An investigation into cellular stress responses reveals how cell compartments called mitochondria use information about the surrounding metabolites and microorganisms to protect themselves from damage. See Letter p.406
Cellular organelles are simultaneously distinct from the rest of the cell and completely reliant on it for their identity and function. A poignant example of this nebulous individualism comes from the mitochondrion, a subcellular entity that began long ago as an independent organism but which, over millennia, has become increasingly dependent on the rest of the cell, and now serves as its energy-generating centre. Interactions between the mitochondrion and the cell provide this organelle with a direct connection to changes in the contents of its surroundings, allowing it to initiate defence mechanisms when things go awry. In this issue, Liu et al.1 (page 406) report the effect of cellular and microbial metabolites on the initiation of protective reactions in mitochondria.
When changes occur that are stressful to mitochondria (for example, changes in their ability to import the molecules required for normal function), they activate a protective program called the mitochondrial unfolded protein response2 (UPRmt). This defence mechanism is designed to maintain normal mitochondrial function, and its activation is typically symptomatic of an imbalance in the cell. A diverse set of stimuli can activate the UPRmt, including disruption of the proteins, import machineries and protease enzymes that help the mitochondrion to function3. Initiation of this defence begins with transportation of the transcription factor ATFS-1 — considered to be the major factor in mitochondrial protection against stress4 — to the nucleus, where it upregulates expression of other genes involved in the response. By contrast, under non-stressed conditions, ATFS-1 is imported into mitochondria, where it is readily degraded by proteases4.
Liu et al. report that when they induced the UPRmt in the nematode worm Caenorhabditis elegans, by genetically or pharmacologically inactivating genes involved in normal mitochondrial function, the worms displayed an aversion to bacteria — their typical source of food. They hypothesized that UPRmt induction might cause such a change in behaviour because the cell interprets the program's activation as symptomatic of a pathogenic attack or as a sign of poor nutrient availability (Fig. 1). If so, this could imply that variations in metabolite levels, which might be caused by the presence of pathogens or a lack of nutrients, play an integral signalling part in the activation of mitochondrial stress responses. These hypotheses are supported by previous observations suggesting a variety of methods by which bacteria are linked to and influence metabolism5 and mitochondrial stress6 in C. elegans.
Next, Liu and colleagues set out to define the intracellular metabolic pathways that are required for UPRmt induction. Using genome-wide screening, the authors identified 45 genes that are needed for the mitochondrial stress response. In this analysis, they found two metabolic pathways required for activation of the UPRmt.
The first, the sphingolipid biosynthesis pathway, is responsible for production of ceramide, a waxy lipid molecule typically found in the cell membrane after being synthesized in another organelle, the endoplasmic reticulum (ER). Ceramide plays a part in multiple physiological processes, including mitochondrial degradation7 and the cell-death program apoptosis8. Although the authors report that ceramide synthesis is required for UPRmt induction, a loss of ceramide synthesis has also been linked to increased longevity in several organisms9,10,11, suggesting a complex role for this molecule in health.
Disruption of a second metabolic pathway, the mevalonate pathway, also blocks activation of the UPRmt. This pathway is required for the synthesis of cholesterol in most multicellular animals, but not in C. elegans12, indicating that alternative metabolic outputs of this pathway (rather than cholesterol synthesis) are required for the mitochondrial stress response in this worm.
The mevalonate pathway is inhibited by statins, drugs that are used in humans to lower cholesterol. Previous work has demonstrated that statins have adverse effects on longevity in C. elegans, affecting the function of the ER and inducing a robust ER stress response13, which acts to protect the ER in the same way as the UPRmt protects mitochondria. Likewise, Liu and co-workers demonstrated that statins affect mitochondrial health. However, in contrast to what is seen in the ER, they found that after treatment with statins the worms could no longer respond to mitochondrial stress. In agreement with this observation, earlier data showed that forcing ATFS-1 to enter the nucleus protects worms from the harmful effects of statins14, suggesting that a breakdown in mitochondrial stress sensing is responsible for the harmful effects of statins in C. elegans. This may have consequences for our understanding of the side effects of statins in humans.
As with statins, Liu et al. found that forcing ATFS-1 to move to the nucleus was sufficient to induce the UPRmt when the sphingolipid biosynthesis pathway was blocked. Clearly, the typical regulation of ATFS-1 is broken after disruption of these two metabolic pathways. But what exactly prevents ATFS-1 from functioning properly?
The authors' data suggest that reducing mitochondrial import of ATFS-1 may not be sufficient to ensure its transportation into the nucleus. Alternatively, mitochondrial import and degradation of ATFS-1 may be directly controlled by metabolite changes in the cell. A closer look at the effect of these metabolites on ATFS-1 import and stability will offer a better understanding of how they affect the function of the UPRmt. Moreover, these results suggest an effect of UPRmt activation, and potentially of ATFS-1, on C. elegans behaviour in response to food sources. How this plays into the general sensing of food availability — for instance, through control of the nervous system15 — is not yet known.
The extent to which this is a general phenomenon or one specific to these particular metabolites also remains unclear. The authors find that a remarkably high percentage of bacterial species in a large panel can induce the UPRmt in C. elegans in the absence of other stressors. By contrast, they find that some classes of microbe are capable of blocking UPRmt activation even during stress. The authors leave the mechanism behind these varying effects open, but the overarching message of this study is clear: different metabolites produced in these differing conditions are most likely to be responsible.
Mitochondria are ancient relics of a purely single-celled world. It is possible that the systems that protect mitochondria were the same as those used to protect colonies of bacteria from their invading neighbours. Perhaps bacterial metabolites have retained the ability to communicate with their long-lost relative, the mitochondrion, which in turn has gained the ability to communicate with its host to alter such complex processes as behaviour. Human physiology also relies on complex interactions with thousands of species of bacterium. It is possible that our own mitochondria will sense and respond to the secondary metabolites produced by these species, and possibly change the behaviour of our cells.
Liu, Y., Buck, S. S., Breen, P. C. & Ruvkun, G. Nature 508, 406–410 (2014).
Zhao, Q. et al. EMBO J. 21, 4411–4419 (2002).
Haynes, C. M., Petrova, K., Benedetti, C., Yang, Y. & Ron, D. Dev. Cell 13, 467–480 (2007).
Nargund, A. M., Pellegrino, M. W., Fiorese, C. J., Baker, B. M. & Haynes, C. M. Science 337, 587–590 (2012).
Virk, B. et al. BMC Biol. 10, 67 (2012).
Pang, S. & Curran, S. P. Cell Metab. 19, 221–231 (2014).
Sentelle, R. D. et al. Nature Chem. Biol. 8, 831–838 (2012).
Stiban, J., Caputo, L. & Colombini, M. J. Lipid Res. 49, 625–634 (2008).
D'Mello, N. P. et al. J. Biol. Chem. 269, 15451–15459 (1994).
Huang, X., Liu, J. & Dickson, R. C. PLoS Genet. 8, e1002493 (2012).
Mosbech, M. B. et al. PLoS ONE 8, e70087 (2013).
Kurzchalia, T. V. & Ward, S. Nature Cell Biol. 5, 684–688 (2003).
Morck, C. et al. Proc. Natl Acad. Sci. USA 106, 18285–18290 (2009).
Rauthan, M., Ranji, P., Pradenas, N. A., Pitot, C. & Pilon, M. Proc. Natl Acad. Sci. USA 110, 5981–5986 (2013).
Schwartz, M. W., Woods, S. C., Porte, D. Jr, Seeley, R. J. & Baskin, D. G. Nature 404, 661–671 (2000).
About this article
Cite this article
Wolff, S., Dillin, A. The stressful influence of microbes. Nature 508, 328–329 (2014). https://doi.org/10.1038/nature13220
PLOS ONE (2017)
Pathologie Biologie (2015)