Mild impairment of mitochondrial function increases longevity and pathogen resistance through ATFS-1-driven activation of p38-regulated innate immunity

  1. Juliane C. Campos
  2. Ziyun Wu
  3. Paige D. Rudich
  4. Sonja K. Soo
  5. Meeta Mistry
  6. Julio C.B. Ferreira
  7. T. Keith Blackwell
  8. Jeremy M. Van Raamsdonk

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

While mitochondrial function is essential for life in all multicellular organisms, a mild impairment of mitochondrial function can extend longevity. By understanding the molecular mechanisms involved, these pathways might be targeted to promote healthy aging. In studying two long-lived mitochondrial mutants in C. elegans, we found that disrupting subunits of the mitochondrial electron transport chain resulted in upregulation of genes involved in innate immunity, which we found to be dependent on not only the canonical p38-mediated innate immune signaling pathway but also on the mitochondrial unfolded protein response. Both of these pathways are absolutely required for the increased resistance to bacterial pathogens and extended longevity of the long-lived mitochondrial mutants, as is the FOXO transcription factor DAF-16. This work demonstrates that both the p38-mediated innate immune signaling pathway and the mitochondrial unfolded protein response can act on the same innate immunity genes to promote resistance to bacterial pathogens, and that input from the mitochondria can extend longevity by signaling through these two pathways. Combined, this indicates that multiple evolutionarily conserved genetic pathways controlling innate immunity also function to modulate lifespan.

Significance Statement

In this work, we explore the relationship between mitochondrial function, aging and innate immunity. We find that mild impairment of mitochondrial function results in upregulation of genes involved in innate immunity, increased resistance to bacterial pathogens and lifespan extension, all of which are dependent on two evolutionarily conserved signaling pathways. This work demonstrates how changes in functional status of the mitochondria can trigger activation of innate immunity, and that the underlying mechanisms are important for the longevity of the organism. This work advances our understanding of connections between metabolism and immunity. As the pathways studied here are conserved up to mammals, these insights may help us to understand the role of mitochondrial health, innate immunity and lifespan in humans.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    __Reviewer #1 __ __1. One key citation missing from the current manuscript is from Hwang et al. 2014 (PMID 25288734). This study has already described that the isp-1 mutant strain survives longer during P. aeruginosa infection. This citation also describes that the gene expression profile of isp-1 mutants animals includes a considerable number of pathogen-responsive genes that are similarly induced during infection. While the current manuscript does go into the mechanism of this resistance with more detail, they should amend the language to more appropriately reflect previous work, notably the above reference.

    __

    We apologize for the oversight and have added the suggested citation. Hwang et al. show that *isp-1 *worms have increased resistance to bacterial pathogens that is dependent on HIF-1/HIF1 and AAK-2/AMPK. In future work, it will be interesting to examine whether HIF-1 and AAK-2 act in concert with, or independently of, ATFS-1 and the p38-mediated innate immune signaling pathway to mediate pathogen resistance and longevity in *isp-1 *worms. We will add these points to our discussion.

    __2. The authors suggest that ROS activation of the p38 MAPK pathway is likely not the mechanism that explains the resistance of long-lived mitochondrial mutant animals due to their reduced food intake. However, is ROS production nonetheless involved? Does antioxidant treatment suppress the increased resistance during infection of isp-1 and/or nuo-6 mutant animals?

    __

    To address this question, we will treat wild-type, *isp-1 *and *nuo-6 *worms with antioxidant and then measure resistance to bacterial pathogens using the *P. aeruginosa *strain PA14 slow kill assay. For the antioxidant treatment, we will use 10 mM Vitamin C as we have previously shown that this concentration is effective at reducing ROS in *isp-1 *worms to decrease *isp-1 *lifespan (Van Raamsdonk and Hekimi 2012, PNAS). Although antioxidant treatment can have pleiotropic effects, if this decreases survival of bacterial pathogen exposure, it will suggest that the elevated ROS production in *isp-1 *and *nuo-6 *worms may contribute to their enhanced bacterial pathogen resistance.

    __3. (line 278-282): the authors should elaborate on how the p38 MAPK pathway plays a permissive role. It is intriguing that ATFS-1 and ATF-7 are both bZIP transcription factors that could theoretically heterodimerize and that they share common immune gene targets. The authors do indicate that the binding sites for ATFS-1 and ATF-7 are very different and are likely acting distinctly but some speculation would nonetheless strengthen this statement.

    __

    While ATFS-1 and ATF-7 were shown to bind to the promoter regions of the same innate immunity genes, the apparent consensus binding sites are different suggesting that they bind to different regions of the promoter. One way in which the p38 MAPK pathway may be playing a permissive role is that ATF-7 binding and relief from its repressor activity is required for any transcription of p38-mediated innate immunity target genes to occur. This is consistent with our data showing that disruption of *nsy-1, sek-1, pmk-1 *or *atf-7 *decreases the expression of innate immunity genes in wild-type worms. In contrast, it may be that the role of ATFS-1 is for enhanced expression of innate immunity genes such that when ATFS-1 is bound to the promoter region, or perhaps enhancer elements, the baseline expression of innate immunity genes that results from the binding of ATF-7 is increased. This idea is supported by our data showing that disruption of *atfs-1 *does not affect the expression of innate immunity genes in wild-type worms but prevents *nuo-6 *mutants from having increased expression. We will update our manuscript to include these points.

    __4. The authors suggest that reduced food consumption of nuo-6 and isp-1 animals may suppress ROS-induced activation of the p38 innate immune pathway. It is intriguing that dietary restriction was previously shown to increase resistance to infection, presumably through p38-independent mechanisms (PMID 30905669). It would be interesting to measure host survival of nuo-6 and isp-1 mutant animals that are dietary-restricted to see if the enhanced survival rates conferred by mitochondrial stress and DR are additive or not.

    __

    According to this suggestion, we will compare the bacterial pathogen resistance of wild-type, *isp-1 *and *nuo-6 *worms that have undergone dietary restriction to the same strains under ad libitum conditions. This will determine the extent to which their enhancement of pathogen resistance might be additive.

    __5. Figure 2: It is intriguing that loss of p38 signaling appears to have different effects in nuo-6 versus isp-1 animals. Specifically, loss of p38 signaling in isp-1 mutants renders them more sensitive to infection than wild-type, whereas it generally suppresses survival rates back to wild-type levels in the nuo-6 mutant background. Even within the nuo-6 mutant group, loss of SEK-1 has more dramatic effects on nuo-6 mutant animals than does loss of NSY-1, PMK-1 or ATF-7(gf). This is despite the fact that the nsy-1, sek-1, and pmk-1 alleles that are used in this study are all reported to be null. Can the authors speculate on these differences?

    __

    While the *isp-1 *and *nuo-6 *mutations both alter mitochondrial function, they affect different components of the electron transport chain. *isp-1 *mutations affect Complex III (Feng et al. 2001, Dev. Cell), while *nuo-6 *mutations affect Complex I (Yang and Hekimi 2010, Aging Cell). Although these mutants both have increased lifespan and a similar slowing of physiologic rates, it is not uncommon to observe differences between these mutants. For example, while treatment with the antioxidant NAC completely reverts *nuo-6 *lifespan to wild-type, it only partially reduces *isp-1 *lifespan (Yang and Hekimi 2010, PLoS Biology), suggesting that *nuo-6 *lifespan may be more dependent on ROS than *isp-1. *We have recently shown that deletion of *atfs-1 *reduces *nuo-6 *lifespan, but completely prevents *isp-1 *worms from developing to adulthood (Wu et al. 2018, BMC Biology), suggesting that *isp-1 *worms are more dependent on ATFS-1 than *nuo-6 *worms. Disruption of *sek-1 *has a greater impact on pathogen resistance than *nsy-1 *and *pmk-1 *because SEK-1 is absolutely required for innate immune signaling, while some partial redundancy exists for NSY-1 and PMK-1. We will add these points to our manuscript.

    __6. One of the main conclusions from this study is that ATFS-1 likely binds directly to innate immune genes that are in common with ATF-7. Since this is such a pivotal finding, the authors should validate some candidate genes from the referenced ChIP seq datasets using ChIP qPCR. Also, are there predicted ATFS-1 binding sites (PMID 25773600) in these promoters?

    __

    Our data shows that activation of ATFS-1 increases the expression of innate immunity genes without increasing activation of p38. The simplest explanation for this observation is that ATFS-1 can upregulate the same innate immunity genes as ATF-7. Accordingly, we hypothesized that ATFS-1 and ATF-7 can bind to the same promoter. Fortunately, two previous ChIP-Seq studies, from well-established laboratories who have extensive experience studying ATFS-1 and ATF-7, had already determined which genes are bound by these two transcription factors (Nargund et al. 2015, Molecular Cell; Fletcher et al. 2019, PLoS Genetics). Comparing the results of these two published studies confirmed our hypothesis by demonstrating that the same innate immunity genes are bound by both ATF-7 and ATFS-1 in vivo. In order to provide additional support for the conclusion that ATFS-1 and ATF-7 can bind to the same genes, we will examine the genetic sequence of innate immunity genes that were shown to be bound by both ATFS-1 and ATF-7 in the published ChIP-seq studies to identify predicted binding sites for ATFS-1 and ATF-7, while noting that the ATFS-1-associated sequence is an enriched motif and not an established binding site. If we are able to identify the predicted binding sites for these two transcription factors in the same gene, it will provide further support for the conclusion that these transcription factors can both bind to the same innate immunity genes.

    __Reviewer #2:

    (1) The authors state that the p38 MAPK PMK-1 is not activated in the long-lived mitochondrial mutants. However, it might be better to state that there is "no enhanced activation" of PMK-1, since they clearly show in nuo-6 and isp-1 mutants the presence of phosphorylated PMK-1 (Fig. 4A), which would indicate an activated form of PMK-1 in these mutants.__

    According to this suggestion, we will change the text to indicate that there is no enhanced activation of PMK-1 in *nuo-6 *and *isp-1 *worms.

    __(2) Are the food-intake behaviors of all mutants in liquid culture (Fig. 4B-F) the same as their food-intake behaviors on solid agar media, the environment where pathogen resistance was measured?

    __

    We previously compared assays measuring food intake on solid agar media versus the liquid culture approach used in the current study to determine which method is the most robust (Wu et al. 2019, Cell Metabolism). While both assays produced similar results, performing the food intake assay on solid agar plates was much more variable as it is challenging to scrape off all of the uneaten bacteria from solid plates in order to measure it. Since the approach of measuring food intake in liquid media produces more consistent and reliable results, we chose to use this assay for the current study. We will update our manuscript to include this justification.

    (__3) Does the p38 pathway single mutant nsy-1 or sek-1 live shorter than wild type on dead E. coli OP50 (Fig. S9) than they do on live OP50 (Fig. 3)? If so, what might that mean? These mutants are also living shorter than wild type on PA14 (Fig. 2), but live as long as wild type on OP50 (Fig. 3). What is in the live OP50 that allows these mutants to live like wild type?

    __In a previous publication, we found that *sek-1 *mutants live shorter than wild-type worms, and *nsy-1 *live slightly shorter than wild-type worms in a lifespan assay performed in liquid medium with dead OP50 bacteria (Wu et al. 2019, Cell Metabolism). In the current study, we performed lifespan assays on solid NGM plates with live OP50 bacteria and observed a wild-type lifespan in *sek-1 *and *nsy-1 *worms. Since there are multiple experimental variables that are different between the previous and current study, most notably liquid versus solid media, the lifespan results cannot be directly compared. In the case of measuring survival of these strains on PA14, the simplest explanation is that they are dying sooner because their innate immune signaling pathway is disrupted, and so they are less able to mount an immune response against the pathogenic bacteria. We will update our manuscript to include these points.

    __At the same time, wouldn't it be simpler to call the multiple antibiotic-treated OP50 as "dead bacteria", instead of "non-proliferating bacteria"? Some of the antibiotics used to treat OP50 are bactericidal and not bacteriostatic.

    __

    We previously monitored the OD600 of the antibiotic-treated, cold-treated OP50 that we used in our experiment, and found that there is only a very small decrease in OD600 after 10 days (Moroz et al. 2014, Aging Cell). Since dead bacteria are rapidly broken down leading to a decrease in OD600, this result is consistent with the bacteria being alive but not proliferating. We will include this point in our manuscript.

    __(4) Since nuo-6 and isp-1 do not always behave exactly the same in their dependence on certain genes (e.g., Fig. 2C vs Fig 2D), what happens in isp-1; atfs-1 double mutants? Do these mutants behave in the same manner as nuo-6; atfs-1?

    __

    This is an interesting question. Unfortunately, *isp-1;atfs-1 *mutants arrest during development (Wu et al. 2018, BMC Biology), which is why we only examined the effect of *atfs-1 *deletion in *nuo-6 *mutants. We will update the manuscript to note this point.

    __Regarding nuo-6; atfs-1, why does the double mutant live shorter on PA14 than either single mutant (Fig. 6A)? Is this because atfs-1 is needed to activate the p38 MAPK-dependent and -independent pathways? __

    It is possible that the *nuo-6 *mutation makes worms more sensitive to bacterial pathogens, perhaps due to decreased energy production, and that activation of ATFS-1 is required not only to enhance their resistance to pathogens but also to increase their resistance back to wild-type levels. In a previous study, we showed that loss of ATFS-1 slows down the rate of nuclear localization of DAF-16. Thus, loss of *atfs-1 *may also be decreasing resistance to bacterial pathogens by diminishing the general stress resistance imparted by the DAF-16-mediated stress response pathway. We will update the manuscript to include these points.

    __In Fig. 7B, the atfs-1(gof) appears to have slightly more phosphorylated p38 compared to wild type, although it is not statistically significant?

    __

    While there is a trend towards a very modest increase in phosphorylated p38 in the constitutively-active *atfs-1 *mutant compared to wild-type, quantification of four biological replicates indicated that the difference is not significant. This result is consistent with the fact that the levels of phosphorylated p38 are not significantly increased in *nuo-6 *or *isp-1 *mutants, both of which show activation of ATSF-1. We have provided raw images of all of these Western blots in our supplementals. In addition, we will repeat these Western blots to determine if this difference becomes significant with additional replicates.

    __In Fig. 6B, the atfs-1 loss-of-function single mutant also increases the expression of Y9C9A.8, but suppresses it in a nuo-6 mutant background? What might that mean?

    __

    It is possible that in wild-type animals disruption of *atfs-1 *causes a compensatory upregulation of specific stress response genes. We have previously shown that deletion of atfs-1 results in upregulation of chaperone genes involved in the cytoplasmic unfolded protein response (hsp-16.11, hsp-16.2; Wu et al. 2018; BMC Biology). Perhaps Y9C9A.8 is acting in a similar way. In *nuo-6, *the upregulation of Y9C9A.8 is driven by activation of ATFS-1, and thus is prevented by *atfs-1 *deletion. We will add these points to the manuscript.

    __Reviewer #3:

    1. Some studies propose that OP50 offers some toxicity to worms which is not observed in other bacterial strains like HT115. The authors should test the role of the p38-innate immune signaling pathway in nuo-6 and isp-1 lifespan using other non-pathogenic E. coli strains.

    __

    To determine if the effect of disrupting the p38-mediated innate immune signaling pathway on the lifespan of *isp-1 *and *nuo-6 *mutants was simply the result of losing protection against OP50 bacteria, we examined the effect of *nsy-1, sek-1 *and *atf-7(gof) *mutations on *isp-1 *and *nuo-6 *lifespan using non-proliferating bacteria. We found that even when no proliferating bacteria are present, disruption of the p38-mediated innate immune signaling pathway markedly decreases *isp-1 *and *nuo-6 *lifespan. This suggests that the p38-mediated innate immune signaling pathway is required for their long lifespan independently of its ability to protect against bacterial infection. Similarly, we have previously shown that lifespan extension resulting from dietary restriction is dependent on the p38-mediated innate immune signaling pathway even when non-proliferating bacteria are used (Wu et al. 2019, *Cell *Metabolism). We will clarify this important point in the manuscript.

    __

    1. The authors should measure food intake in worms exposed to pathogenic bacteria, given that reduced bacterial intake may be related to reduced mortality.

    __

    Unfortunately, it is not feasible to perform the food intake assay using the pathogenic bacteria because the bacteria cause death thereby complicating the calculation of food consumed per worm (which requires at least 3 days to assess). As an alternative to measuring food intake, we will attempt to measure intestinal accumulation of P. aeruginosa, which is a balance between food intake and other factors. To do this we will use a *P. aeruginosa *strain that expresses GFP and quantify the amount of intestinal fluorescence in wild-type, *isp-1 *and *nuo-6 *worms that have been grown on the GFP-labelled P. aeruginosa.

    __3) The authors should check if ROS is required for the activation of the p38-mediated innate immune signaling pathway and reduction in food intake.

    __

    To determine if the elevated ROS that is present in *isp-1 *and *nuo-6 *worms affects activation of the p38-mediated innate immune signaling pathway, we will treat wild-type, *isp-1 *and *nuo-6 *worms with Vitamin C and measure the ratio of phosphorylated p38 to total p38 by Western blotting. Similarly, to examine the effect of ROS on food intake, we will treat wild-type, *isp-1 *and *nuo-6 *worms with Vitamin C and then quantify its effect on food intake. For these experiments, we will use 10 mM Vitamin C as we have previously shown that this concentration is effective at reducing ROS in *isp-1 *worms to decrease *isp-1 *lifespan (Van Raamsdonk and Hekimi 2012, PNAS).

    __4) Since ATFS-1 and the p38 pathway control food intake, how related to dietary restriction the phenotypes the authors are studying are?

    __

    While the lifespan extension that results from mild impairment of mitochondrial function and the lifespan extension resulting from dietary restriction are both dependent on the p38-mediated innate immune signaling pathway, these interventions modulate innate immunity gene expression in opposite directions. We previously reported that dietary restriction primarily downregulates innate immunity genes (Wu et al. 2019 Cell Metabolism). Here, we show that mutations in *isp-1 *or *nuo-6 *primarily result in upregulation of innate immunity genes. To more globally examine gene expression changes between dietary restriction and mild impairment of mitochondrial function, we compared differentially expressed genes. We found that there was very little overlap of either upregulated or downregulated genes between dietary restriction and *isp-1/nuo-6 *mutants. We will add a supplementary figure to demonstrate this, and add these points to our manuscript.

    __

    1. Somewhat related to the previous points, I am not so sure whether the changes in food intake are cause or consequence of the alterations in the innate immunity-related genes. Reduced food intake is depicted in Fig. 8 as the cause of the activation of the p38 pathway, but there is not enough evidence to unequivocally prove that. In fact, food intake might be controlled by the p38 or ATFS-1 pathway or by a common regulator such as ROS.

    __

    We apologize that we didn’t make this clearer. In our previous work, we showed that dietary restriction results in decreased activation of the p38 pathway (Wu et al. 2019, Cell Metabolism). Here, we show that activation of ATFS-1 results in decreased food intake. Based on our previous study, this decrease in food intake should similarly decrease p38 pathway activation. In Figure 8, we have depicted ATFS-1 inhibiting food intake, and food intake activating the p38-mediated innate immune signaling pathway. Combined, our model suggests that activation of ATFS-1 should act to decrease p38-mediated innate immune signaling. We will clarify this in the figure legend.

    __6) I am not so convinced of the role of DAF-16. In fact, in Fig. 5A daf-16 mutation reduces pathogen resistance and that could represent a toxic effect of the mutation. Furthermore, the results in Fig. 4D do not exclude the possibility that daf-16 and isp-1 act in parallel.

    __

    We agree that the role of DAF-16 could be non-specific. While we show that disruption of daf-16 leads to decreased bacterial pathogen survival in *isp-1 *worms, it also decreases bacterial pathogen survival in wild-type worms. Since DAF-16 is known to be required for general resistance to stress, the decreased survival when *daf-16 *is disrupted could be due to a general toxic effect of reducing general stress resistance. This conclusion is consistent with our observation that DAF-16 is not involved in the upregulation of innate immunity genes in *isp-1 *worms. We will emphasize these points in our manuscript.

    __

    1. Loss of innate immunity related genes may result in toxicity and sensitize worms to pathogenic bacteria. This is further supported by an even lower resistance to pathogens in the double mutants mainly in Fig. 2D.

    __

    We agree. Our data confirms that disruption of the p38-mediated innate immune signaling pathway makes worms more susceptible to bacterial pathogens. We will emphasize this point.

    __

    1. The blots are saturated, particularly in Fig. 4A, and this can be masking the differences in p38 phosphorylation. In fact, the fact that p38 phosphorylation is not changed is contradictory to the other results. How is p38 regulated by mitochondrial mutations then? I am concerned that p38 is actually not altered and the changes in gene expression are exclusively due to ATFS-1. The interaction with the p38 pathway demonstrated genetically could be due to the toxicity elicited by the loss of function mutations in this pathway.__

    To address this concern, we will repeat the Western blotting experiment to compare the ratio of phosphorylated p38 to total p38 between wild-type, *isp-1 *and *nuo-6 *worms. We will take multiple exposures to ensure that the blots are not over-saturated. Having already completed four replicates, we believe that there is not a major change in p38 activation. Our data suggests that the p38-mediated innate immunity pathway is playing a permissive role such that it is required for baseline expression of innate immunity genes, but that activation of ATFS-1 is driving the enhanced expression of innate immunity genes that we observe in the long-lived mitochondrial mutants and constitutively active *atfs-1 *mutants. We will update our manuscript to clarify this.

    __ **Minor concerns**

    1. Lines 167 and 174: What are these p values referred to?

    __

    The p-values indicate the significance of the overlap between the two gene sets. Given the size of the two gene sets, this is the probability that the observed number of overlapping genes would result by picking genes at random. We will clarify this in the manuscript.

    __2) Line 258: I partially agree with the conclusions, since the functions may not necessarily be associated with innate immune signaling but rather other functions of p38.

    __

    Since *isp-1 *and *nuo-6 *worms have extended longevity even when grown on non-proliferating bacteria this indicates that their long life is not dependent on their enhanced resistance to bacterial pathogens. Similarly, since disruption of genes in the p38-mediated innate immune signaling pathway decrease *isp-1 *and *nuo-6 *lifespan even when the worms are grown on non-proliferating bacteria, this suggests that this pathway enhances longevity independently of its ability to increase innate immunity.

    __

    1. Why in figures 4D and E different mutants were used?

    __

    We only used *isp-1 *mutants to examine the effect of *daf-16 *because we were unable to generate *nuo-6;daf-16 *mutants due to close proximity of the two genes on the same chromosome. We only used *nuo-6 *mutants to examine the effect of *atfs-1 *because *isp-1;atfs-1 *worms arrest during development. We will include this explanation in our manuscript.

    __

    1. Line 498: revise writing.

    __

    We will rewrite this sentence to improve clarity.

    __

    1. Show blots in Fig. 7B.

    __

    We will provide an image of a representative Western blot in Figure 7, and will provide the raw images for all of Western blots in our supplementals.

    __

    1. It would be interesting to know where the activation of the immune-related genes by the mitochondrial mutations is happening, whether this is a cell autonomous or cell non-autonomous mechanism.

    __While it would be interesting to explore whether specific tissues are important in sensing mitochondrial impairment in order to upregulate genes involved in innate immunity, it is beyond the scope of this manuscript. Previous work has shown that knocking down the expression of the cytochrome c oxidase gene *cco-1 *in neurons can activate the ATFS-1 target gene *hsp-6 *in the intestine (Durieux et al., 2011). Based on this, one could hypothesize that a similar cell non-autonomous mechanism might be involved. We will note this possible future direction in our discussion.

  5. Review Commons

    Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    ** Reviewer #1**

    1. One key citation missing from the current manuscript is from Hwang et al. 2014 (PMID 25288734). This study has already described that the isp-1 mutant strain survives longer during P. aeruginosa infection. This citation also describes that the gene expression profile of isp-1 mutants animals includes a considerable number of pathogen-responsive genes that are similarly induced during infection. While the current manuscript does go into the mechanism of this resistance with more detail, they should amend the language to more appropriately reflect previous work, notably the above reference.

    We apologize for the oversight and have added the suggested citation. Hwang et al. show that isp-1 worms have increased resistance to bacterial pathogens that is dependent on HIF-1/HIF1 and AAK- 2/AMPK. In future work, it will be interesting to examine whether HIF-1 and AAK-2 act in concert with, or independently of, ATFS-1 and the p38-mediated innate immune signaling pathway to mediate pathogen resistance and longevity in isp-1 worms. We will add these points to our discussion.

    1. The authors suggest that ROS activation of the p38 MAPK pathway is likely not the mechanism that explains the resistance of long-lived mitochondrial mutant animals due to their reduced food intake. However, is ROS production nonetheless involved? Does antioxidant treatment suppress the increased resistance during infection of isp-1 and/or nuo-6 mutant animals?

    To address this question, we will treat wild-type, isp-1 and nuo-6 worms with antioxidant and then measure resistance to bacterial pathogens using the P. aeruginosa strain PA14 slow kill assay. For the antioxidant treatment, we will use 10 mM Vitamin C as we have previously shown that this concentration is effective at reducing ROS in isp-1 worms to decrease isp-1 lifespan (Van Raamsdonk and Hekimi 2012, PNAS). Although antioxidant treatment can have pleiotropic effects, if this decreases survival of bacterial pathogen exposure, it will suggest that the elevated ROS production in isp-1 and

    nuo-6 worms may contribute to their enhanced bacterial pathogen resistance.

    1. (line 278-282): the authors should elaborate on how the p38 MAPK pathway plays a permissive role. It is intriguing that ATFS-1 and ATF-7 are both bZIP transcription factors that could theoretically heterodimerize and that they share common immune gene targets. The authors do indicate that the binding sites for ATFS-1 and ATF-7 are very different and are likely acting distinctly but some speculation would nonetheless strengthen this statement.

    While ATFS-1 and ATF-7 were shown to bind to the promoter regions of the same innate immunity genes, the apparent consensus binding sites are different suggesting that they bind to different regions of the promoter. One way in which the p38 MAPK pathway may be playing a permissive role is that ATF- 7 binding and relief from its repressor activity is required for any transcription of p38-mediated innate immunity target genes to occur. This is consistent with our data showing that disruption of nsy-1, sek-1, pmk-1 or atf-7 decreases the expression of innate immunity genes in wild-type worms. In contrast, it may be that the role of ATFS-1 is for enhanced expression of innate immunity genes such that when ATFS-1 is bound to the promoter region, or perhaps enhancer elements, the baseline expression of innate immunity genes that results from the binding of ATF-7 is increased. This idea is supported by our data showing that disruption of atfs-1 does not affect the expression of innate immunity genes in wild- type worms but prevents nuo-6 mutants from having increased expression. We will update our manuscript to include these points.

    1. The authors suggest that reduced food consumption of nuo-6 and isp-1 animals may suppress ROS- induced activation of the p38 innate immune pathway. It is intriguing that dietary restriction was previously shown to increase resistance to infection, presumably through p38-independent mechanisms (PMID 30905669). It would be interesting to measure host survival of nuo-6 and isp-1 mutant animals that are dietary-restricted to see if the enhanced survival rates conferred by mitochondrial stress and DR are additive or not.

    According to this suggestion, we will compare the bacterial pathogen resistance of wild-type, isp-1 and nuo-6 worms that have undergone dietary restriction to the same strains under ad libitum conditions. This will determine the extent to which their enhancement of pathogen resistance might be additive.

    1. Figure 2: It is intriguing that loss of p38 signaling appears to have different effects in nuo-6 versus isp-1 animals. Specifically, loss of p38 signaling in isp-1 mutants renders them more sensitive to infection than wild-type, whereas it generally suppresses survival rates back to wild-type levels in the nuo-6 mutant background. Even within the nuo-6 mutant group, loss of SEK-1 has more dramatic effects on nuo-6 mutant animals than does loss of NSY-1, PMK-1 or ATF-7(gf). This is despite the fact that the nsy-1, sek-1, and pmk-1 alleles that are used in this study are all reported to be null. Can the authors speculate on these differences?

    While the isp-1 and nuo-6 mutations both alter mitochondrial function, they affect different components of the electron transport chain. isp-1 mutations affect Complex III (Feng et al. 2001, Dev. Cell), while nuo-6 mutations affect Complex I (Yang and Hekimi 2010, Aging Cell). Although these mutants both have increased lifespan and a similar slowing of physiologic rates, it is not uncommon to observe differences between these mutants. For example, while treatment with the antioxidant NAC completely reverts nuo-6 lifespan to wild-type, it only partially reduces isp-1 lifespan (Yang and Hekimi 2010, PLoS Biology), suggesting that nuo-6 lifespan may be more dependent on ROS than isp-1. We have recently shown that deletion of atfs-1 reduces nuo-6 lifespan, but completely prevents isp-1 worms from developing to adulthood (Wu et al. 2018, BMC Biology), suggesting that isp-1 worms are more dependent on ATFS-1 than nuo-6 worms. Disruption of sek-1 has a greater impact on pathogen resistance than nsy-1 and pmk-1 because SEK-1 is absolutely required for innate immune signaling, while some partial redundancy exists for NSY-1 and PMK-1. We will add these points to our manuscript.

    1. One of the main conclusions from this study is that ATFS-1 likely binds directly to innate immune genes that are in common with ATF-7. Since this is such a pivotal finding, the authors should validate some candidate genes from the referenced ChIP seq datasets using ChIP qPCR. Also, are there predicted ATFS-1 binding sites (PMID 25773600) in these promoters?

    Our data shows that activation of ATFS-1 increases the expression of innate immunity genes without increasing activation of p38. The simplest explanation for this observation is that ATFS-1 can upregulate the same innate immunity genes as ATF-7. Accordingly, we hypothesized that ATFS-1 and ATF-7 can bind to the same promoter. Fortunately, two previous ChIP-Seq studies, from well-established laboratories who have extensive experience studying ATFS-1 and ATF-7, had already determined which genes are bound by these two transcription factors (Nargund et al. 2015, Molecular Cell; Fletcher et al. 2019, PLoS Genetics). Comparing the results of these two published studies confirmed our hypothesis by demonstrating that the same innate immunity genes are bound by both ATF-7 and ATFS-1 in vivo. In order to provide additional support for the conclusion that ATFS-1 and ATF-7 can bind to the same genes, we will examine the genetic sequence of innate immunity genes that were shown to be bound by both ATFS-1 and ATF-7 in the published ChIP-seq studies to identify predicted binding sites for ATFS-1 and ATF-7, while noting that the ATFS-1-associated sequence is an enriched motif and not an established binding site. If we are able to identify the predicted binding sites for these two transcription factors in the same gene, it will provide further support for the conclusion that these transcription factors can both bind to the same innate immunity genes.

    Reviewer #2

    1. The authors state that the p38 MAPK PMK-1 is not activated in the long-lived mitochondrial mutants. However, it might be better to state that there is "no enhanced activation" of PMK-1, since they clearly show in nuo-6 and isp-1 mutants the presence of phosphorylated PMK-1 (Fig. 4A), which would indicate an activated form of PMK-1 in these mutants.**

    According to this suggestion, we will change the text to indicate that there is no enhanced activation of PMK-1 in nuo-6 and isp-1 worms.

    1. Are the food-intake behaviors of all mutants in liquid culture (Fig. 4B-F) the same as their food- intake behaviors on solid agar media, the environment where pathogen resistance was measured?

    We previously compared assays measuring food intake on solid agar media versus the liquid culture approach used in the current study to determine which method is the most robust (Wu et al. 2019, Cell Metabolism). While both assays produced similar results, performing the food intake assay on solid agar plates was much more variable as it is challenging to scrape off all of the uneaten bacteria from solid plates in order to measure it. Since the approach of measuring food intake in liquid media produces more consistent and reliable results, we chose to use this assay for the current study. We will update our manuscript to include this justification.

    1. Does the p38 pathway single mutant nsy-1 or sek-1 live shorter than wild type on dead E. coli OP50 (Fig. S9) than they do on live OP50 (Fig. 3)? If so, what might that mean? These mutants are also living shorter than wild type on PA14 (Fig. 2), but live as long as wild type on OP50 (Fig. 3). What is in the live OP50 that allows these mutants to live like wild type?

    In a previous publication, we found that sek-1 mutants live shorter than wild-type worms, and nsy-1 live slightly shorter than wild-type worms in a lifespan assay performed in liquid medium with dead OP50 bacteria (Wu et al. 2019, Cell Metabolism). In the current study, we performed lifespan assays on solid NGM plates with live OP50 bacteria and observed a wild-type lifespan in sek-1 and nsy-1 worms. Since there are multiple experimental variables that are different between the previous and current study, most notably liquid versus solid media, the lifespan results cannot be directly compared. In the case of measuring survival of these strains on PA14, the simplest explanation is that they are dying sooner because their innate immune signaling pathway is disrupted, and so they are less able to mount an immune response against the pathogenic bacteria. We will update our manuscript to include these points.

    At the same time, wouldn't it be simpler to call the multiple antibiotic-treated OP50 as "dead bacteria", instead of "non-proliferating bacteria"? Some of the antibiotics used to treat OP50 are bactericidal and not bacteriostatic.

    We previously monitored the OD600 of the antibiotic-treated, cold-treated OP50 that we used in our experiment, and found that there is only a very small decrease in OD600 after 10 days (Moroz et al. 2014, Aging Cell). Since dead bacteria are rapidly broken down leading to a decrease in OD600, this result is consistent with the bacteria being alive but not proliferating. We will include this point in our manuscript.

    1. Since nuo-6 and isp-1 do not always behave exactly the same in their dependence on certain genes (e.g., Fig. 2C vs Fig 2D), what happens in isp-1; atfs-1 double mutants? Do these mutants behave in the same manner as nuo-6; atfs-1?

    This is an interesting question. Unfortunately, isp-1;atfs-1 mutants arrest during development (Wu et al. 2018, BMC Biology), which is why we only examined the effect of atfs-1 deletion in nuo-6 mutants. We will update the manuscript to note this point.

    Regarding nuo-6; atfs-1, why does the double mutant live shorter on PA14 than either single mutant (Fig. 6A)? Is this because atfs-1 is needed to activate the p38 MAPK-dependent and -independent pathways?

    It is possible that the nuo-6 mutation makes worms more sensitive to bacterial pathogens, perhaps due to decreased energy production, and that activation of ATFS-1 is required not only to enhance their resistance to pathogens but also to increase their resistance back to wild-type levels. In a previous study, we showed that loss of ATFS-1 slows down the rate of nuclear localization of DAF-16. Thus, loss of atfs-1 may also be decreasing resistance to bacterial pathogens by diminishing the general stress resistance imparted by the DAF-16-mediated stress response pathway. We will update the manuscript to include these points.

    In Fig. 7B, the atfs-1(gof) appears to have slightly more phosphorylated p38 compared to wild type, although it is not statistically significant?

    While there is a trend towards a very modest increase in phosphorylated p38 in the constitutively-active atfs-1 mutant compared to wild-type, quantification of four biological replicates indicated that the difference is not significant. This result is consistent with the fact that the levels of phosphorylated p38 are not significantly increased in nuo-6 or isp-1 mutants, both of which show activation of ATSF-1. We have provided raw images of all of these Western blots in our supplementals. In addition, we will repeat these Western blots to determine if this difference becomes significant with additional replicates.

    In Fig. 6B, the atfs-1 loss-of-function single mutant also increases the expression of Y9C9A.8, but suppresses it in a nuo-6 mutant background? What might that mean?

    It is possible that in wild-type animals disruption of atfs-1 causes a compensatory upregulation of specific stress response genes. We have previously shown that deletion of atfs-1 results in upregulation of chaperone genes involved in the cytoplasmic unfolded protein response (hsp-16.11, hsp-16.2; Wu et al. 2018; BMC Biology). Perhaps Y9C9A.8 is acting in a similar way. In nuo-6, the upregulation of Y9C9A.8 is driven by activation of ATFS-1, and thus is prevented by atfs-1 deletion. We will add these points to the manuscript.

    Reviewer #3

    1. Some studies propose that OP50 offers some toxicity to worms which is not observed in other bacterial strains like HT115. The authors should test the role of the p38-innate immune signaling pathway in nuo-6 and isp-1 lifespan using other non-pathogenic E. coli strains.**

    To determine if the effect of disrupting the p38-mediated innate immune signaling pathway on the lifespan of isp-1 and nuo-6 mutants was simply the result of losing protection against OP50 bacteria, we examined the effect of nsy-1, sek-1 and atf-7(gof) mutations on isp-1 and nuo-6 lifespan using non- proliferating bacteria. We found that even when no proliferating bacteria are present, disruption of the p38-mediated innate immune signaling pathway markedly decreases isp-1 and nuo-6 lifespan. This suggests that the p38-mediated innate immune signaling pathway is required for their long lifespan independently of its ability to protect against bacterial infection. Similarly, we have previously shown that lifespan extension resulting from dietary restriction is dependent on the p38-mediated innate immune signaling pathway even when non-proliferating bacteria are used (Wu et al. 2019, Cell Metabolism). We will clarify this important point in the manuscript.

    1. The authors should measure food intake in worms exposed to pathogenic bacteria, given that reduced bacterial intake may be related to reduced mortality.

    Unfortunately, it is not feasible to perform the food intake assay using the pathogenic bacteria because the bacteria cause death thereby complicating the calculation of food consumed per worm (which requires at least 3 days to assess). As an alternative to measuring food intake, we will attempt to measure intestinal accumulation of P. aeruginosa, which is a balance between food intake and other factors. To do this we will use a P. aeruginosa strain that expresses GFP and quantify the amount of intestinal fluorescence in wild-type, isp-1 and nuo-6 worms that have been grown on the GFP-labelled P. aeruginosa.

    1. The authors should check if ROS is required for the activation of the p38-mediated innate immune signaling pathway and reduction in food intake.

    To determine if the elevated ROS that is present in isp-1 and nuo-6 worms affects activation of the p38- mediated innate immune signaling pathway, we will treat wild-type, isp-1 and nuo-6 worms with Vitamin C and measure the ratio of phosphorylated p38 to total p38 by Western blotting. Similarly, to examine the effect of ROS on food intake, we will treat wild-type, isp-1 and nuo-6 worms with Vitamin C and then quantify its effect on food intake. For these experiments, we will use 10 mM Vitamin C as we have previously shown that this concentration is effective at reducing ROS in isp-1 worms to decrease isp-1 lifespan (Van Raamsdonk and Hekimi 2012, PNAS).

    1. Since ATFS-1 and the p38 pathway control food intake, how related to dietary restriction the phenotypes the authors are studying are?

    While the lifespan extension that results from mild impairment of mitochondrial function and the lifespan extension resulting from dietary restriction are both dependent on the p38-mediated innate immune signaling pathway, these interventions modulate innate immunity gene expression in opposite directions. We previously reported that dietary restriction primarily downregulates innate immunity genes (Wu et al. 2019 Cell Metabolism). Here, we show that mutations in isp-1 or nuo-6 primarily result in upregulation of innate immunity genes. To more globally examine gene expression changes between dietary restriction and mild impairment of mitochondrial function, we compared differentially expressed genes. We found that there was very little overlap of either upregulated or downregulated genes between dietary restriction and isp-1/nuo-6 mutants. We will add a supplementary figure to demonstrate this, and add these points to our manuscript.

    1. Somewhat related to the previous points, I am not so sure whether the changes in food intake are cause or consequence of the alterations in the innate immunity-related genes. Reduced food intake is depicted in Fig. 8 as the cause of the activation of the p38 pathway, but there is not enough evidence to unequivocally prove that. In fact, food intake might be controlled by the p38 or ATFS-1 pathway or by a common regulator such as ROS.

    We apologize that we didn’t make this clearer. In our previous work, we showed that dietary restriction results in decreased activation of the p38 pathway (Wu et al. 2019, Cell Metabolism). Here, we show that activation of ATFS-1 results in decreased food intake. Based on our previous study, this decrease in food intake should similarly decrease p38 pathway activation. In Figure 8, we have depicted ATFS-1 inhibiting food intake, and food intake activating the p38-mediated innate immune signaling pathway. Combined, our model suggests that activation of ATFS-1 should act to decrease p38-mediated innate immune signaling. We will clarify this in the figure legend.

    1. I am not so convinced of the role of DAF-16. In fact, in Fig. 5A daf-16 mutation reduces pathogen resistance and that could represent a toxic effect of the mutation. Furthermore, the results in Fig. 4D do not exclude the possibility that daf-16 and isp-1 act in parallel.

    We agree that the role of DAF-16 could be non-specific. While we show that disruption of daf-16 leads to decreased bacterial pathogen survival in isp-1 worms, it also decreases bacterial pathogen survival in wild-type worms. Since DAF-16 is known to be required for general resistance to stress, the decreased survival when daf-16 is disrupted could be due to a general toxic effect of reducing general stress resistance. This conclusion is consistent with our observation that DAF-16 is not involved in the upregulation of innate immunity genes in isp-1 worms. We will emphasize these points in our manuscript.

    1. Loss of innate immunity related genes may result in toxicity and sensitize worms to pathogenic bacteria. This is further supported by an even lower resistance to pathogens in the double mutants mainly in Fig. 2D.

    We agree. Our data confirms that disruption of the p38-mediated innate immune signaling pathway makes worms more susceptible to bacterial pathogens. We will emphasize this point.

    1. The blots are saturated, particularly in Fig. 4A, and this can be masking the differences in p38 phosphorylation. In fact, the fact that p38 phosphorylation is not changed is contradictory to the other results. How is p38 regulated by mitochondrial mutations then? I am concerned that p38 is actually not altered and the changes in gene expression are exclusively due to ATFS-1. The interaction with the p38 pathway demonstrated genetically could be due to the toxicity elicited by the loss of function mutations in this pathway.

    To address this concern, we will repeat the Western blotting experiment to compare the ratio of phosphorylated p38 to total p38 between wild-type, isp-1 and nuo-6 worms. We will take multiple exposures to ensure that the blots are not over-saturated. Having already completed four replicates, we believe that there is not a major change in p38 activation. Our data suggests that the p38-mediated innate immunity pathway is playing a permissive role such that it is required for baseline expression of innate immunity genes, but that activation of ATFS-1 is driving the enhanced expression of innate immunity genes that we observe in the long-lived mitochondrial mutants and constitutively active atfs-1 mutants. We will update our manuscript to clarify this.

    Minor concerns

    1. Lines 167 and 174: What are these p values referred to?**

    The p-values indicate the significance of the overlap between the two gene sets. Given the size of the two gene sets, this is the probability that the observed number of overlapping genes would result by picking genes at random. We will clarify this in the manuscript.

    1. Line 258: I partially agree with the conclusions, since the functions may not necessarily be associated with innate immune signaling but rather other functions of p38.

    Since isp-1 and nuo-6 worms have extended longevity even when grown on non-proliferating bacteria this indicates that their long life is not dependent on their enhanced resistance to bacterial pathogens. Similarly, since disruption of genes in the p38-mediated innate immune signaling pathway decrease isp- 1 and nuo-6 lifespan even when the worms are grown on non-proliferating bacteria, this suggests that this pathway enhances longevity independently of its ability to increase innate immunity.

    1. Why in figures 4D and E different mutants were used?

    We only used isp-1 mutants to examine the effect of daf-16 because we were unable to generate nuo- 6;daf-16 mutants due to close proximity of the two genes on the same chromosome. We only used nuo- 6 mutants to examine the effect of atfs-1 because isp-1;atfs-1 worms arrest during development. We will include this explanation in our manuscript.

    1. Line 498: revise writing.

    We will rewrite this sentence to improve clarity.

    1. Show blots in Fig. 7B.

    We will provide an image of a representative Western blot in Figure 7, and will provide the raw images for all of Western blots in our supplementals.

    1. It would be interesting to know where the activation of the immune-related genes by the mitochondrial mutations is happening, whether this is a cell autonomous or cell non-autonomous mechanism.

    While it would be interesting to explore whether specific tissues are important in sensing mitochondrial impairment in order to upregulate genes involved in innate immunity, it is beyond the scope of this manuscript. Previous work has shown that knocking down the expression of the cytochrome c oxidase gene cco-1 in neurons can activate the ATFS-1 target gene hsp-6 in the intestine (Durieux et al., 2011). Based on this, one could hypothesize that a similar cell non-autonomous mechanism might be involved. We will note this possible future direction in our discussion.

  6. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Campos et al provide evidence that mild mitochondrial dysfunction in C. elegans induces genes involved in innate immunity and promotes bacterial pathogen resistance and longevity, while inhibits food intake through an ATFS-1-mediated mechanism. The manuscript is well-written and the experiments are well-performed and reported. However, there are several points that need to be addressed before the manuscript can be published.

    Major concerns

    1. Some studies propose that OP50 offers some toxicity to worms which is not observed in other bacterial strains like HT115. The authors should test the role of the p38-innate immune signaling pathway in nuo-6 and isp-1 lifespan using other non-pathogenic E. coli strains.
    2. The authors should measure food intake in worms exposed to pathogenic bacteria, given that reduced bacterial intake may be related to reduced mortality.
    3. The authors should check if ROS is required for the activation of the p38-mediated innate immune signaling pathway and reduction in food intake.
    4. Since ATFS-1 and the p38 pathway control food intake, how related to dietary restriction the phenotypes the authors are studying are?
    5. Somewhat related to the previous points, I am not so sure whether the changes in food intake are cause or consequence of the alterations in the innate immunity-related genes. Reduced food intake is depicted in Fig. 8 as the cause of the activation of the p38 pathway, but there is not enough evidence to unequivocally prove that. In fact, food intake might be controlled by the p38 or ATFS-1 pathway or by a common regulator such as ROS.
    6. I am not so convinced of the role of DAF-16. In fact, in Fig. 5A daf-16 mutation reduces pathogen resistance and that could represent a toxic effect of the mutation. Furthermore, the results in Fig. 4D do not exclude the possibility that daf-16 and isp-1 act in parallel.
    7. Loss of innate immunity related genes may result in toxicity and sensitize worms to pathogenic bacteria. This is further supported by an even lower resistance to pathogens in the double mutants mainly in Fig. 2D.
    8. The blots are saturated, particularly in Fig. 4A, and this can be masking the differences in p38 phosphorylation. In fact, the fact that p38 phosphorylation is not changed is contradictory to the other results. How is p38 regulated by mitochondrial mutations then? I am concerned that p38 is actually not altered and the changes in gene expression are exclusively due to ATFS-1. The interaction with the p38 pathway demonstrated genetically could be due to the toxicity elicited by the loss of function mutations in this pathway.

    Minor concerns

    1. Lines 167 and 174: What are these p values referred to?
    2. Line 258: I partially agree with the conclusions, since the functions may not necessarily be associated with innate immune signaling but rather other functions of p38.
    3. Why in figures 4D and E different mutants were used?
    4. Line 498: revise writing.
    5. Show blots in Fig. 7B.
    6. It would be interesting to know where the activation of the immune-related genes by the mitochondrial mutations is happening, whether this is a cell autonomous or cell non-autonomous mechanism.

    Significance

    This study provides significant advance in mechanistic aspects of lifespan regulation in worms, linking mitochondrial metabolism, food intake, innate immunity, resistance to pathogen infections and longevity. The work presents novel mechanistic insights that could be applied to understand how mild mitochondrial dysfunction leads to increased lifespan. Overall, the audience interested in this study are expected to be aging biologists and possibly immunologists with particular interest in mechanistic aspects of longevity and innate immunity, as well as C. elegans as a model organism. I am part of this group of scientists with particular interest in studying the interplay between metabolism and aging.

  7. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary:

    Campos et al. show that mild mitochondrial impairment promotes C. elegans resistance against the bacterial pathogen Pseudomonas aeruginosa PA14, which is associated with increased expression of a subset of innate immunity genes in the animal. Interestingly, upregulation of the innate immunity genes in the mitochondrial electron transport chain mutants, nuo-6 (complex I) and isp-1 (complex III), does not appear to involve enhanced activation of the p38 MAPK PMK-1, which has been previously implicated in anti-bacterial immunity (Jeong et al, EMBO J 2017, 36, 1046). Because the authors also show that this increased pathogen resistance and expression of innate immunity genes in at least one of the mitochondrial mutants (nuo-6) only partly depend on the p38 PMK-1 pathway, this would argue for the involvement of another pathway. The authors show that this other pathway involves the mitochondrial unfolded protein response (mitoUPR) through activation of the transcription factor atfs-1, which not only upregulates a subset of innate immunity genes, but also presumably decreases pathogen intake. Together their data suggest that the p38 PMK-1 pathway and mitoUPR act in parallel to promote the enhanced pathogen resistance of mitochondrial mutants.

    Moreover, while they show that the FOXO transcription factor daf-16 is also required for the enhanced pathogen resistance of mitochondrial mutants (i.e,, isp-1), they rule out daf-16 involvement in the activation of innate immunity genes. Instead, daf-16 decreases pathogen intake and upregulates other stress-response genes. Thus, this study highlights the requirement for multiple pathways to promote pathogen resistance through multiple mechanisms.

    Major comments:

    (1) The authors state that the p38 MAPK PMK-1 is not activated in the long-lived mitochondrial mutants. However, it might be better to state that there is "no enhanced activation" of PMK-1, since they clearly show in nuo-6 and isp-1 mutants the presence of phosphorylated PMK-1 (Fig. 4A), which would indicate an activated form of PMK-1 in these mutants.

    (2) Are the food-intake behaviors of all mutants in liquid culture (Fig. 4B-F) the same as their food-intake behaviors on solid agar media, the environment where pathogen resistance was measured?

    (3) Does the p38 pathway single mutant nsy-1 or sek-1 live shorter than wild type on dead E. coli OP50 (Fig. S9) than they do on live OP50 (Fig. 3)? If so, what might that mean? These mutants are also living shorter than wild type on PA14 (Fig. 2), but live as long as wild type on OP50 (Fig. 3). What is in the live OP50 that allows these mutants to live like wild type?

    At the same time, wouldn't it be simpler to call the multiple antibiotic-treated OP50 as "dead bacteria", instead of "non-proliferating bacteria"? Some of the antibiotics used to treat OP50 are bactericidal and not bacteriostatic.

    (4) Since nuo-6 and isp-1 do not always behave exactly the same in their dependence on certain genes (e.g., Fig. 2C vs Fig 2D), what happens in isp-1; atfs-1 double mutants? Do these mutants behave in the same manner as nuo-6; atfs-1?

    Regarding nuo-6; atfs-1, why does the double mutant live shorter on PA14 than either single mutant (Fig. 6A)? Is this because atfs-1 is needed to activate the p38 MAPK-dependent and -independent pathways? In Fig. 7B, the atfs-1(gof) appears to have slightly more phosphorylated p38 compared to wild type, although it is not statistically significant?

    In Fig. 6B, the atfs-1 loss-of-function single mutant also increases the expression of Y9C9A.8, but suppresses it in a nuo-6 mutant background? What might that mean?

    Some of my comments can be easily addressed with written comments. Others might require generation of a strain, like the isp-1; atfs-1 double mutant, prior to any assays.

    Significance

    Please see the above summary for the significance of this manuscript to the field. Importantly, this study highlights the requirement for multiple pathways to promote pathogen resistance through multiple mechanisms. Readers interested in aging, mitochondrial function, innate immunity and stress responses should find this study thought-provoking. I include myself in this group of readers, since I study the genetics of C. elegans aging and stress responses.

  8. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    The manuscript by Campos et al. describe the association between long-lived mitochondrial mutants and increased resistance to pathogen infection. The authors discover that mitochondrial electron transport chain mutants (nuo-6 and isp-1) display increased expression of many genes involved in innate immunity that are regulated by the p38 signaling pathway. Consistent with this finding, mito mutants displayed increased survival during infection. p38 signaling was found to be required for these innate immune gene inductions during mitochondrial stress and for their increased survival during infection. P38 signaling was also found to be required for the increased lifespan of isp-1 and nuo-6 mutant animals. Intriguingly, p38 signaling does not appear to be affected in these mitochondrial mutants, despite being required for the increase in immunity/host resistance. The authors discover that mitochondrial stress animals exhibit reduced feeding which they argue may suppress any activation of the p38 pathway caused by ROS. The mitochondrial UPR was also found to be required for the increase in innate immune gene expression in isp-1 and nuo-6 mutant animals, as well as their extended survival. The authors conclude that ATFS-1 can act in parallel to p38 signaling by directly binding to common innate immune target genes. In support of this, ATFS-1 and ATF-7 appear to bind to shared target genes but likely at independent sites due to their different consensus sequences.

    1. One general consideration is that some of the key concepts outlined in this manuscript have already been described previously and are therefore not entirely novel conceptually. For example, one key citation missing from the current manuscript is from Hwang et al. 2014 (PMID 25288734). This study has already described that the isp-1 mutant strain survives longer during P. aeruginosa infection. This citation also describes that the gene expression profile of isp-1 mutants animals includes a considerable number of pathogen-responsive genes that are similarly induced during infection. While the current manuscript does go into the mechanism of this resistance with more detail, they should amend the language to more appropriately reflect previous work, notably the above reference.
    2. The authors suggest that ROS activation of the p38 MAPK pathway is likely not the mechanism that explains the resistance of long-lived mitochondrial mutant animals due to their reduced food intake. However, is ROS production nonetheless involved? Does antioxidant treatment suppress the increased resistance during infection of isp-1 and/or nuo-6 mutant animals?
    3. (line 278-282): the authors should elaborate on how the p38 MAPK pathway plays a permissive role. It is intriguing that ATFS-1 and ATF-7 are both bZIP transcription factors that could theoretically heterodimerize and that they share common immune gene targets. The authors do indicate that the binding sites for ATFS-1 and ATF-7 are very different and are likely acting distinctly but some speculation would nonetheless strengthen this statement.
    4. The authors suggest that reduced food consumption of nuo-6 and isp-1 animals may suppress ROS-induced activation of the p38 innate immune pathway. It is intriguing that dietary restriction was previously shown to increase resistance to infection, presumably through p38-independent mechanisms (PMID 30905669). It would be interesting to measure host survival of nuo-6 and isp-1 mutant animals that are dietary-restricted to see if the enhanced survival rates conferred by mitochondrial stress and DR are additive or not.
    5. Figure 2: It is intriguing that loss of p38 signaling appears to have different effects in nuo-6 versus isp-1 animals. Specifically, loss of p38 signaling in isp-1 mutants renders them more sensitive to infection than wild-type, whereas it generally suppresses survival rates back to wild-type levels in the nuo-6 mutant background. Even within the nuo-6 mutant group, loss of SEK-1 has more dramatic effects on nuo-6 mutant animals than does loss of NSY-1, PMK-1 or ATF-7(gf). This is despite the fact that the nsy-1, sek-1, and pmk-1 alleles that are used in this study are all reported to be null. Can the authors speculate on these differences?
    6. One of the main conclusions from this study is that ATFS-1 likely binds directly to innate immune genes that are in common with ATF-7. Since this is such a pivotal finding, the authors should validate some candidate genes from the referenced ChIP seq datasets using ChIP qPCR. Also, are there predicted ATFS-1 binding sites (PMID 25773600) in these promoters?

    Significance

    As mentioned in my comments, some of the findings of the current manuscript have been shown before. Nonetheless, the authors do describe new insights into the mechanism of how mitochondrial stress signaling promotes host resistance to infection, which is noteworthy.

    This manuscript would be of value to researchers in the fields of mitochondrial biology, mitochondrial stress signaling (including the UPRmt field), host-pathogen interactions, and longevity determination.

    My expertise is in stress signaling in the context of longevity and host-pathogen interactions.

  9. Version published to 10.1101/2021.03.31.437812v1 on bioRxiv