Oocyte mitochondria link maternal environment to offspring phenotype

During maturation oocytes undergo a recently discovered mitochondrial proteome remodeling event in flies1, frogs1, and humans2. This oocyte mitochondrial remodeling, which includes substantial changes in electron transport chain (ETC) subunit abundance1,2, is regulated by maternal insulin signaling1. Why oocytes undergo mitochondrial remodeling is unknown, with some speculating that it might be an evolutionarily conserved mechanism to protect oocytes from genotoxic damage by reactive oxygen species (ROS)2. In Caenorhabditis elegans, we previously found that maternal exposure to osmotic stress drives a 50-fold increase in offspring survival in response to future osmotic stress3. Like mitochondrial remodeling, we found that this intergenerational adaptation is also regulated by insulin signaling to oocytes3. Here, we used proteomics and genetic manipulations to show that insulin signaling to oocytes regulates offspring’s ability to adapt to future stress via a mechanism that depends on ETC composition in maternal oocytes. Specifically, we found that maternally expressed mutant alleles of nduf-7 (complex I subunit) or isp-1 (complex III subunit) altered offspring’s response to osmotic stress at hatching independently of offspring genotype. Furthermore, we found that expressing wild-type isp-1 in germ cells (oocytes) was sufficient to restore offspring’s normal response to osmotic stress. Chemical mutagenesis screens revealed that maternal ETC composition regulates offspring’s response to stress by altering AMP kinase function in offspring which in turn regulates both ATP and glycerol metabolism in response to continued osmotic stress. To our knowledge, these data are the first to show that proper oocyte ETC composition is required to link a mother’s environment to adaptive changes in offspring metabolism. The data also raise the possibility that the reason diverse animals exhibit insulin regulated remodeling of oocyte mitochondria is to tailor offspring metabolism to best match the environment of their mother.


Results
While diverse animals (including humans) remodel their mitochondrial proteome in oocytes 1,2 , the physiological relevance of this phenomena is unknown.A recent study speculated that ETC remodeling prevents the generation of reactive oxygen species (ROS) that might otherwise cause heritable genetic damage 2 .However, this hypothesis is untested.Studies in flies and frogs have demonstrated that oocyte mitochondrial proteome remodeling is linked to maternal insulin signaling 1 .If mitochondrial proteome remodeling in oocytes exists to prevent ROS-mediated genotoxicity, then it remains unclear why such a process would be linked to an environmentally responsive somatic signaling pathway like insulin signaling and not simply part of normal gametogenesis.Recently, we found that maternal C. elegans exposure to osmotic stress leads to a 50-fold increase in offspring survival in response to future osmotic stress via a mechanism that also depends on insulin signaling to oocytes 3 .Since then, reduced insulin signaling to oocytes was similarly found to promote an offspring adaptation to nutrient stress in fruit flies 4 .Based on these findings across different species and for different environmental stressors, we hypothesized that changes in insulin signaling to oocytes might be an evolutionarily conserved mechanism to promote adaptive changes in offspring metabolism that prepare offspring for their mother's current, external environment.Furthermore, we hypothesized that these effects might be transmitted to offspring by remodeling mitochondrial ETC composition in oocytes.

Maternal environment alters offspring mitochondrial protein abundance
If a mother's environment regulates offspring's ability to adapt to osmotic stress via changes in oocyte mitochondrial protein composition, then we hypothesized that some changes in mitochondrial protein composition might still be detectable in offspring.To test this hypothesis we performed global proteomics profiling on ~50,000 embryos (n = 8 replicates each) collected from parents that were exposed to normal laboratory conditions (50 mM NaCl), and parents exposed to mild osmotic stress (300 mM NaCl).We detected peptides matching 7,161 C. elegans proteins and found that the abundance of 371 proteins was significantly (padj < 0.01) changed in the offspring of parents exposed to 300 mM NaCl (Fig. 1a and Supplementary Table 1).These 371 proteins included glycerol-3-phosphate dehydrogenase GPDH-2 3 , the O-GlcNAc transferase OGT-1 5 , and the late embryo abundant protein LEA-1 6 (Fig. 1a), all of which participate in the osmotic stress response.In addition to these known osmotic stress proteins, we discovered that the majority of regulated proteins were related to animal metabolism and that mitochondria was the most common cellular localization site for differentially abundant proteins (Fig 1b and Supplementary Table 1).Differentially abundant mitochondrial proteins included two subunits of ETC complex II (SDHA-1 and SDHA-2) and two subunits of complex V (ASG-2 and MAI-2; Fig. 1a).Thus, parental exposure to osmotic stress alters the abundance of metabolic and mitochondria-localized proteins in offspring.
When we compared these proteomics data to our previous embryo RNA-seq data (from parents exposed to either 50 mM or 300 mM NaCl 3 ), we found that only 106 of the 371 proteins had a corresponding change in mRNA abundance (Supplementary Fig. 1a).It is possible that the observed discrepancy between proteomic and RNA-seq data is an artifact of rapid mRNA degradation of mRNA contributed from oocytes.Thus, we also manually dissected oocytes from wild-type animals exposed to either 50 mM or 300 mM NaCl and performed oocyte RNA-seq to capture "earlier" mRNA expression patterns (Supplementary Fig. 1b).We identified 169 differentially expressed genes in oocytes from parents exposed to osmotic stress (Supplementary Fig. 1c and Supplementary Table 2).However, only 5 of these 169 differentially expressed genes overlapped with changes in protein abundance in embryos (Supplementary Fig. 1d).Interestingly, gpdh-2 did not exhibit an increase in mRNA abundance in either oocytes or embryos (Supplementary Figure 1) 3 , even though GPDH-2 protein is more abundant in the offspring of parents exposed to 300 mM NaCl (Fig. 1a) and GPDH-2 is functionally required for animals to intergenerationally adapt to osmotic stress 3 .Proteomics therefore identifies functionally relevant changes in offspring protein abundance that cannot be inferred from RNA-seq data alone.

Oocyte ETC composition affects offspring response to stress
Our proteomics data suggested that changes in mitochondrial ETC composition or function might be required for animals to intergenerationally adapt to osmotic stress.To test this hypothesis, we took C. elegans mutants with known point mutations in ETC complex I (gas-1 7 , nuo-6 8 , nduf-7 9 ), complex II (mev-1 10 , sdha-2), complex III (isp-1 11 ), and complex V (asg-2), and assayed whether they could intergenerationally adapt to osmotic stress.As we found previously 3 , >99% of wild-type embryos collected from parents grown under normal laboratory conditions (50 mM NaCl) enter a reversible and stress-resistant state of near suspended animation in response to 500 mM NaCl after hatching (Fig. 1c).This response helps animals survive otherwise lethal exposure to osmotic stress 3 .By contrast, parental exposure to mild osmotic stress (300 mM NaCl) generates offspring that are resistant to osmotic stress, and bypass this state of near suspended animation 3 (Fig. 1c).Point mutations in gas-1(fc21), nuo-6(qm200), nduf-7(et19), mev-1(kn1), and isp-1(qm150) resulted in animals that were unable to adapt to osmotic stress (Fig. 1c).ASG-2 and SDHA-2 proteins were elevated in the offspring of wildtype parents exposed to 300 mM NaCl (Fig. 1a), and we also found that mutations in asg-2(tm1472) and sdha-2(tm1420) reduced animals capacity to adapt to osmotic stress relative to wild-type offspring from parents exposed to 300 mM NaCl (Fig. 1c).(Note, in C. elegans ASG-2 and SDHA-2 are partially redundant with ASG-1 and SDHA-1 respectively).In contrast to mutations in subunits of complexes I, II, III, and IV, we found that mutations in Coenzyme Q synthesis (clk-1(qm30)), mitochondrial fission (drp-1), and mitochondrial fusion (eat-3) genes did not disrupt animal's ability to adapt to osmotic stress, while a mutation in the mitochondrial fusion gene fzo-1 slightly enhanced the adaptive response (Fig. 1c).We conclude that proper function of complexes I, II, III, and V of the mitochondrial oxidative phosphorylation (OxPhos) machinery is required for animals to intergenerationally adapt to osmotic stress.
Previous studies found that animals with ETC mutations (nuo-6 and isp-1) are resistant to and survive exposure to 500 mM NaCl better than wild-type animals 12,13 .However, our findings indicate that the same mutant animals cannot bypass entry into a state of near suspended animation in response to 500 mM NaCl even when their parents were exposed to 300 mM NaCl (Fig. 1c).Since we previously found that stress-induced near suspended animation promotes animal survival at 500 mM NaCl 3 , we hypothesized that these two seemingly contradictory results might be because OxPhos dysfunction favors animal entry into a protective suspended animation in response to osmotic stress.If true, then OxPhos mutants should enter suspended animation at NaCl concentrations that are normally tolerated by wild-type animals, in the same way that 100% of insulin receptor (daf-2) mutant animals enter suspended animation in response to 300 mM NaCl even though 0% of wild-type animals enter this state under the same conditions 3 .To test this, we exposed wild-type, nuo-6, nduf-7, gas-1, mev-1, and isp-1 mutant animals to 300 mM, 350 mM, and 400 mM NaCl, and quantified the number of mobile and developing animals after 48 hours of exposure to osmotic stress.Consistent with our hypothesis, we found that >50% of isp-1(qm150) mutants entered the near suspended animation state at 300 mM NaCl, a condition where 100% of wild-type animals develop normally (Fig 1d).gas-1(fc21), nduf-7(et19), nuo-6(qm200), and mev-1(kn1) mutants were more prone than wild-type animals to enter developmental arrest in response to 350 mM and 400 mM NaCl, but these effects were relatively subtle compared to isp-1(qm150) mutants (Fig. 1e).To rule out that the observed phenotype was due to background mutations in the isp-1(qm150) strain, we used CRISPR/Cas9 to recreate the P225S mutation in isp-1 in wild-type animals, and confirmed that this newly and independently generated mutant behaves the same as isp-1(qm150) (Fig. 1f).To test if ETC mutants cause defects in animals' response to osmotic stress due to a global loss of ATP abundance or due to defects in glycerol metabolism we profiled the abundance of these metabolites in wild-type and isp-1 mutant embryos from parents exposed to 50 mM and 300 mM NaCl (isp-1 mutants were the most sensitive to osmotic stress (Fig. 1d-e)).However, we found that the global ATP/ADP ratio was slightly increased in isp-1 mutant animals and that isp-1 mutants had no changes in either baseline glycerol abundance or the ability to increase glycerol abundance in response to osmotic stress (Supplementary Figure 2).We conclude that ETC function regulates C. elegans decision to enter a protective near suspended animation like state in response to osmotic stress and the effects of ETC dysfunction on animals' response to osmotic stress are not due to a global loss of ATP abundance or due to a defect in glycerol metabolism.
We previously found that insulin signaling to oocytes regulated whether or not C. elegans enter a suspended animation like state in response to osmotic stress 3 .To test if ETC subunits might also function in oocytes to regulate offspring response to osmotic stress we expressed a wild-type copy of isp-1 specifically in germ cells using the pie-1 promoter in isp-1(qm150) mutant animals.
We found that this germ cell specific expression was sufficient to rescue the defects caused by the isp-1(qm150) mutation (Fig. 1f).We conclude that altering germline ETC subunit composition is sufficient to alter offspring response to osmotic stress at hatching.
If mitochondrial function or composition in oocytes regulates offspring response to osmotic stress after hatching, then disruption of maternal, but not paternal, mitochondrial function should disrupt offspring adaptation to osmotic stress.To test this, we used a series of genetic crosses similar to those we previously used to find that the insulin-like receptor DAF-2 functions in oocytes to regulate C. elegans intergenerational adaptation to osmotic stress 3 .Briefly, we previously found that 100% of daf-2(e1370) homozygous mutant animals enter a state of near suspended animation in response to even mild osmotic stress (300 mM NaCl -see Fig. 2a and 3 ).This sensitivity to developmental arrest is due to a loss of insulin signaling to intestinal cells which drives arrest 3 .However, when daf-2 mutant mothers are crossed with wild-type males it results in offspring that are resistant to developmental arrest even at 500 mM NaCl (Fig. 2a    and 3 ).This unusual maternal effect that is distinct from either the maternal or paternal phenotype is because heterozygous daf-2/+ offspring have functional insulin signaling to intestinal cells which reveals a previously masked maternal effect of reduced insulin signaling to oocytes on offspring response to osmotic stress 3 .In other words, the loss of insulin signaling to oocytes triggers the same intergenerational adaptation to osmotic stress as maternal exposure to 300 mM NaCl even when parents are not exposed to any osmotic stress (See Burton et al., 2017 for details 3 ).To test if test if ISP-1 is required maternally to promote offspring adaptation to osmotic stress we first attempted to generate daf-2(e1370); isp-1(qm150) homozygous double mutant animals.However, we found that such double mutant animals were sterile.Nonetheless, we were able construct daf-2(e1370); isp-1(qm150/+) animals that were heterozygous for the isp-1(qm150) mutation.We found that when daf-2(e1370); isp-1(qm150/+) mutants were crossed with wild-type males then 99% of offspring that inherited the isp-1(qm150) mutation from their mothers arrested their development at 500 mM NaCl (Fig. 2a -column 7).This effect was due to the maternal presence of the isp-1(qm150) mutation because genetically identical offspring that inherited the isp-1(qm150) mutation from their fathers showed no defect in adapting to 500 mM NaCl (Fig. 2a -column 6).We conclude that the function of maternally expressed isp-1 regulates offspring response to osmotic stress.Importantly, the offspring enter the state of suspended animation in response to osmotic stress only after hatching.To our knowledge, this is the first example of maternal OxPhos machinery affecting an offspring phenotype that arises after completing embryonic development.
To test if our findings for isp-1(qm150) were unique to ETC complex III or the isp-1(qm150) mutation, we performed analogous crosses and tests with the nduf-7(et19) mutants that were most similar to wild-type animals in their naïve response to osmotic stress (Fig. 1f).In this case, daf-2(e1370); nduf-7(et19) double mutants were homozygous viable.When these double mutant mothers were crossed with wild-type males, none of the offspring bypassed suspended animation in response to 500 mM NaCl (Fig. 2b, column 5).However, genetically identical offspring that inherited the nduf-7 (et19) mutation from the father were able to adapt to 500 mM NaCl (Fig. 2b, column 6).Thus, maternal dysfunction in either ETC complex I (nduf-7) or complex III (isp-1) blocks the intergenerational adaptation to osmotic stress.

ETC mutations alter the abundance of diverse mitochondrial proteins
Disruptions in diverse subunits of complexes I, II, III, and V all blocked animals' ability to intergenerationally adapt to osmotic stress even though these mutations have very different effects on other animal phenotypes like aging.We hypothesized that one possible explanation for this discrepancy is because all of these mutations might themselves alter the abundance of ETC complexes or subunits either directly or indirectly as part of biological mechanisms to compensate for or tolerate mutations in these ETC genes.In this case then any mitochondrial remodeling that occurs to allow toleration of mutations in these diverse ETC subunits might also prevent the normal mitochondrial remodeling in oocytes/embryos that occurs in response to osmotic stress.If true, then we also hypothesized that mutations in these ETC subunits should also result in mitochondrial remodeling whether or not animals are exposed to osmotic stress.
To test this hypotheses, we performed unbiased, global proteomics on wild-type, isp-1(qm150), nuo-6(qm200), and nduf-7(et19) mutant embryos,.We found that 146 proteins exhibited significant changes in abundance in all three mutant backgrounds relative to wild-type embryos (Fig. 2c).Forty-five of these 146 proteins localize to mitochondria, including multiple subunits of complex I.These findings are consistent with our hypothesis that mutations in diverse ETC subunits all result in substantial changes in the abundance of mitochondrial proteins, either indirectly or as part of a mechanism to compensate for abnormal ETC function.Of the 146 proteins that change in abundance in all three ETC mutants, twenty-four were also among the 371 proteins that exhibited altered abundance in the offspring of parents exposed to 300 mM NaCl (Fig. 1a), suggesting that mutations in ETC subunits also affect the abundance of dozens of proteins that are regulated by osmotic stress (Fig. 1a and Supplementary Table 3).Notably, complex 1 mutations (nuo-6, nduf-7) reduced the abundance of most other complex I subunits (including nuo-6 and nduf-7, respectively; Figs.2d, 2e, and Supplementary Table 3), and 80% of the protein changes in nduf-7(et19) mutants were also observed in nuo-6(qm200) mutants.By contrast, the isp-1(qm150) mutation affected 289 proteins that were not altered in either complex I mutant (nuo-6, nduf-7), nor did it alter isp-1 protein levels (Fig. 2f).We conclude that mutations in diverse ETC subunits change the abundance of many (distinct) mitochondrial proteins, including those that are regulated by maternal exposure to 300 mM NaCl.These findings are consistent with a model in which mutations in diverse ETC subunits all block animals ability to intergenerationally adapt to osmotic stress because they have abnormal compositions of ETC subunits that are part of animals normal adaptation to osmotic stress.

ETC function modifies animal stress response via AAK-2
To identify additional genes involved in the intergenerational adaptation to osmotic stress, we performed two independent mutagenesis screens.One screen was performed on wild-type animals to identify mutants that could bypass suspended animation when exposed to 500 mM NaCl.The second screen was performed on isp-1(qm150) animals to identify suppressor mutants that could bypass suspended animation when exposed to 350 mM NaCl.AMP kinase catalytic subunit aak-2 alleles were identified from both screens (Fig. 3a-d), and an independent aak-2(ok529) deletion allele behaved identically to mutations isolated from our screens (Fig. 3c).AAK-2 is canonically activated in vivo by phosphorylating T243 14,15 (equivalent to mammalian AMPK T172 [14][15][16] ) and can be noncanonically activated by oxidation of cysteines by ROS 17,18 .To test if AAK-2 phosphorylation or cysteine oxidation might be required to regulate animal's response to osmotic stress, we obtained an existing T243A allele of aak-2 generated by CRISPR/Cas9 and, separately, generated a C201S allele of aak-2 based on the conservation of C201 (in C. elegans) across species.We found that T243A mutants behaved like aak-2 null mutants, while C201S mutants behaved like weak partial loss-of-function mutants (Fig. 3c).We conclude that the canonical phosphorylation site (T243) for activation of AAK-2 is required for AAK-2 to promote entry into a state of near suspended animation in response to osmotic stress.
To test if aak-2 functions maternally or in offspring, we crossed aak-2(ok529) mutant mothers with wild-type males, and quantified offspring response to 500 mM NaCl.Offspring from this cross entered suspended animation similar to wild-type animals (Fig. 3e).AMP-kinase activity is well-established to be regulated by changes in ETC activity and these findings support a model in which changes in ETC function regulate C. elegans response to osmotic stress by modifying AAK-2 activity in offspring.

AAK-2 regulates ATP and Glycerol metabolism in response to stress
Among its many functions, AMP-kinase activity regulates animal metabolism in response to stress 19 .To better understand how AMP-kinase activity regulates C. elegans intergenerational adaptation to osmotic stress, we performed global metabolomics profiling on wild-type and aak-2(ok529) mutant animals before exposure to osmotic stress, after 24 hours of exposure to osmotic stress (500 mM NaCl), and after 48 hours of exposure to osmotic stress (500 mM NaCl).
Wild-type and aak-2 mutant starved animals were used as controls to differentiate between metabolic shifts resulting from a general lack of food from those that specifically occur in response to osmotic stress.Two significant findings emerged from these comparisons.First, we found that aak-2 was required for maintaining ATP levels in response to both osmotic stress and starvation (Fig. 4a-c).This is consistent with the known function of AMP-kinase signaling.
Second, glycerol levels were approximately 2 to 8-fold higher in aak-2 mutants than in wild-type animals (Fig. 4d-f).These data are consistent with our previous results showing that increased glycerol production in offspring promotes resistance to 500 mM NaCl 3 , and indicate that glycerol abundance is regulated by AAK-2 activity.While it is known that glycerol abundance increases in response to osmotic stress 3,20 , we unexpectedly found that animals in suspended animation metabolized most of their glycerol within 72 hours, and did so in an aak-2 dependent manner (Fig. 4b).These findings suggest that glycerol is more than just an osmolyte that prevents water loss during osmotic stress and is metabolized by animals that have entered a near suspended animation-like state in response to otherwise lethal osmotic conditions.
We previously found that the glycerol-3-phosphate dehydrogenase GPDH-2 is required for animals to adapt to osmotic stress and bypass the near suspended animation like state in response to 500 mM NaCl 3 .To test if aak-2 mutants also require GPDH-2 to bypass suspended animation in response to 500 mM NaCl, we placed wild-type, aak-2, gpdh-2, and aak-2; gpdh-2 double mutant embryos at 500 mM NaCl and quantified the number of animals that were mobile and developing after 48 hours.Consistent with our previous findings and with aak-2 mutants having increased glycerol abundance when compared to wild-type animals, we found that gpdh-2 was required for aak-2 mutants to develop at 500 mM NaCl (Fig. 4g).We conclude that AAK-2 regulates glycerol abundance and the intergenerational adaptation to osmotic stress via a GPDH-2 dependent mechanism.
In conclusion, our results support a model (Fig. 4h) where maternal exposure to mild osmotic stress causes reduced insulin signaling to oocytes 3 .Reduced insulin signaling to oocytes changes the abundance of mitochondrial and ETC proteins in oocytes (as it does in flies and frogs 1 ), which in turn alters offspring AAK-2 activity to promote animals' metabolic adaptation to osmotic stress.To our knowledge, this is the first demonstration that changes in oocyte mitochondria ETC composition can drive long-lasting changes in offspring phenotype, without any changes in offspring genotype.These findings change our understanding of mitochondrial inheritance; they suggest that oocytes inherit more than just mtDNA from the mother, and that changes in oocyte ETC composition can transmit environmental information to offspring that tailors offspring metabolism to best match the environment experienced by their mothers.
Intergenerational and transgenerational responses to environmental stress have been reported in diverse species 3,[21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38] .Mechanistic follow up studies of these findings to date have largely focused on epigenetic or small RNA based mechanisms (reviewed in 39 ).Here, we report a largely novel mechanism regulating an intergenerational adaptation to stress, the transmission of altered mitochondria to offspring via oocytes.This mechanism would neatly explain multiple unanswered questions in the field of intergenerational effects including: Why do some of the most robust known models of intergenerational adaptations to stress (>10-fold increases in offspring survival) only transmit maternally but not paternally 3,21,22,24,40 ?And how can environmental information be maintained through gametogenesis and early embryogenesis when there is extensive epigenetic remodeling 41 ?
The complete extent to which insulin signaling to oocytes and mitochondrial proteome remodeling in oocytes regulates offspring metabolism across species, including in humans, remains unknown.However, the fact that oocyte mitochondrial remodeling in oocytes continues to be identified in all taxa investigated to date, including in humans 2 , suggests that is potentially an evolutionarily ancient phenomena that plays a major role in animal fitness and physiology.
Interestingly, recent studies of D. melanogaster found that reduced insulin signaling to oocytes increases offspring's resistance to nutrient stress via a mechanism that depends on changes in the NAD + /NADH ratio 4 .Coupled with our findings in C. elegans, these two studies indicate that insulin signaling to oocytes regulates two distinct intergenerational responses to stress (osmotic, nutrient) across two diverse taxa (C.elegans and D. melanogaster) 3,4 .While this general mechanism was the same across these two species, we found that the C. elegans adaptation to osmotic stress is mediated by an AMP-kinase dependent mechanism which is normally regulated by changes in the ratio of ATP compared to ADP and AMP.By contrast, the D. melanogaster response to nutrient stress appears to be regulated by a different metabolic ratio, the NAD + /NADH ratio 4 .This difference suggests the exciting possibility that mitochondrial remodeling in oocytes might be a tunable mechanism that responds to many different environmental inputs across species and that it might drive different and tailored metabolic responses in offspring that tune offspring metabolism to maximize fitness in the environment experienced by their mothers -i.e. the environment in which they are about to be born into.
Future studies of this phenomena will be critical in determining if oocyte mitochondrial remodeling is in fact tunable in response to different stimuli and the extent to which it impacts offspring metabolism across species.Nonetheless, because a similar mitochondrial remodeling event occurs in human oocytes 2 , these results might help explain why and individual's risk for metabolic disease (e.g., Type 2 diabetes) has already been epidemiologically linked to a mother's environment 42,43 .
Offspring embryo isolation.Animals grown in normal NGM conditions were collected in M9 and then washed 3x in diH20.Washed animals were then treated with a solution of of 5% sodium hypochlorite and 5 M NaOH,for 5 minutes under agitation.The embryos were then collected via centrifugation and treated with a secondary solution of 10% sodium hypochlorite for less than 1 minute and observed at 4x magnification until remaining detritus was degraded.Eggs were then collected and washed 3x in M9 if the eggs were to be utilized is osmotic stress assays or washed 3x in diH2O if eggs were to be utilized in proteomic or metabolomics analysis.
Osmotic stress assay.Adult animals raised under normal NGM conditions were collected in M9 buffer solution and embryos were isolated utilizing our embryo isolation protocol (see above).Eggs were then transferred to osmotically stressful culture plates at various concentrations (300, 350, and 400mM NaCl) and incubated at room temperature for 48 hours.
Hatched were then quantified and removed utilizing vacuum suction.Arrested animals were then collected using M9 buffer and allowed to recover in normal NGM culture conditions for 24 hours and then quantified.Osmotic stress assays using 500mM NaCl culture conditions were incubated for only 24 hours before quantification and recovery.
Adaptation assay.Adolescent mothers (at the L4 larval stage) were transferred in M9 buffer solution from normal NGM cultures to osmotically stressful 300mM NaCl cultures until animals were gravid (approximately 24 hours).After which time the cultures were collected and eggs were isolated using our egg isolation protocol.Harvested eggs were then transferred to 500mM NaCl cultures.After 48 hours hatched animals were quantified and removed utilizing vacuum suction and arrested embryos were collected in M9 buffer solution and allowed to recover in normal NGM culture conditions for 24 hours after which time recovered animals which regain normal activity were quantified.
Experimental crosses.To generate heterozygous mutant strains genetic crosses were performed.Adolescent mothers (at the L4 larval stage) from the parental strains were mated with males from the desired strains on NGM culture agar plates for 24-48 hours.The parent carrying the mutations were changed from maternal to paternal depending on the experimental design.Note that mitochondrial mutant males do not mate efficiently, and mutant mothers do not generate large brood sizes.To ensure successful mating males carrying fluorescent tags were ensure that progeny was actually produced from a mating pair.
Laid eggs were then individually selected and transferred to 500mM NaCl cultures and allowed to develop at room temperature for 48 hours and the percentage of hatched animals was quantified.Data Independent acquisition analyses were performed on Orbitrap Eclipse coupled to Vanquish Neo system (Thermo Fisher Scientific).The FAIMS Pro source (Thermo Fisher Scientific) was located between the nanoESI source and the mass spectrometer.2 μg of digested peptides were separated on a nano capillary column (20 cm × 75 μm I.D., 365 μm O.D., 1.7 μm C18, CoAnn Technologies, Washington, # HEB07502001718IWF) at 300 nL/min.Mobile phase A consisted of LC/MS grade H2O (LS118-500, Fisher Scientific), mobile phase B consisted of 20% LC/MS grade and H2O and 80% LC/MS grade acetonitrile (LS122500, Fisher Scientific), and both mobile phases contained 0.1% formic acid.The LC gradient was: 1% B to 24% B in 110 min, 85% B in 5 min, and 98% B for 5 min, with a total gradient length of 120 min.For FAIMS, the selected compensation voltage (CV) was applied (−40V, -55V, -70V) throughout the LC-MS/MS runs.Full MS spectra were collected at 120,000 resolution (full width half-maximum; FWHM), and MS2 spectra at 30,000 resolutions (FMWH).Both the standard automatic gain control (AGC) target and the automatic maximum injection time were selected.A precursor range of 380-980 m/z was set for MS2 scans, and an isolation window of 50 m/z was chosen with a 1 m/z overlap for each scan cycle.32% HCD collision energy was used for MS2 fragmentation.To generate a hybrid library for directDIA™ analysis in Spectronaut, pooled samples underwent datadependent acquisition employing 11 distinct FAIMS CV settings ranging from -30 to 80 CV.Full MS spectra were collected at 120,000 resolution (full width half-maximum; FWHM), and MS2 spectra at 30,000 resolutions (FMWH).Both the standard automatic gain control (AGC) target and the automatic maximum injection time were selected.Ions were filtered with charge 2-5.An isolation window of 1.6m/z was used with quadruple isolation mode.
Ions were fragmented using higher-energy collisional dissociation (HCD) with a collision energy of 30%.
DIA data was processed in Spectronaut (version 18, Biognosys, Switzerland) using direct DIA.Data was searched against of Caenorhabditis elegans reference proteome including both UniProt and TrEMBL databases.The manufacturer's default parameters were used.Briefly, trypsin/P was set as digestion enzyme and two missed cleavages were allowed.Cysteine carbamidomethylation was set as fixed modification, and methionine oxidation and protein N-terminus acetylation as variable modifications.Identification was performed using a 1% q-value cutoff on precursor and protein levels.Both peptide precursors and protein false discovery rate (FDR) were controlled at 1%. Ion chromatograms of fragment ions were used for quantification.For each targeted ion, the area under the curve between the XIC peak boundaries was calculated.To enhance proteome coverage, the DDA raw files were utilized in Library Extension Runs to generate a hybrid library.
LIMMA Proteomics analysis.Differential abundance of proteins was performed using LIMMA 44 and eBayes tools.For differential abundance of proteomics based on 8 replicates of samples (Fig. 1a), proteins with 0 variance and proteins with >30% missingness in any experimental group were removed from analysis.We examined the remaining missingness in the dataset and found that proteins that had missing values in any samples had lower abundance values than proteins without any missingness.Since the missing values were not missing at random, we used a left-censored imputation method (impute.QRILC).Data were then log2 transformed to prepare for LIMMA analysis.We next used LIMMA eBayes to analyze the log2 transformed data with custom contrasts.Both LIMMA and empirical Bayes were performed using robust methods to avoid one protein or sample having too much influence over the results.The results were multiple testing corrected using the Benjamini-Hochberg method to maintain a 5% false discovery rate.
For proteomics comparisons based on 3 biological replicates (Fig. 3), only proteins that exhibited zero variance across all samples were also removed from the analysis.The remaining data were then transformed via variance normalization stabilization to prepare for the LIMMA eBayes analysis.We next used LIMMA eBayes to analyze the normalized data with custom contrasts.Both LIMMA and empirical Bayes were performed using robust methods to avoid one protein or sample having too much influence over the results.The results were multiple testing corrected using the Benjamini-Hochberg method to maintain a 5% false discovery rate.
Metabolomics.Metabolites were extracted from embryos using a modified Bligh-Dyer.
Briefly, embryos were diluted to 310µL in ice-cold water, to which 690µL of ice-cold chloroform:methanol (1:1, v/v) was added.The sample was vortexed, sonicated for 5 minutes in a water-bath sonicator, and incubated on wet ice for 40 minutes.Samples were then centrifuged at 14,000xg for 10 min at 4C. 450µL of the upper, aqueous layer was collected and dried in a vacuum evaporator.Of the remaining aqueous layer, 25-50 µL of each sample was collected and pooled to serve as a pooled quality control sample in LCMS and GCMS analysis.Dried extracts were resuspended in 45µL of LCMS grade water containing 0.5ug/mL D5 glutamate (DLM-556, Cambridge Isotopes) as an internal standard.
For measurement of glycerol, samples were dried post LCMS analysis and subjected to derivatization with 30 μL of methoxyamine (11.4 mg/mL) in pyridine and 70 μL of MTBSFA+1%TMCS TBDMS for GCMS analysis on an Agilent 7890GC/5977bMSD as described previously (PMID: 35981545, PMID: 31747582).The oven program was: initial 95°C ramp to 118°C at 40°C/min then hold for 2 min, then ramp to 300°C at 12°C/min, and hold at 300°C for 5 min.A neat standard of glycerol was derivatized and used to determine retention time (10.77min) and fragment ions (170.8,188.7, 289, and 377).170.8m/z was used as the quantifier ion, and others used to qualify the peak.Data were analyzed in Skyline.
Oocyte dissections Animals were raised to adolescence (L4 stage) under normal culture conditions on NGM and then split into separate cultures, control or 300mM NaCl media, until day 1 of adulthood.Animals were then collected in M9 buffer and washed 3x.Single animals were then transferred to a microscope slide and dissected in a 10uL drop of paralytic 0.5 mg/mL tetramisole in M9.The procedure was performed using a dissection microscope at 4x magnification.The first incision was performed using a number 11 surgical scalpel on the paralyzed animal just posterior to the pharynx, exposing the intestine and the anterior gonad arm.The exposed gonad arm was then removed utilizing a 25 G x 5/8 needle and transferred to a 10uL solution of pure M9.The -1,-2, and -3 oocytes were then separated from the gonad arm using a 34 gauge RN needle, resulting the separation of the oocytes connected together with a small amount of gonadal sheath.These cells were then transferred via eyelash pick into 8uL of SMART seq solution and frozen at -80 C.
Oocyte RNA-seq and analysis.RNA was prepped from oocytes using the Direct-zol RNA microprep kit (ZYMO Research) and was sequenced using an Illumina NovaSeq 6000.Reads were aligned to the WBcel235 reference genome and quantified with STARsolo.A gene matrix composed of a row for each gene and column for each sample was built and brought into R using SummarizedExperiment. Genes with fewer than ten reads across all samples were removed.Differential analysis of the quantified genes was performed with DESeq2  2a.LIMMA analysis for global proteomics (Fig. 1a) and DESeq2 analysis for RNA-seq (Supplementary Fig. 1) are described above in separate sections.No statistical method was used to predetermine sample size.The experiments were not randomized.The investigators were not blinded to allocation during experiments and outcome assessment.
Proteomics.C.elegans samples were homogenized on the Bead Ruptor Elite (Cat# 19-042E, Omni International) for 30 s in 4% SDS solution containing 1x HALT Protease (Cat# 78442, Thermo Fisher Scientific).Samples were sonicated and clarified via centrifugation and transferred to a Protein LoBind Eppendorf tube.Proteins were quantified using the Pierce BCA Protein Assay Kit (Cat# 23227, Thermo Fisher Scientific) and 100 µg of protein was aliquoted for digestion.Protein digestion utilized the S-Trap (Cat# CO2-Mini, Protifi) platform to remove any SDS prior to LCMS/MS analysis.Briefly proteins were reduced with Dithiothreitol (DTT) for 20 minutes, alkylated with Iodoacetamide (IAA) for 20 minutes and digested overnight with Trypsin/Lys-C (Cat# V5072, Promega) at a ratio of 50:1 (protein:enzyme (w/w)).Peptides were eluted off the S-Trap Column and dried down in a Genevac SpeedVac.Dried samples were then cleaned up with C-18 reverse phase columns from Harvard Apparatus(Cat# 74-4601) and dried down before resuspension for LC-MS/MS analysis.Dried samples were resuspended in 50 µL 0.1% FA (LS118-1, Fisher Scientific) and diluted with 50 µL of 0.1% TFA (LS119-500, Fisher Scientific).

Figure 1 .
Figure 1.Proper OxPhos function is required for C. elegans to intergenerationally adapt to

45 .
Statistical analysis.ANOVA analysis with post hoc p-value calculations was used for Figs.