The effects of Bacillus subtilis on Caenorhabditis elegans fitness after heat stress

Abstract Microbes can provide their hosts with protection from biotic and abiotic factors. While many studies have examined how certain bacteria can increase host lifespan, fewer studies have examined how host reproduction can be altered. The nematode Caenorhabditis elegans has been a particularly useful model system to examine how bacteria affect the fitness of their hosts under different contexts. Here, we examine how the bacterium Bacillus subtilis, compared to the standard C. elegans lab diet, Escherichia coli, affects C. elegans survival and reproduction after experiencing a period of intense heat stress. We find that under standard conditions, nematodes reared on B. subtilis produce fewer offspring than when reared on E. coli.However, despite greater mortality rates on B. subtilis after heat shock, young adult nematodes produced more offspring after heat shock when fed B. subtilis compared to E. coli. Because offspring production is necessary for host population growth and evolution, the reproductive advantage conferred by B. subtilis supersedes the survival advantage of E. coli. Furthermore, we found that nematodes must be reared on B. subtilis (particularly at the early stages of development) and not merely be exposed to the bacterium during heat shock, to obtain the reproductive benefits provided by B. subtilis. Taken together, our findings lend insight into the importance of environmental context and interaction timing in shaping the protective benefits conferred by a microbe toward its host.

genotype, and hosts of different genotypes may also differ in the level of benefit that they receive from microbial association (Murfin et al., 2015;Parker, Hrček, McLean, & Godfray, 2017). Taken together, the hosts, the microbes, and the environment all shape the nature of these interactions over ecological time, which in turn may shape the evolutionary trajectories of the host and microbial populations. Here, we utilize Caenorhabditis elegans, a well-characterized invertebrate model amenable to a range of experimental manipulations, to test the effects of environmentally obtained bacteria on host fitness under stress.
The nematode C. elegans has been extensively used as a model system to study host-microbe associations (Clark & Hodgkin, 2014;Kurz & Ewbank, 2000;Zhang et al., 2017). Caenorhabditis elegans ( Figure 1) has a natural interaction with microbes in that it feeds on bacteria and fungi in decomposing plant matter (Frézal & Félix, 2015). The nematode has a grinder in the pharynx region that crushes microbes that it consumes; however, some microbes survive the grinder and colonize the gut of the nematode (Gibson et al., 2015;Portal-Celhay & Blaser, 2012). Some of these persistent microbes are pathogenic to the host (Couillault & Ewbank, 2002), some are commensal (Clark & Hodgkin, 2014), and some are beneficial (Zhang et al., 2017). Specifically, some microbes have been shown to increase nematode lifespan under environmental stresses (Donato et al., 2017;Grompone et al., 2012;Gusarov et al., 2013;Leroy et al., 2012;Nakagawa et al., 2016), an important finding given that C. elegans is a model system to study longevity and aging (Cabreiro & Gems, 2013;Garigan et al., 2002;Johnson, 2008). Two studies found that the soil bacterium, Bacillus subtilis, was able to increase C. elegans survivorship after heat shock relative to exposure to the standard lab diet, Escherichia coli (Donato et al., 2017;Gusarov et al., 2013). These studies found that B. subtilis nitric oxide (NO) production and biofilm formation in the host's gut resulted in elevated host lifespan postheat stress. For this interaction to impact host-microbe evolution, the bacteria would not only need to increase survival but would also need to increase host reproduction after heat shock. Caenorhabditis elegans generally exhibit little to no fecundity after exposure to intense heat stress (Aprison & Ruvinsky, 2014), and it is unclear if interactions with B. subtilis could mitigate this substantial fitness loss. In this study, we measure the effects of C. elegans interactions with B. subtilis on nematode fitness (encompassing both survival and fecundity) after a stressful heat event. Additionally, we determine how exposure to B. subtilis at different time points during development affects C. elegans fitness.

| Nematode and bacterial strains
All nematodes in this study were the C. elegans N2 Bristol strain, which were maintained on E. coli OP50 prior to experiments. We used B. subtilis 168 and E. coli OP50 in all experiments. In the first two experiments, we also used B. subtilis Δnos, which lacks the ability to produce NO. We obtained both B. subtilis strains from the study examining the role of NO in C. elegans longevity and survivorship postheat stress (Gusarov et al., 2013) . We grew both B. subtilis strains and E. coli on nematode growth medium (NGM) plus glucose (2%) and arginine (0.5 mM) for all experiments (Gusarov et al., 2013). For experiments involving fecundity, we transferred nematodes to GFP-labeled OP50 (OP50-GFP, grown on NGM) to allow them to produce offspring. We used OP50-GFP to control for the bacterium that nematodes were exposed to during heat shock recovery. GFP-labeled E. coli is different from E. coli OP50 but is still relatively neutral with respect to C. elegans fitness.

| Survival of 6-day-old nematodes on B. subtilis and E. coli
We first compared the short-term survivorship over the 6 hr postheat stress of nematodes across B. subtilis 168, B. subtilis Δnos, and E. coli by performing the heat shock experiment done in the previous studies examining the role of B. subtilis on host lifespan after heat stress (Donato et al., 2017;Gusarov et al., 2013) . We surface sterilized C. elegans N2 eggs using an established alkaline hypochlorite protocol (Stiernagle, 2006) and reared L1 larvae on E. coli until they reached L4/young adult (on day 3). We then transferred nematodes to either B. subtilis, B. subtilis Δnos, or E. coli.
When nematodes were 5 days old, we transferred them to new plates of the appropriate bacteria to prevent mixing of generations. Prior to heat shock, all nematodes were kept at 20°C. On the next day, when they were 6-day-old adults, we heat shocked the nematodes in an incubator set at 34°C. After 3 hr, we removed a set of replicate plates for each bacterial treatment from the incubator and scored survival by prodding with a platinum pick to determine signs of movement (Donato et al., 2017;Gusarov et al., 2013;King et al., 2016)

| Survival and fecundity of 3-day-old nematodes on B. subtilis and E.coli
Because nematodes cease egg production after about 6 days (Altun & Hall, 2009), to assess fitness effects of B. subtilis association, here, we heat shocked nematodes when they were young adults and still capable of producing offspring. Specifically, we investigated 3-dayold nematode survival and fecundity under standard and heat shock conditions on B. subtilis, B. subtilis Δnos, and E. coli. We surface sterilized C. elegans N2 eggs and reared L1 larvae on each bacterium for 3 days at 20°C until they reached young adulthood. We then placed nematodes in an incubator set at either 20°C (standard temperature) or set at 34°C for 6 hr. We used three replicate populations per bacterium per temperature setting, for a total of 18 populations with approximately 200 nematodes per population.
To measure survival, here, and in all subsequent experiments, we determined the proportion of nematodes that were alive 6 hr after the heat shock period (based on prodding, as above), not how long they lived after heat shock (i.e., survivorship or lifespan). Afterward, we washed nematodes from each replicate population with M9 and transferred all adults from each population to another plate seeded with E. coli OP50-GFP to produce offspring, where they were maintained at 20°C. Two days postheat shock, we counted the larvae on each plate. The total number of adults was counted prior to heat shock. Since all plates were transferred using the same protocol, we assumed a similar number of nematodes were transferred to OP50-GFP to produce offspring. We did not count the number of live adults 2 days postheat shock because nematodes could have produced offspring and subsequently died before being counted.
We calculated the average number of offspring per heat shocked adult to determine relative differences in fecundity between treatments. This measure accounts for both difference in survival and difference in fecundity.

| Assaying importance of exposure window, experiment 1
Similar to the survival and fecundity assay described above, we surface sterilized N2 eggs and reared L1 larvae on either B. subtilis or E. coli for 3 days, with six replicates per bacterium. We then transferred three B. subtilis replicates to E. coli and the other three replicates to B. subtilis, and similarly transferred the E. coli replicates to either E. coli or B. subtilis, for a total of 12 populations with approximately 100 nematodes per population. We heat shocked nematodes at 34°C for 6 hr, scored survival after 6 hr as above, then transferred them to plates seeded with E. coli OP50-GFP to lay eggs, where they were maintained at 20°C. We quantified larval offspring 2 days later.

| Assaying importance of exposure window, experiment 2
Similar to the first exposure assay, we surface sterilized N2 eggs and transferred about 100-200 larvae to a new lawn of bacteria each day as indicated in Figure 2 (with four replicates per treatment, for a total of 20 populations). After reaching adulthood, we heat shocked nematodes for 6 hr at 34°C, scored survival after 6 hr as above, then transferred nematodes to plates seeded with E. coli OP50-GFP to lay eggs, where they were maintained at 20°C. We quantified larval offspring 2 days after.

| Colonization of day 2 larvae and adult nematodes
Following a modified protocol from Vega & Gore (2017), we determined whether day 2 larvae reared on B. subtilis harbored live bacterial cells. Briefly, after surface sterilizing N2 eggs, we transferred roughly 100 larvae to a lawn of B. subtilis. The following day (day 2 of Figure 2), we washed larvae three times with cold 0.01% Triton X-100 in M9 and incubated them in bleach (1:1,000 diluted) for 15 min at 4°C to remove surface bacteria. We then treated larvae with a solution of 0.25% sodium dodecyl sulfate (SDS) + 3% dithiothreitol (DTT) for 20 min. After washing with 0.01% Triton X-100 in M9, we transferred about 10-20 larvae to a well of a 96-well plate (five wells total), each well containing a small amount of sterile silicon carbide grit and 0.01% Triton X-100 in M9. We briefly disrupted the samples using a Qiagen TissueLyser II homogenizer. After plating the content onto LB plates, we grew the bacteria for 2 days before quantifying colony-forming units. For colonization of adults, we reared surfaced-sterilized N2 eggs on B. subtilis until adulthood, then heat shocked nematodes for 6 hr at 34°C. We subsequently washed and homogenized the nematodes using the same protocol as the day 2 larvae, crushing five adults in each of five wells of the 96-well plate.

| Statistical analysis
To analyze short-term survivorship of 6-day-old hosts under different bacterial treatments, we used a Cox proportional hazards model with the Coxph function of the Survival package in R (Therneau & Grambsch, 2000). For subsequent experiments, we analyzed survival using a generalized linear model (GLM) with a binomial distribution and logit link function. For fecundity, we used a GLM with a normal distribution and identity link function. We then performed contrast tests to compare bacterial treatments. We used JMP Pro (v.13) for the GLM analyses.

| Bacillus subtilis differentially affects survival of old and young adult hosts and provides a reproductive benefit in young adult C. elegans
We performed a heat shock experiment, similar to previous studies (Gusarov et al., 2013), that allowed us to directly compare wild-type then transferred them to one of the three bacterial strains, where they remained for another 3 days before shifting to 34°C (heat shock temperature) for 6 hr. We found that there was a small but nonsignificant decrease in host survival when nematodes were exposed to B. subtilis Δnos compared to hosts exposed to wildtype B. subtilis ( Figure 3; 2 1 = 3.36, p = 0.07). However, we found a significantly large difference in host survival between wild-type B. subtilis and E. coli immediately after 6 hr of heat shock ( 2 1 = 28.05, p < 0.001). Therefore, B. subtilis conferred greater host survival following heat shock, but the protective effects of NO can, at most, only account for a portion of the benefits conferred by B. subtilis.
Since the nematodes above, as in previous studies, were heat shocked at postreproductive age (egg laying ceases after about 6 days from the time nematodes hatch (Altun & Hall, 2009)), we could not determine whether these bacteria affected host fecundity after heat shock. To this end, we examined the survival and fecundity of young adult C. elegans when reared on B. subtilis 168, B. subtilis Δnos, and E. coli OP50 under standard and heat shock conditions. Three-dayold nematodes reared on their respective bacterium at 20°C were either left at the standard temperature or heat shocked. We found no difference in host survival under standard conditions (measured at the same time we scored survival of heat shocked nematodes) regardless of the hosts' bacterial association (Figure 4a). However, nematodes reared on E. coli produced more offspring than on either

| Development on B. subtilis is necessary for reproductive benefit
To determine whether the decrease in survival and increase in offspring production on B. subtilis was due to host larval development on B. subtilis or simply due to exposure to the bacterium during heat shock, we compared nematodes reared on B. subtilis that were then heat shocked on E. coli, and vice versa. Nematodes that developed on B. subtilis and were heat shocked on E. coli exhibited the lowest survival ( Figure 5a; 2 1 = 97.10, p < 0.001). However, development on B. subtilis resulted in higher fecundity after heat shock regardless of which bacterium nematodes were exposed to during heat stress ( Figure 5b; 2 1 = 15.96, p < 0.001). The reproductive benefit conferred by B. subtilis was therefore predominantly dependent upon the development of hosts on B. subtilis.

| Early exposure to B. subtilis is more beneficial for hosts than later exposure
We then asked whether the age at which nematodes are exposed to B. subtilis has an effect on the hosts' survival and fecundity.
We varied exposure time to B. subtilis by transferring nematodes to the indicated bacterium each day (Figure 2). We found that the time at which the host is exposed to B. subtilis affects both survival and fecundity upon heat stress (Figure 6a; 2 4 = 114.61, p < 0.001; Figure 6b; 2 4 = 35.02, p < 0.001). Specifically, exposure to B. subtilis during the first 2 days of host development is critical for nematodes to obtain the reproductive benefit conferred by B. subtilis upon heat stress (Figure 6b, early exposure treatment vs. all other treatments).
F I G U R E 3 Survivorship of older adult (6-day-old) nematodes under heat shock. Nematodes were reared on Escherichia coli OP50 until L4/young adult stage (at about 3 days old), then subsequently transferred to the indicated bacterium. They were heat shocked at 34°C 3 days later. After 3 and 6 hr, replicate plates were removed from the heat and scored for survival. Error bars indicate standard errors. There were three replicate populations per time point per bacterium, each population containing ~20 nematodes Furthermore, compared to when nematodes were exposed to B. subtilis throughout development and during heat shock, early exposure to B. subtilis was more beneficial in terms of fecundity and survival ( Figure 6, full exposure vs. early exposure). Overall, these results demonstrate that nematodes gained the most benefits when exposed to B. subtilis early, whereas later exposure to B. subtilis conferred no greater benefit than exposure to E. coli alone (Figure 6b).

| Bacterial colonization of host
Because nematodes benefit the most when either exposed to B. subtilis completely or early on in development, we asked whether these benefits were associated with live B. subtilis in the nematode gut. We    (Aprison & Ruvinsky, 2014. Even though hosts undergo some sperm damage, they can produce viable offspring if allowed sufficient time to recover and reproduce (Aprison & Ruvinsky, 2014). Thus, given this reproduction after heat shock, there is potential for B. subtilis protection to shape host evolution.

| D ISCUSS I ON
Our survival results for heat shocked older adult nematodes support those of previous studies, in that B. subtilis led to increased host survivorship compared to E. coli (Gusarov et al., 2013;Donato et al., 2017 Investment in longevity is hypothesized to trade-off with reproduction (Mukhopadhyay & Tissenbaum, 2007). While we did not measure total host lifespan, our heat shock data suggests that increased survival of C. elegans on E. coli led to a reproductive cost, the converse of which is true for hosts on B. subtilis. Furthermore, because nematodes have more offspring on E. coli than B. subtilis in the absence of heat shock (Figure 4b), the interaction between B. subtilis and C. elegans is context-dependent-hosts incur a cost to harboring B. subtilis in the absence of heat stress, but benefit reproductively under heightened temperatures. Furthermore, while our heat shock temperature (34°C) is much higher than the range at which C. elegans is generally reared (15-25°C), hosts were able to reproduce at a rate at which the population could at least replace itself on B. subtilis, compensating for the reduced number of surviving adults compared to E. coli.
Exposure to beneficial microbes during the early stages of host development could be important for host resistance to biotic and environmental stresses during adulthood. For example, a study found that prior diet can affect C. elegans preference for harmful Burkholderia bacteria (Cooper, Carlson, & LiPuma, 2009). Another study examining the consequences of early exposure to pathogens in C. elegans found increased resistance to pathogens and heat stress during adulthood (Leroy et al., 2012). Host fecundity may also differ depending on the bacteria the host is exposed to during development and adulthood. Our study provides support for this phenomenon: exposure to B. subtilis during early stages of development was enough for C. elegans to remain reproductively viable after a period of heat shock (Figure 6b). By contrast, exposure to B. subtilis as an older larva or adult did not benefit hosts greatly when they underwent heat stress (Figure 5b/6b). This suggests that exposure to the bacterium at an early point during nematode development may be critical in priming the host to respond to heat shock as an adult. The bacterium may have entered nematodes as spores or formed spores upon entrance, F I G U R E 5 Survival and fecundity of young adult nematodes heat shocked on either Bacillus subtilis or Escherichia coli. Nematodes were reared on the first bacterium indicated before being transferred to the bacterium they were heat shocked on when they reached adulthood. (a) Survival after 6 hr and (b) fecundity of nematodes under heat shock conditions. Error bars indicate standard errors. Each data point represents a replicate population, with each population consisting of ~100 nematodes. *p < 0.05, **p < 0.01, ***p < 0.001  and between treatment groups, thus resulting in approximately equivalent levels of fecundity between treatments and replicates.
However, we observed substantial variance between replicates and significant differences between treatments ( Figure 6). Therefore, the results are more likely due to B. subtilis colonization than nutrient acquisition via B. subtilis digestion. This variance is also consistent with previous work examining bacterial growth in C. elegans, where stochasticity has a significant effect on bacterial abundance (Vega & Gore, 2017). Finally, a recent study has shown that B. subtilis extends C. elegans lifespan postheat shock through the production of biofilm (Donato et al., 2017), demonstrating that live cells are present in and actively colonizing the host gut. Taken together, these results suggest that the fitness benefits conferred by B. subtilis postheat shock is likely largely due to host-microbe interactions within the host.
Our study demonstrates that interacting with the appropriate microbe under stressful conditions can benefit hosts in terms of reproduction, which could have significant implications for host population growth and evolution in the long term. This could select for association with the microbe in future generations, leading to the potential for coevolution of the partners within the framework of a mutualistic symbiosis. Importantly, the mechanistic nature of the beneficial interaction between C. elegans and B. subtilis may not dictate the system's capacity for mutualistic coevolution. Studies have shown that certain bacteria may serve roles in C. elegans distinguishable from diet (Berg et al., 2016;Cabreiro & Gems, 2013;Dirksen et al., 2016;Gerbaba, Green-Harrison, & Buret, 2017;Zhang et al., 2017). Because B. subtilis can survive within the host, the host and microbe have the potential for mutualistic coevolution when their fitness aligns. However, even if the fitness differences we observed are due to nutrients obtained via B. subtilis digestion and not due to the impact of having maintained association with live bacteria, then mutualistic coevolution is still possible. For example, leaf-cutter ants and the fungi they cultivate have been coevolving with each other for millions of years, even though the fungi serve primarily as the ant's food source (Schultz & Brady, 2008;Weber, 1966).
While more work is necessary to determine the mechanism by which B. subtilis increases C. elegans fecundity after heat stress, our work provides further evidence for the critical role that bacteria can play in the evolution and ecology of their hosts.

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests.

AUTH O R CO NTR I B UTI O N S
KLH, NMG, and LTM designed the study. KLH performed the experiments. KLH and LTM analyzed the data. KLH wrote the manuscript with input from NMG and LTM.

DATA ACCE SS I B I LIT Y
All relevant data are available at the Dryad Digital Repository (https://doi.org/10.5061/dryad.7kc589v).