Insulin-like peptides and the mTOR-TFEB pathway protect Caenorhabditis elegans hermaphrodites from mating-induced death

Lifespan is shortened by mating, but these deleterious effects must be delayed long enough for successful reproduction. Susceptibility to brief mating-induced death is caused by the loss of protection upon self-sperm depletion. Self-sperm maintains the expression of a DAF-2 insulin-like antagonist, INS-37, which promotes the nuclear localization of intestinal HLH-30/TFEB, a key pro-longevity regulator. Mating induces the agonist INS-8, promoting HLH-30 nuclear exit and subsequent death. In opposition to the protective role of HLH-30 and DAF-16/FOXO, TOR/LET-363 and the IIS-regulated Zn-finger transcription factor PQM-1 promote seminal-fluid-induced killing. Self-sperm maintenance of nuclear HLH-30/TFEB allows hermaphrodites to resist mating-induced death until self-sperm are exhausted, increasing the chances that mothers will survive through reproduction. Mothers combat males’ hijacking of their IIS pathway by expressing an insulin antagonist that keeps her healthy through the activity of pro-longevity factors, as long as she has her own sperm to utilize.


Introduction
The battle between the sexes is played out across the animal kingdom: males compete with one another to pass on their own genomes, often at the expense of female health, and females, in turn, must devise strategies to reduce the deleterious effects of mating in order to successfully complete reproduction. Sexual antagonism is evident in Caenorhabditis elegans, where signals from males kill hermaphrodites through three distinct mechanisms: germline activation (Shi and Murphy, 2014), seminal fluid transfer (Shi and Murphy, 2014), and, at high doses, male pheromone toxicity (Maures et al., 2014;Shi et al., 2017). Caenorhabditis elegans are hermaphrodites and thus do not require males to reproduce; in fact, males are normally rare in nature (~1/1000). Brief bursts of mating with males can be beneficial for the species due to increased outcrossing (Kimble, 1988), but males are quickly eliminated by dispersal (Wegewitz et al., 2008;Anderson et al., 2010) and by their own pheromones (Shi et al., 2017).
While germline activation is necessary for reproduction and the resulting lifespan decrease is shared both across sexes and species, seminal fluid (SF) killing appears to be used specifically for the purpose of 'sperm competition' (that is, competition between males) within the species: seminal fluid transfer induces death of mothers just after they have produced all of that male's progeny, eliminating the possibility of the mother re-mating with a different male (Shi and Murphy, 2014). We noted previously that mating-induced death occurs only after C. elegans hermaphrodites produce all their progeny (Shi and Murphy, 2014); this delay is necessary for successful reproduction, but the underlying mechanism allowing this balance was previously unknown. In an accompanying manuscript, Booth et al. report that self-sperm are protective against short-term mating-induced death specifically in hermaphroditic mothers (Booth et al., 2019). Here, we describe the molecular mechanism that hermaphrodites use to prevent early SF-induced death until self-sperm are depleted. This pathway utilizes competing insulin-like peptide agonists and antagonists of the DAF-2/Insulin receptor as well as mTOR signaling to regulate a set of protective (DAF-16/FOXO and HLH-30/TFEB) and detrimental (PQM-1) transcription factors, which in turn determine susceptibility to seminal-fluidinduced death.

Results
Self-sperm protects young hermaphrodites from male seminal-fluidinduced death The presence of self-sperm in young hermaphrodites protects them from brief mating-induced death (Booth et al., 2019; Figure 1A,B). Unlike older hermaphrodites, young hermaphrodites that mate during a brief interaction with males are protected from this mating-induced death (Booth et al., 2019; Figure 1A,B). By contrast, young feminized (i.e. lacking self-sperm; fog-2(q71), fem-1(hc17)) worms exhibit significant acceleration of death after brief mating with males (Booth et al., 2019; Figure 1C, Figure 1-figure supplement 1A), suggesting that self-sperm are protective.
Previously we found that mating induces hermaphrodite shrinking through male sperm activation of the hermaphrodite's germline, and that separately, transfer of male seminal fluid (SF) causes DAF-16/FOXO cytoplasmic localization and inactivation, subsequently leading to death of the mother (Shi and Murphy, 2014). To determine whether brief mating acts through the male sperm/germline pathway or through the seminal fluid pathway, we mated fer-6 males, which are spermless but still contain seminal fluid ( Figure 1D,E; Argon and Ward, 1980;Gems and Riddle, 1996), or fog-2 males, which contain male sperm and thus essentially are like wild type (Schedl and Kimble, 1988), with fog-2 (self-spermless) females; both sperm-containing and spermless males induce death in young and old fog-2 mothers at equal rates, suggesting that the death-inducing factor is not sperm. The dispensability of male sperm suggests that brief mating-induced death acts through the seminal fluid pathway.
Because high doses of male pheromones can also kill hermaphrodites (Maures et al., 2014;Shi et al., 2017), we tested whether pheromone-less daf-22 (Ludewig and Schroeder, 2013) males can also induce killing after brief mating. However, daf-22 males induce the same brief mating lifespan effect as wild-type males. Similarly, 2 hr treatment on high-dose male pheromone-conditioned plates ('MCP'; plates pre-treated with 30 males, then removed prior to experiment) does not shorten the lifespan of either wild-type or fog-2 females ( Figure 1F-I). Together, these data suggest that self-sperm protects against brief mating-induced death in a manner that is independent of both male sperm and pheromone effects, and acts through the seminal fluid/IIS pathway.

Brief mating-induced death acts through the IIS/FOXO seminal fluid pathway
We previously showed that Insulin-IGF-1 signaling (IIS) mediates seminal fluid-induced killing by inducing DAF-16/FOXO nuclear exit (Shi and Murphy, 2014). We find that daf-16 mutants are also susceptible to brief mating, even when young (Figure 2A), suggesting that DAF-16 activity is critical for resistance to brief mating. Long-lived daf-2/Insulin receptor mutants, although susceptible to long-term mating, are better protected than wild-type worms on Day 8 of adulthood ( Figure 2B), exhibiting the opposite phenotype as daf-16 mutants. Furthermore, DAF-16::GFP in fog-2 (selfspermless) mutants moves out of the nucleus upon brief mating ( Figure 2C-D).

Self-sperm modulate expression of ins-37
We reasoned that the factor that provides self-sperm protection against seminal fluid-induced killing should already be present in hermaphrodites prior to mating; therefore, we compared the expression profiles of fog-2 (self-spermless) and wild-type hermaphrodites at L4, a stage after all self-sperm have been made. Not surprisingly, wild-type worms express more major sperm proteins and other sperm-associated genes than do self-spermless fog-2 mutants (  Supplementary file 2). However, of the remaining differences, the conserved insulin-like peptide ins-37 is strikingly upregulated in wild-type compared to fog-2 L4 hermaphrodites ( Figure 3A,B), suggesting that this germline-expressed (Ebbing et al., 2018) signaling molecule is only highly expressed when self-sperm are present. The identification of an insulin-like peptide is also consistent with our finding that brief mating utilizes the seminal fluid/IIS pathway ( Figure 1D-E).
Elimination of a factor that confers protection from seminal-fluid-induced killing should render young, sperm-containing worms susceptible. Indeed, RNAi of ins-37 in wild-type (N2) worms shortens their lifespan after brief mating ( Figure 3C, Figure 3-figure supplement 2B), despite the presence of self-sperm, suggesting that ins-37 is key for self-sperm-mediated protection from mating.
INS-37 may act as a self-sperm-induced antagonist of the insulin receptor, DAF-2, consistent with a recent report that INS-37 acts a DAF-2 antagonist during development (Zheng et al., 2018). Together, our data suggest that self-sperm protect hermaphrodites from brief mating-induced death through their expression of the INS-37 antagonist, which prevents DAF-2/Insulin receptor activation and the subsequent loss of expression of protective genes by the DAF-16/FOXO transcription factor (Murphy et al., 2003;Tepper et al., 2013).
The HLH-30::GFP strain overexpresses functional HLH-30 protein, so we wondered whether these animals are better protected from brief mating-induced death; indeed, no lifespan reduction was observed in the HLH-30::GFP strain after mating for 12 hr, a length of mating that shortens WT lifespan ( Figure  TOR promotes seminal-fluid-induced killing TOR signaling has been implicated in longevity regulation from yeast through humans, including C. elegans (Kapahi and Hansen, 2010;Kaeberlein and Shamieh, 2010). TOR negatively regulates TFEB/HLH-30's nuclear localization and activity (Nakamura et al., 2016). Through genome-wide transcriptional comparisons, we observed that most positive regulators of TOR signaling were modestly downregulated in conditions in which worms are protected from brief mating ( Figure 5A,B). Therefore, we wondered whether TOR signaling is also involved in longevity regulation in response to mating. Indeed, we found that reduction of the worm TOR homolog, let-363, prevented the lifespan decrease caused by brief mating in young spermless (fog-2) females ( Figure  Similarly, RNAi reduction of other TOR pathway components, including rict-1, daf-15, and raga-1, also prevented brief mating-induced death in Day 3 fog-2 females, and   Figure 5C). Together, these results suggest that TOR promotes SF-mediated killing through its inhibition of TFEB/HLH-30 nuclear localization and activation.

PQM-1 is required for seminal-fluid-induced killing
Previously, we found that the Zn-finger transcription factor PQM-1 acts downstream of IIS and antagonistically to DAF-16/FOXO nuclear localization under normal conditions (Tepper et al., 2013). pqm-1 mutant males are immune to both mating-induced and male pheromone-induced death (Shi et al., 2017), but PQM-1's role in hermaphroditic death was unknown. Unlike daf-16 mutants, which are susceptible to brief mating even as young adults (Figure 2A), pqm-1 mutants are impervious to brief mating ( Figure 6A); in fact, they are even resistant to long-term (24 hr) mating ( Figure 6B, Figure 6-figure supplement 1A-B). While DAF-16 leaves the nucleus upon mating (Shi and Murphy, 2014) (Figure 2C-D), PQM-1 translocates into the nucleus ( Figure 6C-E). These results suggest that PQM-1 activity is key for seminal fluid-induced killing, acting in an opposite manner to DAF-16/FOXO.

Insulin-like peptide agonists mediate SF killing
Because self-sperm-mediated protection from SF-induced death correlated with the expression of an antagonist of DAF-2, ins-37, we reasoned that a competing, agonist insulin signal might induce the IIS pathway upon Seminal Fluid transfer. Indeed, upon mating, fog-2 hermaphrodites (which are killed by brief mating) express strikingly higher levels of ins-8 than do pqm-1 hermaphrodites, which are not affected by brief mating ( Figure 7A; Supplementary file 4). We confirmed the transcriptome analysis using a transcriptional reporter: mating significantly increased the expression level of ins-8p::gfp, particularly in the anterior part of the intestine ( Figure 7B). INS-7, which has been previously shown to act as a DAF-2 agonist that shortens lifespan (Murphy et al., 2003;Murphy et al., 2007), and functions in a feed-forward loop to reduce DAF-2 activity (Murphy et al., 2007), is also consistently induced but to a lesser degree. We hypothesized that INS-8 might act as a DAF-2 agonist under mating conditions, and thus might induce seminal-fluid-mediated death; if true, reduction of ins-8 should prevent SF-induced death, even in self-spermless animals. Indeed, while brief mating of Day three fog-2 females with wild-type males shortens lifespan ( Figure 7E), knockdown of ins-8 ( Figure 7C) or ins-7 ( Figure 7D

A model of self-sperm protection from Seminal Fluid-induced killing
The presence of self-sperm in hermaphroditic mothers slows the deleterious effects of mating with males, whose seminal fluid pathway 'hijacks' the mother's longevity and  (Shi and Murphy, 2014;Lin et al., 2001) and HLH-30 nuclear localization and activation. INS-7 induction in the intestine and its feed-forward effect on DAF-2 coordinates the IIS pathway across the entire animal (Murphy et al., 2007), which causes rapid deceleration of stress responsive pathways. PQM-1 appears to drive HLH-30 out of the nucleus in an antagonistic manner, as PQM-1 does with DAF-16/FOXO (Tepper et al., 2013). Like HLH-30, PQM-1 is negatively regulated by TOR signaling (Dowen et al., 2016), but since PQM-1's role is detrimental, the loss of each of these factors prevents mating-induced death (Figure 8). Together, this network of signaling pathways and transcription factors coordinates both the resistance and susceptibility to mating-induced death.  The following figure supplements are available for figure 5:  The ability of hermaphroditic mothers to slow mating-induced decline as long as she has her own self-sperm suggests that hermaphrodites have developed strategies to combat the deleterious effects of mating with males, staving off early death until mating becomes beneficial to the mother. While early mating is detrimental to the mothers, late mating (i.e., after all self-sperm are utilized) is beneficial, as it allows the mother to use her remaining oocytes. Moreover, the involvement of TOR suggests that nutritional status may also influence mating-induced death. Thus, a dynamic system that is responsive to changing sperm levels and nutrient signals may allow such a shift. The expression of a DAF-2 antagonist, INS-37, allows hermaphrodites to resist and delay C. elegans' main mechanism of competition with other males, seminal fluid-induced killing, until self-sperm are exhausted, increasing the chances that the mother will survive through the end of her reproductive period. Such strategies, coupled with the increased attraction of males that accompany the depletion of self-sperm (Morsci et al., 2011), enable hermaphrodites to delay both the odds of mating, through behavioral changes (Morsci et al., 2011) and the detrimental effects of mating (through molecular signaling), before the full utilization of their own sperm. This increases the chance that they will propagate their own genome preferentially before succumbing to mating. The balance shifts in the male's favor as the hermaphrodite's sperm are depleted, causing her death shortly after producing his progeny, and preventing further mating by other males. This insulin peptide-based arms race regulates a transition from preventing to promoting the activation of the insulin signaling pathway, resulting first in protection from death during early reproduction, and then accelerated death of the mother immediately after reproduction. Together, these signals tune and optimize reproduction for the hermaphroditic mother, even when confronted with deleterious effects from males.  nuclear and cytoplasmic GFP intensity was measured in the most anterior nuclei (easily recognizable) of the intestine for worms in each group. Averages were compared using a t-test. (F) Knocking down pqm-1 restores nuclearly localized HLH-30 in mated hermaphrodites. Quantitation of nuclear HLH-30:: GFP in mated hermaphrodites treated with pqm-1 RNAi (same as Figure 5H). (G) Knocking down pqm-1 blocks the lifespan decrease normally induced by 2 hr mating in hlh-30(tm1978) mutants.

Experimental methods
Mating setup 60 mm NGM plates were used to set up group mating. Each 60 mm NGM plate was seeded with OP50 to make a~3 cm diameter bacterial lawn 2 days before mating. About 50 hermaphrodites and 150 young (day 1 À day 2 of adulthood) males were transferred onto the plate. Two hours later (for all the brief mating lifespan assays) or 12/24 hr later (for long term mating assays), the hermaphrodites were transferred onto newly seeded 60 mm NGM plates in the absence of males for lifespan assays.

Lifespan assay
Lifespan assays were performed at 20C. No FUdR was added to the plates. About 25 synchronized Day 1 hermaphrodites were transferred onto each plate. The hermaphrodites were transferred daily onto new seeded plates in the first week of the lifespan assay. Afterwards, they were transferred once every 2 days. Group mating was set up at the beginning of Day 3 of adulthood (or as noted in text) for brief mating lifespan assays. When RNAi was used in lifespan assay, RNAi treatment always started from eggs for all the experiments in this study. Worms were moved onto standard mating plates (NGM seeded with OP50, no RNAi is included) for 2 hr mating to eliminate any potential influence on male mating efficiency caused by RNAi. Then, worms were moved back onto RNAi plates for the remaining of lifespan assays. KaplanÀMeier analysis with logÀrank (MantelÀCox) method was performed to compare the lifespans of different groups. 'Bagged' worms were censored on day of the event.

Male-conditioned plates (MCP) setup
MaleÀconditioned plates for lifespan assays were prepared similar to the previous description (Maures et al., 2014). Briefly, 60 ml of OP50 was dropped onto each 35 mm NGM plate to make a bacterial lawn of~25 mm diameter. 30 young Day 1 wildÀtype males (fogÀtwo males) were transferred onto each plate. Two days later, they were removed and 25 hermaphrodites for lifespan assays were immediately transferred onto these male-conditioned plates (MCP). These MCPs were prepared throughout the course of the lifespan assays to ensure fresh MCP plates ere available.
GFP fluorescent worm imaging and quantitation HLHÀ30::GFP (BC11288) Multiple HLHÀ30::GFP strains are available at CGC. Since we need to mate the worms and monitor their postÀmating lifespans, strains with roller markers are not ideal for our purpose, because roller phenotype could potentially affect the mating efficiency especially in the brief mating set up. Therefore, we chose BC11288, in which gfp was inserted after the start codon of the longest hlhÀ30 transcript. Another advantage of BC11288 is that its fluorescence is very strong, allowing us to directly image the freely moving worms under the scope, which could reveal HLHÀ30 localization more accurately, since mountingÀinduced stress could make HLHÀ30 more nuclear (Lapierre et al., 2013). 1-2 min of videos were taken using iphone connected directly with Leika M205 FA. Worms in focus and with proper exposure were scored in two ways: i) each worm was assigned a category based on HLHÀ30::GFP localization (strongly nuclear, weakly nuclear, and diffuse); a ChiÀsquare test was used to determine the significance between different groups; ii) how many recognizable intestinal  nucleus can be counted. The average number of intestinal nuclei with nuclearly localized HLHÀ30:: GFP was compared with each other using the unpaired tÀtest. When HLH-30::GFP was assayed in the mutant background, the following cross strategy was used: HLH-30::GFP hermaphrodites were crossed with mutant males (i.e. fog-2), F1 males with positive GFP signal were crossed with mutant fog-2 hermaphrodites again, hermaphrodites from F2 generation were directly used in the assay. Half of the were fog-2 homozygous mutant, half of them only have one copy of -mutant allele, thus have functional self-sperm.

PQMÀ1::GFP (OP201)
Worms were mated on day 3 of adulthood for 24 hr. 20-30 worms of each group were imaged by Nikon Eclipse 90i on day 6 of adulthood. PQMÀ1::GFP localization was quantified in two ways. i) each worm was assigned a category based on PQMÀ1::GFP localization (nuclear, cytoplasmic, and diffuse(both nuclear and cytoplasmic)). ChiÀsquare test was used to determine the significance between different groups. ii) the mean gfp intensity of the most anterior intestinal nuclei and the Figure 7 continued to unmated). (D) fog-2;ins-7(RNAi) unmated: 18.2 ± 0.6 days; mated with fog-2 males: 17.7 ± 1.1 days, p=0.9380; mated with fer-6 males: 17.2 ± 0.7 days, p=0.3994 (compared to unmated). (E) Controls for Figure 7C-D. fog-2 on vector control RNAi unmated: 18.4 ± 0.9 days; mated with fog-2 males: 13.3 ± 0.7 days, p<0.0001; mated with fer-6 males: 15.4 ± 0.6 days, p=0.0032 (compared to unmated). (F) Knocking down ins-7 and ins-8 also restores nuclearly localized HLH-30 in mated hermaphrodites. . Model of self-sperm mediated protection from seminal fluid killing Prior to mating, the presence of selfsperm maintains high ins-37 expression. INS-37 antagonizes DAF-2 activity and promotes DAF-16 and HLH-30 maintenance in the nucleus. Self-sperm also inhibits TOR signaling, which regulates HLH-30 nuclear localization. Seminal fluid transfer upon mating increases expression of ins-8 and ins-7; these agonists activate DAF-2, promoting PQM-1 nuclear localization and DAF-16 and HLH-30 nuclear exit, resulting in premature death. INS-7 acts in a feed-forward loop (Murphy et al., 2007), further accelerating DAF-2 activation. DOI: https://doi.org/10.7554/eLife.46413.017 nearby cytoplasmic region was measured by image J (DIC picture of the same worm was used to determine the position and the size of the nuclei, then switch to GFP picture to measure gfp intensity. The same outline of the nucleus was used and moved to the cytoplasmic region next to the nucleus, and the gfp intensity of this cytoplasmic region was measured.) The Nuc/Cyto ratio was calculated using the two values. TÀtest analysis was performed to compare the Nuc/Cyto PQMÀ1:: GFP intensity of mated and unmated worms.

DAF-16::GFP (TJ356)
Worms were mated on day 1 of adulthood for 24 hr. 20-30 worms of each group were imaged by Nikon Eclipse 90i next day. DAFÀ16::GFP localization (nuclear, cytoplasmic, and diffuse(both nuclear and cytoplasmic)). ChiÀsquare test was used to determine the significance between different groups. RNAi treatment started from eggs.

PinsÀ8::GFP (HT1704)
Worms were mated on day 3 of adulthood for 24 hr. 20-30 worms of each group were imaged by Nikon Eclipse 90i on day 5 of adulthood. Image J was used to measure the mean and the maximum GFP intensity of the whole body area. TÀtest analysis was performed to compare the GFP intensity of different group of worms.

Microarrays
About 200 L4 hermaphrodites were collected for each condition and replicate ( Figure 3A  Four biological replicates were performed. RNA was extracted by the heatÀvortexing method. Two-color Agilent microarrays were used for expression analysis. Significant differentially expressed gene sets were identified using SAM (Tusher et al., 2001). g:Profiler was used for GO term analysis and enriched TF motifs prediction (Reimand et al., 2016).
The following datasets were generated: Author (