RESEARCH NEWS & VIEWS called xenophagy. Consequently, autophagy is involved in several diseases 6, but the ATG16L1T300A variant is associated only with Crohn’s disease. Caspases are endoproteases that hydrolyse peptide bonds at specific sequences7. Depending on the target protein, this can lead to the protein’s destruction or to the generation of an active protein. Active caspase 3 is required for the cellular signalling pathway that leads to apoptotic cell death; this pathway is initiated by the activity of other caspases that respond to signals from cell-surface death receptors or irreparable organelle dysfunction. Caspase-3 activity during apoptosis irreversibly sets in motion a sequence of events that leads to the demise of the cell. By contrast, low-level caspase-3 activity that is insufficient to trigger apoptosis has homeostatic and protective functions, including guarding stressed organs against cell death8. Through clever alignment of the ATG16L1 protein sequence from several species, Murthy et al. predicted and then directly demonstrated that the T300A variant protein (or the equivalent variant in mice, T316A) was highly sensitive to cleavage by caspase 3 (Fig. 1). The authors also show that, in human cells carrying the ATG16L1T300A risk variant, signalling initiated by binding of the protein tumour-necrosis factor-α (TNF-α) to its cell-surface receptor (a death receptor), cellular stress caused by starvation, or infection with the pathogenic gut bacterium Yersinia enterocolitica all resulted in caspase-3-dependent degradation of ATG16L1T300A, and consequently in impaired autophagy and xenophagy responses to these stresses. In the case of Y. enterocolitica infection, there was also an increased production of inflammatory cytokine proteins, such as interleukin-1β and TNF-α, in mice engineered to express the ATG16L1T316A variant. In stark contrast, the authors show that autophagosome formation was not impaired by the presence of ATG16L1 T300A when autophagy was directly induced without caspase-3 activation. These findings suggest that the ATG16L1T300A risk variant, through its sensitivity to caspase-3-mediated cleavage, disables a host’s ability to properly respond to environmental challenges that require a compensatory autophagy response. Such challenges may include infections, inflammatory stimuli and metabolic disturbances, all of which induce autophagy to remove pathogens or inflammatory organelles (such as inflammasomes5) and to provide nutrients to overcome and survive cellular stress. Caspase 3 is commonly activated during these stress conditions. The cellular state known as endoplasmic reticulum (ER) stress can also activate both caspase 3 and autophagy, and is commonly observed in intestinal epithelial cells of individuals with Crohn’s disease9. Previous studies in mice 10 showed that loss of
compensatory autophagy in ER-stressed intestinal epithelium, owing to deletion of the Atg16l1 gene, results in Crohn’s-diseaselike inflammation of the small intestine. The identification of the sensitivity of this protein variant to caspase-3-mediated destruction may provide insight into how specific environmental exposures convert genetic disease risk into clinical symptoms. Murthy and colleagues’ study also helps to integrate other observations from work on Crohn’s disease. For example, caspase-3 activation can be regulated by inhibitor-of-apoptosis proteins such as XIAP, variants of which are the cause of a single-gene form of Crohn’s disease11. XIAP, in turn, directly interacts with other proteins involved in microbial sensing that are associated with genetic risk for Crohn’s disease4,11, including RIPK2 and NOD2. The new data might also help to explain why agents targeting TNF-α are highly effective therapies for Crohn’s disease2 — in light of the authors’ Y. enterocolitica infection studies, it seems plausible that TNF-α fuels an inflammatory feed-forward loop by inducing caspase-3mediated degradation of ATG16L1T300A, which in turn results in increased TNF-α secretion. Although further studies are necessary to test the authors’ biochemical and cellular observations in animal models and patients with
Crohn’s disease, their observations provide an appealing and integrating hypothesis for how this common genetic element engenders disease risk. ■ Arthur Kaser is in the Division of Gastroenterology and Hepatology, Department of Medicine, Addenbrooke’s Hospital, University of Cambridge, Cambridge CB2 0QQ, UK. Richard S. Blumberg is in the Division of Gastroenterology, Hepatology and Endoscopy, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115, USA. e-mails: [email protected]; [email protected] Hampe, J. et al. Nature Genet. 39, 207–211 (2007). Maloy, K. J. & Powrie, F. Nature 474, 298–306 (2011). Murthy, A. et al. Nature 506, 456–462 (2014). Jostins, L. et al. Nature 491, 119–124 (2012). Saitoh, T. et al. Nature 456, 264–268 (2008). Levine, B., Mizushima, N. & Virgin, H. W. Nature 469, 323–335 (2011). 7. McIlwain, D. R., Berger, T. & Mak, T. W. Cold Spring Harb. Perspect. Biol. 5, a008656 (2013). 8. Khalil, H. et al. Mol. Cell. Biol. 32, 4523–4533 (2012). 9. Deuring, J. J. et al. Gut http://dx.doi.org/10.1136/ gutjnl-2012-303527 (2013). 10. Adolph, T. E. et al. Nature 503, 272–276 (2013). 11. Uhlig, H. H. Gut 62, 1795–1805 (2013). 1. 2. 3. 4. 5. 6.
This article was published online on 19 February 2014.
ATM O SPH ER I C S CI E N CE
Involatile particles from rapid oxidation How tiny aerosol particles form and grow from vapours produced by vegetation has been a mystery. The finding that highly oxygenated products form directly from volatile organic compounds may offer the solution. See Letter p.476 GORDON MCFIGGANS
O
n page 476 of this issue, Ehn et al.1 report their identification of large yields of highly oxygenated compounds when volatile organic compounds emitted from biological sources are exposed to atmospherically relevant conditions. This observation may help to close the gap between the measured mass of organic aerosol particles in the atmosphere and that predicted by models. It might also forge a mechanistic link between biogenic volatile organic compounds and the formation of aerosol particles, and provide insight into one of the main climate feedback cycles. Organic aerosol particles are widespread in the global atmosphere2. They consist of primary particles, which are emitted directly into the atmosphere, and secondary particles,
4 4 2 | N AT U R E | VO L 5 0 6 | 2 7 F E B R UA RY 2 0 1 4
© 2014 Macmillan Publishers Limited. All rights reserved
which form from the oxidation products of anthropogenic and biogenic volatile organic compounds. In the overall aerosol budget, the contribution of secondary particles that form from the oxidation of natural plant emissions is complex and highly uncertain. Secondary organic aerosol (SOA) probably dominates primary aerosol and, in some conditions, biogenic sources have been estimated3 to contribute up to 90% of SOA. Understanding the roles of naturally occurring atmospheric particles in the climate system has proved difficult. Because of mechanistic uncertainties, ‘bottom-up’ models that predict the concentrations of atmospheric aerosol particles by explicitly describing the emission and oxidation of their precursors have long been unable to predict the observed mass of SOA. Geographically widespread, long-term observations of high numbers of ultrafine
STEPHEN J. KRASEMANN/SPL
NEWS & VIEWS RESEARCH
Figure 1 | Forest haze. Ehn and colleagues’ discovery1 of a mechanism by which highly oxygenated compounds form directly from biogenic emissions, such as those produced by boreal forests, may help to explain how naturally occurring ultrafine particles form and grow in the atmosphere.
particles in the atmosphere, particularly over forested regions (Fig. 1), has stimulated much research into how such aerosol particles form and grow4,5. So far, compounds of sufficiently low volatility to explain the growth of ultrafine particles have not been identified or quantified. To address these problems, Ehn and colleagues made concurrent, direct measurements of the mass spectra of gaseous organic compounds in a chamber, and of the masses of ‘seed’ particles injected into the chamber, in experiments conducted under atmospherically reasonable conditions. They studied several biogenic volatile organic compounds (BVOCs), such as monoterpenes (members of the terpene family of naturally occurring hydrocarbons), and found that the oxidation products of these compounds had a high ratio of oxygen to carbon atoms. The products were irreversibly taken up by the seed particles, even at very low particle loadings. This irreversible condensation is expected for compounds that exhibit extremely low volatility, as highly oxygenated compounds generally do. Possibly the most surprising of the authors’ findings is that such compounds are formed at high yields at an early stage of oxidation — in oxidation reactions of the primary compounds, rather than in subsequent oxidation reactions. Observed atmospheric particle-growth rates are generally higher than previously proposed oxidation mechanisms can support, which has led to speculation that other mechanisms might play a part6. Ehn and co-workers’ observation of a hitherto unmeasured class of gas-phase oxidation products obviates the need to invoke such mechanisms. The authors convincingly argue that this direct pathway to low-volatility compounds can contribute to observed particle formation and explain particle growth in boreal forest regions,
consistent with recent predictions7. Ehn et al. propose that the products are formed efficiently in the reaction of endocyclic alkenes (chemical structures found in many abundant terpenes) with ozone. They suggest a mechanism consisting of a chain reaction in which a hydrogen atom is rapidly removed from a terpene molecule, forming a free radical to which an oxygen molecule attaches; this cycle repeats several times before terminating (see Extended Data Fig. 9a of the paper1). Their hypothesis may explain why models better approximate biogenic SOA if they allow the products of the first oxidation reactions of terpenes to be involatile8. In any case, inclusion of such products in atmospheric oxidation mechanisms can only improve our predictive capability and potentially close the gap between bottom-up models and atmospheric observations. The hypothesized mechanism also provides a means by which the particleforming potential of BVOCs can be affected by anthropogenic emissions, because the extent of the chain reaction will be influenced by atmospheric levels of compounds produced by human activities. The current study should provoke valuable developments in several areas. Attempts should be made to directly establish the oxidation mechanisms that so efficiently generate highly oxidized compounds, to identify the products and the kinetics of the participating reactions. This information will enable detailed analyses of the sensitivity of the SOAformation process to different conditions, and of the sensitivities of components of previously proposed oxidation mechanisms to the newly described process. Although Ehn and colleagues’ logic is convincing, they did not directly determine the volatility of the observed products. It would
therefore be beneficial if their attribution of complete involatility (or extremely low volatility) to the products is substantiated by measurement. That said, measurement will be difficult: the compounds are likely to have several hydroperoxide groups (HOO) and/or peroxy acids (HOOC=O) on a monoterpene backbone, and so will probably be difficult to make. Estimating the compounds’ volatilities will also be difficult, because hydroperoxides and peroxy acids are poorly represented in techniques used for such estimations. The current work highlights the importance of accurately measuring atmospheric oxidant levels and BVOC emissions. Anthropogenic changes in the ratio of ozone to hydroxyl radicals (both of which are key oxidizers in the atmosphere), and in concentrations of nitrogen oxides (which affect atmospheric ozone levels) will influence climate through SOA and its effects on concentrations of cloud condensation nuclei, the particles that act as seeds for cloud-droplet formation. Biogenic SOA particles probably have a substantial role in natural climate feedback cycles, whereby emissions of BVOCs are affected by the direct or indirect effects of the particles on the intensity of solar radiation that reaches the ground9,10. The influences of climate and of perturbations to terrestrial vegetation caused by human activities will also affect BVOC emissions. Mechanistic11 and sensitivity12 elements of the feedbacks mentioned above have received considerable recent attention. The mechanistic insights into the production of low-volatility compounds provided by Ehn et al. should lead to a better description of how aerosol particles and cloud condensation nuclei form, and of the associated climate feedbacks following BVOC emission changes, thereby improving the predictive capabilities of climate and Earth system models. ■ Gordon McFiggans is at the School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester M13 9PL, UK. e-mail: [email protected] 1. Ehn, M. et al. Nature 506, 476–479 (2014). 2. Jimenez, J. L. et al. Science 326, 1525–1529 (2009). 3. Hallquist, M. et al. Atmos. Chem. Phys. 9, 5155–5236 (2009). 4. Kulmala, M. et al. J. Aerosol Sci. 35, 143–176 (2004). 5. Kulmala, M. et al. Science 339, 943–946 (2013). 6. Donahue, N. M., Trump, E. R., Pierce, J. R. & Riipinen, I. Geophys. Res. Lett. 38, L16801 (2011). 7. Donahue, N. M. et al. Faraday Discuss. 165, 91–104 (2013). 8. Spracklen, D. V. et al. Atmos. Chem. Phys. 11, 12109–12136 (2011). 9. Kulmala, M. et al. Atmos. Chem. Phys. 4, 557–562 (2004). 10. Mentel, Th. F. et al. Atmos. Chem. Phys. 13, 8755–8770 (2013). 11. Topping, D., Connolly, P. & McFiggans, G. Nature Geosci. 6, 443–446 (2013). 12. Paasonen, P. et al. Nature Geosci. 6, 438–442 (2013).
2 7 F E B R UA RY 2 0 1 4 | VO L 5 0 6 | N AT U R E | 4 4 3
© 2014 Macmillan Publishers Limited. All rights reserved
compensatory autophagy in ER-stressed intestinal epithelium, owing to deletion of the Atg16l1 gene, results in Crohn’s-diseaselike inflammation of the small intestine. The identification of the sensitivity of this protein variant to caspase-3-mediated destruction may provide insight into how specific environmental exposures convert genetic disease risk into clinical symptoms. Murthy and colleagues’ study also helps to integrate other observations from work on Crohn’s disease. For example, caspase-3 activation can be regulated by inhibitor-of-apoptosis proteins such as XIAP, variants of which are the cause of a single-gene form of Crohn’s disease11. XIAP, in turn, directly interacts with other proteins involved in microbial sensing that are associated with genetic risk for Crohn’s disease4,11, including RIPK2 and NOD2. The new data might also help to explain why agents targeting TNF-α are highly effective therapies for Crohn’s disease2 — in light of the authors’ Y. enterocolitica infection studies, it seems plausible that TNF-α fuels an inflammatory feed-forward loop by inducing caspase-3mediated degradation of ATG16L1T300A, which in turn results in increased TNF-α secretion. Although further studies are necessary to test the authors’ biochemical and cellular observations in animal models and patients with
Crohn’s disease, their observations provide an appealing and integrating hypothesis for how this common genetic element engenders disease risk. ■ Arthur Kaser is in the Division of Gastroenterology and Hepatology, Department of Medicine, Addenbrooke’s Hospital, University of Cambridge, Cambridge CB2 0QQ, UK. Richard S. Blumberg is in the Division of Gastroenterology, Hepatology and Endoscopy, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115, USA. e-mails: [email protected]; [email protected] Hampe, J. et al. Nature Genet. 39, 207–211 (2007). Maloy, K. J. & Powrie, F. Nature 474, 298–306 (2011). Murthy, A. et al. Nature 506, 456–462 (2014). Jostins, L. et al. Nature 491, 119–124 (2012). Saitoh, T. et al. Nature 456, 264–268 (2008). Levine, B., Mizushima, N. & Virgin, H. W. Nature 469, 323–335 (2011). 7. McIlwain, D. R., Berger, T. & Mak, T. W. Cold Spring Harb. Perspect. Biol. 5, a008656 (2013). 8. Khalil, H. et al. Mol. Cell. Biol. 32, 4523–4533 (2012). 9. Deuring, J. J. et al. Gut http://dx.doi.org/10.1136/ gutjnl-2012-303527 (2013). 10. Adolph, T. E. et al. Nature 503, 272–276 (2013). 11. Uhlig, H. H. Gut 62, 1795–1805 (2013). 1. 2. 3. 4. 5. 6.
This article was published online on 19 February 2014.
ATM O SPH ER I C S CI E N CE
Involatile particles from rapid oxidation How tiny aerosol particles form and grow from vapours produced by vegetation has been a mystery. The finding that highly oxygenated products form directly from volatile organic compounds may offer the solution. See Letter p.476 GORDON MCFIGGANS
O
n page 476 of this issue, Ehn et al.1 report their identification of large yields of highly oxygenated compounds when volatile organic compounds emitted from biological sources are exposed to atmospherically relevant conditions. This observation may help to close the gap between the measured mass of organic aerosol particles in the atmosphere and that predicted by models. It might also forge a mechanistic link between biogenic volatile organic compounds and the formation of aerosol particles, and provide insight into one of the main climate feedback cycles. Organic aerosol particles are widespread in the global atmosphere2. They consist of primary particles, which are emitted directly into the atmosphere, and secondary particles,
4 4 2 | N AT U R E | VO L 5 0 6 | 2 7 F E B R UA RY 2 0 1 4
© 2014 Macmillan Publishers Limited. All rights reserved
which form from the oxidation products of anthropogenic and biogenic volatile organic compounds. In the overall aerosol budget, the contribution of secondary particles that form from the oxidation of natural plant emissions is complex and highly uncertain. Secondary organic aerosol (SOA) probably dominates primary aerosol and, in some conditions, biogenic sources have been estimated3 to contribute up to 90% of SOA. Understanding the roles of naturally occurring atmospheric particles in the climate system has proved difficult. Because of mechanistic uncertainties, ‘bottom-up’ models that predict the concentrations of atmospheric aerosol particles by explicitly describing the emission and oxidation of their precursors have long been unable to predict the observed mass of SOA. Geographically widespread, long-term observations of high numbers of ultrafine
STEPHEN J. KRASEMANN/SPL
NEWS & VIEWS RESEARCH
Figure 1 | Forest haze. Ehn and colleagues’ discovery1 of a mechanism by which highly oxygenated compounds form directly from biogenic emissions, such as those produced by boreal forests, may help to explain how naturally occurring ultrafine particles form and grow in the atmosphere.
particles in the atmosphere, particularly over forested regions (Fig. 1), has stimulated much research into how such aerosol particles form and grow4,5. So far, compounds of sufficiently low volatility to explain the growth of ultrafine particles have not been identified or quantified. To address these problems, Ehn and colleagues made concurrent, direct measurements of the mass spectra of gaseous organic compounds in a chamber, and of the masses of ‘seed’ particles injected into the chamber, in experiments conducted under atmospherically reasonable conditions. They studied several biogenic volatile organic compounds (BVOCs), such as monoterpenes (members of the terpene family of naturally occurring hydrocarbons), and found that the oxidation products of these compounds had a high ratio of oxygen to carbon atoms. The products were irreversibly taken up by the seed particles, even at very low particle loadings. This irreversible condensation is expected for compounds that exhibit extremely low volatility, as highly oxygenated compounds generally do. Possibly the most surprising of the authors’ findings is that such compounds are formed at high yields at an early stage of oxidation — in oxidation reactions of the primary compounds, rather than in subsequent oxidation reactions. Observed atmospheric particle-growth rates are generally higher than previously proposed oxidation mechanisms can support, which has led to speculation that other mechanisms might play a part6. Ehn and co-workers’ observation of a hitherto unmeasured class of gas-phase oxidation products obviates the need to invoke such mechanisms. The authors convincingly argue that this direct pathway to low-volatility compounds can contribute to observed particle formation and explain particle growth in boreal forest regions,
consistent with recent predictions7. Ehn et al. propose that the products are formed efficiently in the reaction of endocyclic alkenes (chemical structures found in many abundant terpenes) with ozone. They suggest a mechanism consisting of a chain reaction in which a hydrogen atom is rapidly removed from a terpene molecule, forming a free radical to which an oxygen molecule attaches; this cycle repeats several times before terminating (see Extended Data Fig. 9a of the paper1). Their hypothesis may explain why models better approximate biogenic SOA if they allow the products of the first oxidation reactions of terpenes to be involatile8. In any case, inclusion of such products in atmospheric oxidation mechanisms can only improve our predictive capability and potentially close the gap between bottom-up models and atmospheric observations. The hypothesized mechanism also provides a means by which the particleforming potential of BVOCs can be affected by anthropogenic emissions, because the extent of the chain reaction will be influenced by atmospheric levels of compounds produced by human activities. The current study should provoke valuable developments in several areas. Attempts should be made to directly establish the oxidation mechanisms that so efficiently generate highly oxidized compounds, to identify the products and the kinetics of the participating reactions. This information will enable detailed analyses of the sensitivity of the SOAformation process to different conditions, and of the sensitivities of components of previously proposed oxidation mechanisms to the newly described process. Although Ehn and colleagues’ logic is convincing, they did not directly determine the volatility of the observed products. It would
therefore be beneficial if their attribution of complete involatility (or extremely low volatility) to the products is substantiated by measurement. That said, measurement will be difficult: the compounds are likely to have several hydroperoxide groups (HOO) and/or peroxy acids (HOOC=O) on a monoterpene backbone, and so will probably be difficult to make. Estimating the compounds’ volatilities will also be difficult, because hydroperoxides and peroxy acids are poorly represented in techniques used for such estimations. The current work highlights the importance of accurately measuring atmospheric oxidant levels and BVOC emissions. Anthropogenic changes in the ratio of ozone to hydroxyl radicals (both of which are key oxidizers in the atmosphere), and in concentrations of nitrogen oxides (which affect atmospheric ozone levels) will influence climate through SOA and its effects on concentrations of cloud condensation nuclei, the particles that act as seeds for cloud-droplet formation. Biogenic SOA particles probably have a substantial role in natural climate feedback cycles, whereby emissions of BVOCs are affected by the direct or indirect effects of the particles on the intensity of solar radiation that reaches the ground9,10. The influences of climate and of perturbations to terrestrial vegetation caused by human activities will also affect BVOC emissions. Mechanistic11 and sensitivity12 elements of the feedbacks mentioned above have received considerable recent attention. The mechanistic insights into the production of low-volatility compounds provided by Ehn et al. should lead to a better description of how aerosol particles and cloud condensation nuclei form, and of the associated climate feedbacks following BVOC emission changes, thereby improving the predictive capabilities of climate and Earth system models. ■ Gordon McFiggans is at the School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester M13 9PL, UK. e-mail: [email protected] 1. Ehn, M. et al. Nature 506, 476–479 (2014). 2. Jimenez, J. L. et al. Science 326, 1525–1529 (2009). 3. Hallquist, M. et al. Atmos. Chem. Phys. 9, 5155–5236 (2009). 4. Kulmala, M. et al. J. Aerosol Sci. 35, 143–176 (2004). 5. Kulmala, M. et al. Science 339, 943–946 (2013). 6. Donahue, N. M., Trump, E. R., Pierce, J. R. & Riipinen, I. Geophys. Res. Lett. 38, L16801 (2011). 7. Donahue, N. M. et al. Faraday Discuss. 165, 91–104 (2013). 8. Spracklen, D. V. et al. Atmos. Chem. Phys. 11, 12109–12136 (2011). 9. Kulmala, M. et al. Atmos. Chem. Phys. 4, 557–562 (2004). 10. Mentel, Th. F. et al. Atmos. Chem. Phys. 13, 8755–8770 (2013). 11. Topping, D., Connolly, P. & McFiggans, G. Nature Geosci. 6, 443–446 (2013). 12. Paasonen, P. et al. Nature Geosci. 6, 438–442 (2013).
2 7 F E B R UA RY 2 0 1 4 | VO L 5 0 6 | N AT U R E | 4 4 3
© 2014 Macmillan Publishers Limited. All rights reserved
