200 µg ml⁻¹ ampicillin (to inhibit bacterial growth) for 24 h before imaging. Worms were rinsed 3 times in M9 buffer, mounted on 5% agarose pads, immobilized with 1 mmol l⁻¹ levamisole in M9 buffer (Manjarrez and Mailler, 2020) and imaged using a Nikon Eclipse TE2000-S inverted microscope equipped with a Plan Apo 60×/1.25 numerical aperture oil objective (Nikon Instruments Inc.). Images were captured using a Q imaging Exi Blue camera (Teledyne Photometrics) and Q Capture Pro 7 software (Teledyne Photometrics). Scoring of ODR-10::GFP fluorescence intensity was done using a scale of 0–2 where: 0=absent, 1=faint, 2=bright (Lawson et al., 2019; Ryan et al., 2014; Wexler et al., 2020). The scorer was blind to the sex, nutritional status and temperature treatment of the animal in each image. Two biological replicates were performed for each temperature and nutritional condition for a total n=33–58. Data were analyzed using multiple Chi-square tests in Prism 9 (GraphPad).

In the absence of hermaphrodites, males at 20°C will leave a bacterial food patch and engage in mate-searching behavior, the first behavior associated with the male mating drive (Barrios et al., 2008; Lipton et al., 2004). To sense the presence of hermaphrodites, C. elegans males use the touch response, mediated by sensory neurons in the tail fan, and the pheromone response, mediated by sensory neurons in the head (Barr and Sternberg, 1999; Koo et al., 2011; Liu and Sternberg, 1995). As heat-stressed males have abnormal tail anatomy (Nett et al., 2019), they potentially have an impaired touch response. To test whether heat-stressed males either had an impaired touch response or failed to initiate mate searching, we used the leaving assay. In the leaving assay, males were placed either on a plate with food and no hermaphrodites or on a plate with food and FA hermaphrodites (Fig. 1A) (Lipton et al., 2004). FA hermaphrodites do not excrete ascaroside pheromones, but can be still detected by touch. Under ideal temperature conditions and in the presence of food, if FA hermaphrodites are present, males will detect them through the touch response and be retained on the food patch. However, if hermaphrodites are absent, males will quickly initiate mate searching and leave the food patch. In our experiments we used three wild strains: N2, the canonical lab wild-type strain, and two more recently isolated wild strains, JU1171 and LCK34. As expected, we found that unstressed males of all three strains quickly left a food patch without FA hermaphrodites (−FA), but were retained on a food patch where FA hermaphrodites were present (+FA) (Fig. 1B–D). However, we found that heat-stressed males very rarely left a food patch within 24 h whether FA hermaphrodites were present or not (Fig. 1B–D). To directly compare male leaving behavior under different conditions, we calculated probabilities of leaving (P_L) from the leaving data (Lipton et al., 2004). Unstressed males exhibited high P_L values in the absence of FA hermaphrodites and low P_L values in their presence (Fig. 1E–G). However, heat-stressed males demonstrated very low P_L values in the presence or absence of FA hermaphrodites (Fig. 1E–G). There was a trend for heat-stressed males to stay on the food patch longer if FA hermaphrodites were present, but the difference was not statistically significant compared with when FA hermaphrodites were absent. Therefore, while the touch response may still be intact in heat-stressed males, the initiation of mate-searching behavior is significantly lessened compared with their unstressed counterparts.

Because heat-stressed males rarely leave a bacterial food patch even in the absence of hermaphrodites (Fig. 1), we hypothesized that they have an intact feeding drive. To test this possibility directly, we performed food search assays (Fig. 2A) (Ryan et al., 2014). In these assays, males were released at a distance from a food patch on a plate and the percentage of males able to reach the food within 2 h was scored. We found that heat-stressed males from all three wild strains were just as likely to reach a food patch within 2 h as their unstressed counterparts (Fig. 2B–D). We interpret these results to mean that moderate heat stress likely does not reduce feeding drive in C. elegans males.

One possible explanation for male retention on food is that these males have difficulty leaving a bacterial food patch due to motility defects. Therefore, we tested whether heat-stressed males experience changes in crawling on plates or thrashing in liquid. We found that heat-stressed males from all three strains had lower crawling velocities and were less likely to move on plates than their unstressed counterparts (Fig. 3A–F). In addition, heat-stressed males from all three strains had lower thrashing rates in buffer than their unstressed counterparts (Fig. 3G,H). Interestingly, we also found similar changes in motility in heat-stressed hermaphrodites (Fig. 3), suggesting that changes in motility due to moderate heat stress are not sex specific. Overall, there was a general decrease in motility in heat-stressed males that likely affects their ability to perform the behaviors required to mate successfully.

To more directly test whether decreased food leaving behavior is due to the males being trapped by physical properties of the food, we placed males on polystyrene microspheres, which mimic food viscosity without having any nutritional content or chemical signals. These experiments were identical to the food search assays (Fig. 2A) except polystyrene microspheres were dispensed at the site where males were released. In high concentrations, microspheres can be used to fully immobilize C. elegans (Manjarrez and Mailler, 2020). We used a concentration of microspheres that does not fully immobilize worms but creates increased movement difficulty, similar to a bacterial lawn. We found that most males, both unstressed and heat-stressed, escaped the polystyrene microspheres within the first 15 min of the experiment (Fig. 4A–C). Both LCK34 and JU1171 heat-stressed males were equally able to escape entrapment by microspheres as their non-stressed counterparts, while N2 heat-stressed males showed a small but significant delay in escaping entrapment by microspheres compared with their unstressed counterparts (Fig. 4A–C). However, in all three wild strains, heat-stressed males were less likely to reach the food source during the 2 h time frame of the experiment than their unstressed counterparts (Fig. 4D–F). These results suggest that the viscous medium restricted motility and likely more drastically tired heat-stressed males. Therefore, the reduced motility defects we observed in heat-stressed males (Fig. 3) likely contribute in part to the leaving defects we saw in these males (Fig. 1).

Given that heat-stressed males seem to be similarly attracted to food via chemical signals as unstressed males, we asked whether heat-stressed males had an intact response to hermaphrodite pheromones. To test this, we performed a pheromone response assay (Fig. 5A) (Sammut et al., 2015). In this assay, males are released equidistant from two spots: one that had been saturated with buffer conditioned by hermaphrodites of the same genotype (conditioned region) and one that had been saturated with buffer alone (control region). If males had an intact response to hermaphrodite pheromones, we predicted that they would stay in the hermaphrodite-conditioned region longer than the control region. As previously shown, unstressed N2 males did spend more time in the conditioned region than in the control region (Fig. 5D) (Sammut et al., 2015). Surprisingly, unstressed JU1171 and LKC34 males did not spend more time in the conditioned region than in the control region (Fig. 5B,C). We also found that heat-stressed males from all three wild strains spent more time in the conditioned region than in the control region (Fig. 5B–D). In addition, heat-stressed males of all three wild strains spent more time in the conditioned region than their unstressed counterparts (Fig. 5B–D). These results showed that C. elegans male pheromone response was both intact in heat-stressed males and prolonged despite the absence of a hermaphrodite. Notably, this change in pheromone response does not translate to increased reproductive success, as these males are still mostly sterile and produce very few viable progeny at 27°C (Petrella, 2014; Sepulveda and Petrella, 2021).

To determine whether ascarosides were the signaling molecules that elicited attraction in heat-stressed males, we also performed pheromone response assays using buffer conditioned by daf-22(m130) hermaphrodites, which are defective in ascaroside production (Fig. S1A) (Golden and Riddle, 1985). We found that both unstressed and heat-stressed N2 males spent an equal time in the region conditioned by daf-22(m130) hermaphrodites and the unconditioned region (Fig. S1B). These data suggest that the heat-stressed male attraction to hermaphrodite-conditioned buffer that we observed (Fig. 5B–D) was due to ascarosides excreted by the hermaphrodites.

Another component of mating drive that could be defective in heat-stressed males is the ability to search for and find mates that are available when food is also present (Barrios, 2014). To test this possibility, we performed mate search assays, where males are released on a food lawn at a distance from hermaphrodites on the same food lawn (Fig. 6A) (Fagan et al., 2018). We used two different types of hermaphrodites: FA strain-matched hermaphrodites that can be recognized by touch but do not produce any chemical signaling molecules; and living unc-13(e51) hermaphrodites in an N2 background, which have extremely limited motility, but still excrete chemical signaling molecules. We found that heat-stressed males of all three wild strains were significantly less likely to locate FA or Unc hermaphrodites than their unstressed counterparts (Fig. 6B–D). However, we also found that both unstressed and heat-stressed males of all three wild strains were significantly more likely to locate an Unc hermaphrodite than a FA hermaphrodite (Fig. 6B–D). Collectively, these results suggest that heat stress negatively impacts the components of male mating drive responsible for male mate searching on food. These data also support the findings of the pheromone response assays, affirming that male attraction to hermaphrodite chemical signals is intact in heat-stressed males and improves the likelihood of locating a mate.

The results of our previous experiments suggested that heat-stressed males maintain a food drive but have a decreased ability and/or motivation to leave food to search for mates. Specifically, heat-stressed males fail to initiate more widespread searching of hermaphrodites in the absence of hermaphrodite signals, especially when food is present. We wanted to test whether males could initiate mate searching in the absence of hermaphrodite chemical cues but the presence of food in an assay where they would need to crawl significantly shorter distances than in the leaving assay. A shorter travel distance and a smaller food lawn should limit the impact of motility defects on male mating behavior. For this, we designed a food/mate prioritization assay where males are tested under two conditions: +FA Herm plates, where males were place on a plate containing one food spot without FA hermaphrodites and one food spot with FA hermaphrodites, or −FA Herm plates, where males were placed on a plate in which neither food spot had FA hermaphrodites (Fig. 7A). Males were then scored over a 24 h period for which spot they occupied and for copulatory behaviors with FA hermaphrodites. If a male with an intact mating drive is placed on a +FA Herm plate, he will crawl to one of the two food spots. If he locates the spot with FA hermaphrodites first, he will stay there and copulate with them. If he locates the spot without the FA hermaphrodites, he will quickly switch to the spot with the FA hermaphrodites and then subsequently stay on that spot. In contrast, if a male with an intact mating drive is placed on a −FA Herm plate, he should continually switch between spots as he unsuccessfully searches for a mate. This is what we observed for unstressed males (Fig. 7B–D). On +FA Herm plates, ∼50% of the males switched spots at least once, while only ∼20% of males switched spots more than once, suggesting that most unstressed males locate and remain on the spot where FA hermaphrodites are present. However, on −FA Herm plates, ∼75–90% of males switched at least once and 50–80% of males switched more than once. These differing responses represent an active mating drive. Contrastingly, we found that heat-stressed males were significantly less likely to switch between food spots than their unstressed counterparts, with the highest amount of switching shown by JU1171 on −FA Herm plates, where ∼40% of males switched at least once (Fig. 7B–D). JU1171 was the only strain to show a significant difference between plates, and only for the number of males that switched plates at least once (Fig. 7B). These data further support the idea that heat-stressed males are less likely to initiate searching for mates in the presence of food, which may be compounded by their motility defects.

On +FA Herm plates, we also scored whether a male demonstrated active copulatory behavior with the FA hermaphrodites. We found that heat-stressed males were less likely to be copulating with FA hermaphrodites than their unstressed counterparts (Fig. 7E–G). This result is likely a combination of two factors: (1) heat-stressed males are less likely to switch and therefore they are less likely to find a hermaphrodite; and (2) heat-stressed males are less likely to initiate copulatory behavior than unstressed males (Nett et al., 2019). Overall, these two disruptions in male mating drive would significantly impede reproduction in heat-stressed males.

We observed that heat-stressed males can locate a food source like unstressed males (Fig. 2), and that they rarely leave a food source (Figs 1 and 7). We propose that these changes are due to moderate heat stress decreasing motility and reducing male mating drive. An alternative possibility is that heat stress elicits developmental defects causing adult males to respond to food like larval males. Caenorhabditis elegans males do not leave food to engage in mate-searching behavior until they are fully mature, with a functional germline and a complement of sperm (Kleemann et al., 2008; Lipton et al., 2004; Motola et al., 2006). To test whether heat-stressed adult males behave more like L4 males, we performed pharyngeal pumping assays with L4 stage and young adult males (Hobson et al., 2006; Raizen et al., 2012). Pharyngeal pumping rates have been shown to decrease between the L4 and young adult stages, which may reflect a switch from rapid food ingestion to balancing feeding and mating (Gruninger et al., 2006; Liu and Sternberg, 1995). If moderate heat stress causes adult males to exhibit juvenile feeding behavior, we would expect L4 and YA males to have similar pumping rates (pumps min⁻¹) at 27°C. We found that in all three strains, both heat-stressed and unstressed males showed a decrease in pumping rates as they matured from the L4 stage to adulthood (Fig. S2A–C). There were some differences within strains between heat-stressed and unstressed males, but the decrease in pumping between life stages was consistent between strains (Fig. S2A–C). These results suggest that heat-stressed adult males do not exhibit larval-like feeding behavior. Because we observed that heat-stressed males have decreased physical endurance, we also scored males at both 20 and 27°C for lack of pumping. LKC34 and N2 heat-stressed adult males were less likely to be actively pumping than conspecific L4 heat-stressed males (Fig. S2D–F), suggesting that pumping, like other movements dependent upon the neuromusculature, is disrupted during moderate heat stress.

Caenorhabditis elegans males ascertain their nutritional status in part via external chemical cues from a food source (Wexler et al., 2020). In well-fed unstressed males, food signals suppress the expression of the food-associated chemoreceptor odr-10 (Fig. 8A), promoting food-leaving behavior (Wexler et al., 2020). In starved unstressed males, odr-10 expression is high (Fig. 8A), prompting a return to feeding behavior (Wexler et al., 2020). We hypothesized that ODR-10 levels may be higher in heat-stressed males, promoting feeding behavior and retention on a food patch. To test this possibility, we examined males carrying the ODR-10::GFP transgene kyIs53 [Podr-10::odr-10::gfp] (Fig. 8B) (Sengupta et al., 1996). We found that well-fed heat-stressed males had higher levels of ODR-10::GFP than their well-fed unstressed counterparts (Fig. 8C). In fact, well-fed heat-stressed males had levels of ODR-10::GFP equivalent to those of starved unstressed males (Fig. 8C). These data suggest that heat stress increases the amount of ODR-10 in male worms, which may subsequently promote male feeding behavior and retention on a food patch.

To determine whether the increase in ODR-10 we observed in heat-stressed males is directly responsible for the food-leaving defects we observed, we performed leaving assays with two loss of function odr-10 mutants in an N2 background: CX32 [odr-10(ky32) X] and CX3410 [odr-10(ky225) X] (Sengupta et al., 1996). If increased expression of odr-10 in heat-stressed males is primarily what drives male retention on food even in the absence of hermaphrodites, we would expect to see heat-stressed odr-10 mutant males leave food in the absence of hermaphrodites. However, heat-stressed odr-10(ky32) and odr-10(ky225) males very rarely left a food source whether FA hermaphrodites were present or not, and showed low P_L values, similar to what was seen with heat-stressed N2 males (Fig. S3A–C). These findings are complicated by the fact that loss of odr-10 does not have a strong effect on leaving behavior at ideal temperatures. At 20°C, N2, odr-10(ky32) and odr-10(ky225) males left a food patch when FA hermaphrodites were absent (−FA), but were retained on a food patch where FA hermaphrodites were present (+FA) (Fig. S3A–C). As previously shown by others using a similar assay at 20°C (Ryan et al., 2014), in the absence of FA hermaphrodites, odr-10(ky225) males were equally likely to leave food as N2 males (Fig. S3D,E). Contrastingly, at 20°C in the absence of FA hermaphrodites, odr-10(ky32) males were less likely to leave food than N2 males (Fig. S3D,E). We are the first to perform food-leaving experiments with odr-10(ky32) males, and the cause of the discrepancy between mutant phenotypes is unknown, although odr-10(ky32) is a missense mutant while odr-10(ky225) is a deletion mutant (Sengupta et al., 1996). However, this was surprising as both odr-10(ky32) males and hermaphrodites perform similarly to odr-10(ky225) worms in food chemotaxis experiments (Ryan et al., 2014; Sengupta et al., 1996). These results suggest that increased odr-10 expression alone does not account for the reduction in food-leaving behavior we saw in males at elevated temperature.

Here, we endeavored to identify the physical and behavioral changes that explain why C. elegans males exposed to moderate heat stress (27°C) have little to no success in mating. A male physically unable to mate or uninterested in mating with a hermaphrodite stands virtually no chance at reproducing successfully, regardless of the quality of his sexual anatomy or germline-associated processes such as sperm production. We report that heat stress profoundly changes multiple components of male mating drive including: initiation of searching behavior, possible difficulty detecting hermaphrodites by touch, and initiating copulation. In addition, these negative changes to mating behavior are exacerbated by a heat stress-induced decline in male motility. However, heat-stressed males have an intact, and even potentially increased, ability to detect and respond to hermaphrodite secreted signals. In comparison, despite motility defects, heat-stressed males have an intact feeding drive, and possess sufficient motivation to locate and exploit food resources. This combination of behavioral and motility defects leads to heat-stressed males failing to mate in the presence of food, due in part to unbalanced mating and feeding drives. The outcome of these changes is that males are virtually sterile at 27°C.

To reproduce successfully, C. elegans males must locomote. We observed that heat-stressed males displayed reduced thrashing in liquid, were less likely to move or show pharyngeal pumping activity, and had reduced movement velocity on plates. Impaired motility likely contributed to decreased male leaving or switching by heat-stressed males in pheromone response and food/mate prioritization assays, respectively. In addition, when placed on polystyrene microspheres, heat-stressed males were less likely to find a food source, suggesting that additional restriction of movement was more detrimental to heat-stressed males. Because mate searching, copulation and sperm transfer are directly dependent on a male's ability to move, any reductions in motility could result in a reduction in successful mating attempts. This may be especially relevant in the natural environment of C. elegans. Adult, reproductively active C. elegans isolated from the wild are located in rotting fruits, which present an environment variable in terms of physical properties such as viscosity (Schulenburg and Félix, 2017). Similar defects in movement that contribute to infertility at elevated temperature have been observed in other taxonomic groups including fish and insects (Araujo et al., 2018; Rank et al., 2007; Wilson, 2005). Thus, decreased locomotion and movement defects may be widespread consequences of moderate heat stress that are likely to affect population fertility and species survival during global climate change.

While the motility defects seen in heat-stressed males contribute to their decreased fertility, our data indicate that defects in male mating drive also contribute. In specific circumstances heat-stressed males could be motivated to move in ways indistinguishable from those of their non-stressed counterparts. For example, heat-stressed males were able to locate a food spot and move through an unconditioned spot on a plate during pheromone response assays at rates similar to those of unstressed males. In addition, when motivated by hermaphrodite chemical cues, heat-stressed males were able to increase their motility. Thus, when motivated by food (or other chemical cues), males are able to overcome their motility defects and perform searching behaviors. This indicates that the feeding drive is intact in heat-stressed males. However, despite being able to overcome motility defects when encountering hermaphrodite chemical cues, heat-stressed males fail to perform most of the other behaviors indicative of an intact mating drive, which include initiation of mate searching (i.e. leaving food when hermaphrodites are absent) and initiation of copulatory behaviors when hermaphrodites are present. Thus, the male mating drive is impaired in heat-stressed males beyond just what can be explained by their motility defects.

Caenorhabditis elegans male exploratory behavior is regulated by both internal states and sensory cues (Barrios, 2014), any or all of which could be temperature sensitive. We looked at two such states shown to regulate mating behavior: developmental stage and nutritional status. Compared with adults, larval males cannot mate, do not engage in mate-searching behavior, and ingest food faster. As heat-stressed males produce ample functional sperm (Nett et al., 2019), and show feeding behavior in line with their developmental stage, we conclude that the changes in mating behavior we observed are not the result of a delay in maturity similar to a larval-like juvenile state. At ideal temperatures, well-fed males will leave food to initiate mate searching until they perceive themselves to be starving. In food-deprived male worms, a lack of chemical signals from food results in the activation of the food signal chemoreceptor odr-10 in AWA neurons through an insulin-dependent signaling cascade (Wexler et al., 2020). Increased levels of ODR-10 are associated with an increase in feeding behavior in starved males; and upon feeding, ODR-10 levels return to baseline (Ryan et al., 2014; Wexler et al., 2020). Well-fed males at 27°C showed levels of ODR-10 expression equivalent to those of starved males at 20°C, suggesting that heat stress may increase attraction to food and reduce mate searching. However, heat-stressed odr-10(-) males did not show increased leaving behavior in our assays, suggesting that a reduction of odr-10 expression is not sufficient to drive leaving behavior. Thus, it is still unclear to what extent increased expression of ODR-10 is playing a role in driving decreased mate-searching behaviors. Additional studies are needed to fully elucidate the molecular changes that lead to modified behavior in heat-stressed males.

One surprising finding of our study is that heat-stressed males remain in proximity to hermaphrodite chemical signals for longer than their unstressed counterparts, even though there are no hermaphrodites physically present in the area. There is no indication that this is due to changes in motility as the same heat-stressed males moved through unconditioned locations on the plate just as quickly as their unstressed counterparts. This change in behavior could either be a maladaptive interruption of normal decision making or an adaptive response to stress.

In the first scenario, temperature-driven changes in certain neuronal circuits may lead to hyperactivation of signaling downstream of either food or pheromone sensation. This could lead to behaviors such as increased turns, that result in males staying in the location of the food or hermaphrodite pheromones even in the face of either satiation or a lack of hermaphrodite presence, both of which lead to decreased mate searching. Interestingly, pheromone, food and temperature sensing connect in a shared neuronal circuit. The AWA, ASI and AWC neurons are all known to play key roles in both thermosensation/thermotaxis and attraction to food and/or hermaphrodite pheromones (Beverly et al., 2011; Biron et al., 2008; Clark et al., 2006; Goodman and Sengupta, 2018, 2019; Kuhara et al., 2008; Kimura et al., 2004; Mori and Ohshima, 1995; Ramot et al., 2008; Wan et al., 2019; White et al., 2007). Therefore, temperature-induced alterations in any of these neurons could lead to the changes we see in behavior.

In the second scenario, the change in pheromone attraction could be an adaptive response to increase reproductive success in a stressful environment. Hermaphrodites have been shown to increase pheromone excretion and be more likely to outcross when grown at 25°C compared with 20°C (Toker et al., 2022). The authors proposed that this response was a way to promote outcrossing in stressful environments, which has been shown to be an adaptive advantage (Morran et al., 2009a,b, 2011). It may be that heat-stressed males have a similar adaptive behavioral response to sensing hermaphrodite pheromones as a counterpart to increased reception of hermaphrodites to mating. However, under the stress condition we used here, the increased behavioral response to pheromones, which should increase a male's chance of locating a mate, does not ultimately lead to increased male fertility because of downstream heat stress-induced defects in mating drive, primarily the inability to copulate when a hermaphrodite is encountered.

The majority of C. elegans research is confined to the laboratory domesticated wild-type strain N2, with few studies exploiting the growing number of wild strains that more fully capture the genetic diversity of the species (Andersen et al., 2012; Cook et al., 2017; Sterken et al., 2015). Fewer studies still are performed with males from different wild strains. We have previously shown that there are appreciable differences in fertility between different wild strains under heat stress (Nett et al., 2019; Petrella, 2014; Sepulveda and Petrella, 2021). In this study, we have dissected fertility into several behavioral factors, which also differ between wild strains. The question remains, however, how relevant are these differences in the wild? Caenorhabditis elegans reproduces preferentially in rotting fruit and herbaceous plant matter at the soil surface where bacteria are plentiful, but where temperatures can reach those we used in our experiments (Félix and Duveau, 2012; Kiontke et al., 2011; Petersen et al., 2015; Richaud et al., 2018). Few males are present in the wild and outcrossing appears to be infrequent in both ideal environs such as rotting fruit (Richaud et al., 2018) and less desirable locations such as compost heaps (Barrière and Félix, 2005, 2007; Haber et al., 2005; Sivasundar and Hey, 2005). However, a single migratory male from a strain that does comparatively well under heat stress, e.g. JU1171, could fare very well in a population of hermaphrodites where he is more motile, is more likely to search for mates, and transfers sperm more reliably than a different genotype of male (Nett et al., 2019; Seidel et al., 2008, 2011). While evidence of outcrossing suggests it is infrequent in many natural populations of C. elegans, several studies have demonstrated that it is adaptive in novel and stressful environments (Carvalho et al., 2014; Chelo et al., 2014; Morran et al., 2009a,b; Teotonio et al., 2012). Climate change-associated changes in ambient temperature may prove to be one such novel and stressful environment that favors C. elegans males, especially those with superior reproductive features during moderate heat stress.

Some strains were provided by The Caenorhabditis Genetics Center (CGC) which is supported by the National Institutes of Health Office of Research Infrastructure Programs (P40 OD010440). We thank Douglas S. Portman PhD for sharing C. elegans strains. We thank Leigh R. Wexler PhD for imaging guidance. We thank Zoë A. Hilbert PhD for the R code used to analyze male leaving behavior.