EROS is a selective chaperone regulating the phagocyte NADPH oxidase and purinergic signalling

  1. Lyra O Randzavola
  2. Paige M Mortimer
  3. Emma Garside
  4. Elizabeth R Dufficy
  5. Andrea Schejtman
  6. Georgia Roumelioti
  7. Lu Yu
  8. Mercedes Pardo
  9. Kerstin Spirohn
  10. Charlotte Tolley
  11. Cordelia Brandt
  12. Katherine Harcourt
  13. Esme Nichols
  14. Mike Nahorski
  15. Geoff Woods
  16. James C Williamson
  17. Shreehari Suresh
  18. John M Sowerby
  19. Misaki Matsumoto
  20. Celio XC Santos
  21. Cher Shen Kiar
  22. Subhankar Mukhopadhyay
  23. William M Rae
  24. Gordon J Dougan
  25. John Grainger
  26. Paul J Lehner
  27. Michael A Calderwood
  28. Jyoti Choudhary
  29. Simon Clare
  30. Anneliese Speak
  31. Giorgia Santilli
  32. Alex Bateman
  33. Kenneth GC Smith
  34. Francesca Magnani
  35. David C Thomas

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This study focus follows this group's previous work on EROS and NOX2. In this current study the authors examine neutrophil EROS in the generation of superoxide by the NADPH oxidase. They demonstrate how EROS is involved in the maturation of gp91phox and expand our knowledge on the role of EROS in regulating expression of the P2x7 ion channel.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

EROS (essential for reactive oxygen species) protein is indispensable for expression of gp91 phox , the catalytic core of the phagocyte NADPH oxidase. EROS deficiency in humans is a novel cause of the severe immunodeficiency, chronic granulomatous disease, but its mechanism of action was unknown until now. We elucidate the role of EROS, showing it acts at the earliest stages of gp91 phox maturation. It binds the immature 58 kDa gp91 phox directly, preventing gp91 phox degradation and allowing glycosylation via the oligosaccharyltransferase machinery and the incorporation of the heme prosthetic groups essential for catalysis. EROS also regulates the purine receptors P2X7 and P2X1 through direct interactions, and P2X7 is almost absent in EROS-deficient mouse and human primary cells. Accordingly, lack of murine EROS results in markedly abnormal P2X7 signalling, inflammasome activation, and T cell responses. The loss of both ROS and P2X7 signalling leads to resistance to influenza infection in mice. Our work identifies EROS as a highly selective chaperone for key proteins in innate and adaptive immunity and a rheostat for immunity to infection. It has profound implications for our understanding of immune physiology, ROS dysregulation, and possibly gene therapy.

Article activity feed

  1. Version published to 10.7554/elife.76387 on eLife
  2. Version published to 10.1101/2021.09.14.460103v2 on bioRxiv
  3. eLife

    Author Response

    Reviewer #3 (Public Review):

    The study by Randzavola and colleagues provides a follow-up of their previous publication (Thomas DC et al, J Exp Med 2017) describing EROS (Essential for Reactive Oxygen Species or C17Orf62) as a novel chaperone essentially required to support the phagocyte Nox NADPH Oxidase respiratory burst and bacterial killing. Here, the authors extend the investigation of the mechanism underlying EROS effect and show its very early binding in the endoplasmic reticulum and interaction with immature partially glycosylated forms of gp91phox (the catalytic subunit of the Nox complex), allowing the incorporation of heme and subsequent binding of p22phox, which later follows the usual steps for complex maturation. A novel finding was the association of EROS with the OST component of the N-glycosylation machinery. An extended proteome analysis confirmed that EROS is quite specific for the gp91phox/p22phox complex and also for the purinergic P2X7 receptor, which also interacts with EROS (as also shown previously by the authors and further investigated by Ryoden et al. J Immunol 2020). The authors further validate EROS binding to P2X7 and provide evidence that EROS loss-of-function impairs P2X7-associated functions. Particularly, mice with genetically ablated EROS show improved survival to influenza infection.

    A major strength of this line of investigation is the clear functional importance of EROS in the regulation of the protein expression of the Nox complex components. Previous work has clearly shown that human EROS deficiency associated with the severe immunodeficiency Chronic Granulomatous Disease, which is usually caused by genetic deficiency of the Nox complex components. Indeed, the loss of gain of functions of EROS are very clearly associated with major changes in the expression of those components, indicating EROS functional relevance. Moreover, the interplay between the P2X7 receptor and EROS is also relevant, given that this receptor mediates an important arm of innate immunity, namely the nucleotide-driven inflammasome activation. Thus, the authors are likely dealing with some undoubtedly important novel information which may be of broad impact to understand several aspects of the adaptive and even adaptive immunity.

    Enthusiasm for this article, however, is somewhat decreased by some aspects, as follows:

    1. While there is a substantial amount of new data, the corresponding progress in depth of mechanistic insights has not been commensurate, bearing in mind the author's previous work. The novel findings are the more clear documentation of EROS/gp91phox interaction and its time-course during nascent gp91phox protein processing in the ER. Also, their interplay with the OST complex. The extended list of proteins associating with EROS essentially confirms previous findings. Also, the work with P2X7 mostly confirms previous findings, while the novel and interesting experiment with EROS-silenced mice and viral infection needs further work, as commented below.

    We thank the reviewer for this comment and for seeking clarity on novelty. We have addressed this above and in the discussion section. We have not reported the EROS interactome by mass spectrometry in previous work.

    1. Some aspects of these results are less than clearcut. The association between gp91phox and EROS is generally convincing, but for many experiments the authors make wide use of transfections of tagged protein constructs. One can clearly understand that this is possibly the only feasible approach at this time, however these constructs carry the intrinsic problem of possible protein misfolding, which would make them a potentially artificial target of any endoplasmic reticulum chaperone-like protein such as EROS. This would impact exactly on the very mechanism the authors are proposing for EROS effects, i.e., early protein processing.

    We understand Reviewer 3’s concerns about using tagged constructs. However, all transfection experiments depicted in Figure 1 have been done with untagged constructs and in different cell types in both mouse and human systems. The whole approach is also validated by extensive previous work showing the ability of transfected p22phox to augment gp91phox expression (Yu et al., J Biol Chem 1997; PMID: 9341176). All our experiments showed the same result, namely the stabilisation of the 58kDa gp91phox precursor. We have now included data showing we can immunoprecipitate endogenous gp91phox in PLB985 cells and detect endogenous EROS (Figure 3, figure supplement 1A) which confirms the specificity of the association between gp91phox and EROS. In the same sample, we can also detect endogenous p22phox (our positive control) which is well-known to associate as heterodimer with gp91phox. Furthermore, transfection of our constructs does not induce significant ER stress in HEK293 cells. Based on our own data and that of other investigators, we argue that this is a valid and useful approach to demonstrating the ability of EROS to increase gp91phox abundance. Similarly, this is just one of many orthologous techniques used in the manuscript.

    1. The same consideration applies to the experiments in Figure 3 with the OST complex STT3A. The co-localizations shown by the authors are technically acceptable, but their meaning is unclear, given it is expected that the proteins EROS and OST occupy the same compartment, being ER-located proteins, especially if transfected as constructs (tagged or not).

    The experiment has been done to assess the localisation of gp91phox relative to EROS and STT3A which are known to occupy the ER -compartment as pointed by the reviewer. Since HEK293 cells do not express gp91phox, this microscopy analysis allowed to determine if some population of gp91phox could be detected with EROS and STT3A at the ER as opposed to its localization as a mature protein at the plasma membrane and within granules, in phagocytic cells.

    1. It would be important to assess whether cells receiving such constructs depict markers of endoplasmic reticulum stress and/or show impaired survival.

    This has been addressed in Reviewer 3’s recommendation for author point 2.

    1. The experiments with co-transfection in HEK293 cells of EROS, Nox1 and Nox4 provide results at variance with the author's data in their previous work, in which endogenous Nox1 (intestine) and Nox4 (kidney) had no changes in expression in genetically silenced EROS mice.

    We thank the reviewer for this comment and acknowledge that this introduces some ambiguity. In showing the augmentation of NOX1, NOX4 but not p22phox or NOX5 we are demonstrating that it is likely that EROS can bind and stabilise NOX proteins that also require p22phox. In the case of NOX4, this is also supported by our yeast 2 hybrid data. Thus, these data suggest that EROS can bind p22phox-dependent NOX proteins. The key question is whether EROS has a physiological role in controlling the expression of other NOX proteins. Although we addressed this in our previous study, we have done so in a more extensive way in this manuscript. In particular, we note the subsequent publication by Diebold et al. (Methods Mol Biol 2019; PMID: 31172474) which points out that many commercially available antibodies are non-specific. Detailed examination showed this to be the case for the antibody we used in Thomas et al., (J Exp Med, 2017; PMID: 28351984). We therefore undertook specific analysis with the anti-mouse NOX1 antibody clone from Dr C. Yabe-Nishimura and Dr. Misaki Matsumoto.

    Similarly, our work on NOX4 in Thomas et al 2017 (J Exp Med, 2017; PMID: 28351984) suggested that while NOX4 is certainly present in the kidneys of EROS-/- mice, this was a limited analysis as it was not the main focus of the paper, and the conclusion was that there was no drastic effect on NOX4 expression in the same manner as that observed for NOX2. For the revisions to this paper, we examined a cohort of 4 control and 4 EROS-/- mice and showed that EROS does not physiologically regulate NOX4 in the kidney.

    Thus, the use of HEK293, which do not express NOX proteins, as a reductionist system may favour the effect of EROS on NOX1 and NOX4 abundance upon transfection of the constructs. One possible explanation could be that EROS binds to a conserved motifs present on NOX1, NOX2 and NOX4 which is readily accessible in the system we are using.

    1. The article is conceptually divided into two parts. However, there is no clear cross-fertilization between them and they essentially do not integrate.

    Although the reviewer notes that it seems that there are two separate stories, this reflects that we have extensively characterised the function of EROS and found that it specifically and profoundly affects only two distinct pathways in immunity, which is significant in itself. A strength of our manuscript is our extensive granular mass spectrometry approach which shows the specificity of EROS in 2 different cell types in which up to 8000 proteins have been detected. We have therefore placed the control of P2X7 and gp91phox-p22phox in context of the entire proteome. Our paper defines just how specific EROS is in its physiological effects and we therefore focus on the two major pathways that are affected by EROS deficiency. We integrate this in the final figure by showing how the combined lack of gp91phox and P2X7 lead to resistance to influenza A in contrast to the susceptibility to certain bacterial infections.

    1. While the authors claim that "the loss of both ROS and P2X7 signalling leads to resistance to influenza infection", this was not in fact shown in this work. It is known that P2X7 deficiency protects against influenza infection. Thus, it follows naturally that EROS deficiency, which essentially eliminates the expression of P2X7, would have the same effect. However, the role of ROS and gp91phox, i.e. whether or not they add to this equation, remains unclear.

    We thank the reviewer for this comment. The role of phagocyte NADPH oxidase-derived ROS has been explored in gp91phox deficiency and we apologise if this is not made clear in our manuscript. We have now added the following text to the discussion section of the manuscript:

    “A particular strength of our study is that we show marked in vivo sequelae of the lack of P2X7. EROS deficiency leads to profound susceptibility to bacterial infection but protects mice from infection with influenza A. This is likely to reflect the fact that mice that are (i) deficient in gp91phox (ii) deficient in P2X7 (iii) treated with P2X7 inhibitors have improved outcomes following infection with influenza A and raises intriguing questions about the physiological role of EROS. Snelgrove et al showed that gp91phox deficiency improved outcomes in influenza A. gp91phox knockout mice exhibited a reduced influenza titre in the lung parenchyma. Inflammatory infiltrate into the lung parenchyma was markedly reduced and lung function significantly improved (Snelgrove et al., 2006). To et al then showed that the phagocyte NADPH oxidase is activated by single stranded RNA and DNA viruses in endocytic compartments. This causes endosomal hydrogen peroxide generation, which suppresses antiviral and humoral signalling networks via modification of a highly conserved cysteine residue (Cys98) on Toll-like receptor-7. In this study, targeted inhibition of endosomal reactive oxygen species production using cholestanol-conjugated gp91dsTAT (Cgp91ds-TAT) abrogates influenza A virus pathogenicity (To et al., 2017). This group went on to explore infection with a more pathogenic influenza A strain, PR8. Using the same specific inhibitor. Cgp91ds-TAT reduced airway inflammation, including neutrophil influx and alveolitis and enhanced the clearance of lung viral mRNA following PR8 infection (To et al., 2019). This group has also shown that NOX1 (Selemidis et al., 2013) and NOX4 (Hendricks et al., 2022) can drive pathogenic inflammation in influenza A, emphasising the importance of clarifying the roles of EROS in control of expression of these proteins.

    In studies on P2X7, Rosli et al showed that mice infected with 105 PFU of influenza A HKx31 had improved outcomes if they were treated with a P2X7 inhibitor at day 3 post infection and every two days thereafter. Survival was also improved even if the inhibitor is given on day 7 post infection following a lethal dose of the mouse adapted PR8. This was associated with reduced cellular infiltration and pro-inflammatory cytokine secretion in bronchoalveolar lavage fluid, but viral titres were not measured (Rosli et al., 2019). Leyva-Grado et al examined influenza A infection in P2X7 knockout mice. They infected mice with both influenza A/Puerto Rico/08/1934 virus and influenza A/Netherlands/604/2009 H1N1pdm virus. They showed that P2X7 receptor deficiency led to improved survival after infection with both viruses with less weight loss (Leyva-Grado et al., 2017). Production of proinflammatory cytokines and chemokines was impaired and there were fewer cellular hallmarks of severe infection such as infiltration of neutrophils and depletion of CD11b+ macrophages. It is worth noting that the P2X7 knockout strain used in this study was the Pfizer strain in which some splice variants of P2X7 are still expressed (Bartlett et al., 2014). Hence, the dual loss of the phagocyte NADPH oxidase and P2X7 in EROS-/- mice likely confers protection from IAV infection. By reducing the expression of both NOX2 and P2X7, EROS regulates two pathways that may be detrimental in influenza A and we speculate that EROS may physiologically act as a rheostat controlling certain types of immune response.”

  4. eLife

    Evaluation Summary:

    This study focus follows this group's previous work on EROS and NOX2. In this current study the authors examine neutrophil EROS in the generation of superoxide by the NADPH oxidase. They demonstrate how EROS is involved in the maturation of gp91phox and expand our knowledge on the role of EROS in regulating expression of the P2x7 ion channel.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

  5. eLife

    Reviewer #1 (Public Review):

    This is an interesting manuscript providing important new information on the mechanism of action of EROS in the generation of superoxide by the NADPH oxidase of neutrophils. The authors have shown in previous publications that EROS deficiency results in defective NOX2 activity and thus represents a hitherto unrecognised, rare form of chronic granulomatous disease. They now show how EROS is involved in oligosaccharide transfer during the maturation of gp91phox and also extend what is known about the role for EROS in regulating expression of the P2x7 ion channel.

    The results presented in the manuscript are supported by findings from a variety of techniques and for the most part, are convincing and well presented. However, I do have queries about certain aspects of the manuscript.

    1. Figure 1
    The much lower EROS expression when gp91phox is expressed warrants a comment.
    Fig 1 G. Please explain what fold change represents. From F, zero time expression appears much more than the 1.5 fold higher shown in G for the EROS-expressing cells. This needs explaining. With the very high error bars (presumably for the EROS sample although this is not clear) overlapping zero I find it hard to conclude anything from this figure.

    2. P 9 line 9 states that Fig 1H shows that cycloheximide increases expression. Yet it appears from the legend that cycloheximide is present in all samples and it is EROS that increases expression. Please clarify.

    3. Fig 3A&B and p12 1st para. The identification of OST as a binding partner is interesting and a significant novel finding. However, the presentation of this information appears to me to be unduly complex and more information is required. Not all the readers will be familiar with the details of SAINTexpress methodology and more explanation of what is being shown would be helpful. At the least, a supplementary Table of the 59 identified proteins would be helpful, plus information on controls to establish selective pull down by EROS and on how the blue spots in A relate to the proteins. Also please make it clearer which of the proteins in B were identified and the relevance of showing all the steps in the pathway.

    4. Figure 6. This contains a large amount of information. Although interesting, I am concerned that the authors may be trying to include too much at the expense of the necessary detail for some of the experiments. For example, the EROS -/- +ATP scattergram on the left of Fig 6E does not seem to agree with the right hand graph. I would also like to see the mean values for the 5 experiments in Fig 6G shown. Most importantly, insufficient information is given for Fig 6H. I don't think I missed it but I could find no details about the experiment in the Methods section. We need to know more about exactly how many animals in were in each group (death of 1 animal appears to equate to 5% of total - how does this relate to >10 in total), how signs of illness were monitored and related to death, and generally more about the conditions of the experiment. Alternatively, this may be better left to a more detailed study.

  6. eLife

    Reviewer #2 (Public Review):

    Randzavola et al. investigated the role of EROS (Essential for reactive oxygen species) in the early maturation stages of the Nox2 enzyme responsible for ROS production. They find that EROS in a similar manner supports assembly and stability of purinergic membrane receptors such as P2X7, linking Nox2 and inflammasome activation.

    Moreover, they show that inhibition of EROS by affecting ROS and inflammasome activation might ameliorate pathology during influenza A virus infection and has the potential to tune antiviral immune responses in general.

    These data add in an interesting way to previous work on the function of EROS in mice and humans and provide further mechanistic insight into EROS interaction with Nox2 and P2X7 receptor. The conclusions of this paper are mostly well supported by data, but some details on the in vivo influenza A virus infection experiment need to be clarified.

  7. eLife

    Reviewer #3 (Public Review):

    The study by Randzavola and colleagues provides a follow-up of their previous publication (Thomas DC et al, J Exp Med 2017) describing EROS (Essential for Reactive Oxygen Species or C17Orf62) as a novel chaperone essentially required to support the phagocyte Nox NADPH Oxidase respiratory burst and bacterial killing. Here, the authors extend the investigation of the mechanism underlying EROS effect and show its very early binding in the endoplasmic reticulum and interaction with immature partially glycosylated forms of gp91phox (the catalytic subunit of the Nox complex), allowing the incorporation of heme and subsequent binding of p22phox, which later follows the usual steps for complex maturation. A novel finding was the association of EROS with the OST component of the N-glycosylation machinery. An extended proteome analysis confirmed that EROS is quite specific for the gp91phox/p22phox complex and also for the purinergic P2X7 receptor, which also interacts with EROS (as also shown previously by the authors and further investigated by Ryoden et al. J Immunol 2020). The authors further validate EROS binding to P2X7 and provide evidence that EROS loss-of-function impairs P2X7-associated functions. Particularly, mice with genetically ablated EROS show improved survival to influenza infection.

    A major strength of this line of investigation is the clear functional importance of EROS in the regulation of the protein expression of the Nox complex components. Previous work has clearly shown that human EROS deficiency associated with the severe immunodeficiency Chronic Granulomatous Disease, which is usually caused by genetic deficiency of the Nox complex components. Indeed, the loss of gain of functions of EROS are very clearly associated with major changes in the expression of those components, indicating EROS functional relevance. Moreover, the interplay between the P2X7 receptor and EROS is also relevant, given that this receptor mediates an important arm of innate immunity, namely the nucleotide-driven inflammasome activation. Thus, the authors are likely dealing with some undoubtedly important novel information which may be of broad impact to understand several aspects of the adaptive and even adaptive immunity.

    Enthusiasm for this article, however, is somewhat decreased by some aspects, as follows:

    1. While there is a substantial amount of new data, the corresponding progress in depth of mechanistic insights has not been commensurate, bearing in mind the author's previous work. The novel findings are the more clear documentation of EROS/gp91phox interaction and its time-course during nascent gp91phox protein processing in the ER. Also, their interplay with the OST complex. The extended list of proteins associating with EROS essentially confirms previous findings. Also, the work with P2X7 mostly confirms previous findings, while the novel and interesting experiment with EROS-silenced mice and viral infection needs further work, as commented below.
    2. Some aspects of these results are less than clearcut. The association between gp91phox and EROS is generally convincing, but for many experiments the authors make wide use of transfections of tagged protein constructs. One can clearly understand that this is possibly the only feasible approach at this time, however these constructs carry the intrinsic problem of possible protein misfolding, which would make them a potentially artificial target of any endoplasmic reticulum chaperone-like protein such as EROS. This would impact exactly on the very mechanism the authors are proposing for EROS effects, i.e., early protein processing.
    3. The same consideration applies to the experiments in Figure 3 with the OST complex STT3A. The co-localizations shown by the authors are technically acceptable, but their meaning is unclear, given it is expected that the proteins EROS and OST occupy the same compartment, being ER-located proteins, especially if transfected as constructs (tagged or not).
    4. It would be important to assess whether cells receiving such constructs depict markers of endoplasmic reticulum stress and/or show impaired survival.
    5. The experiments with co-transfection in HEK293 cells of EROS, Nox1 and Nox4 provide results at variance with the author's data in their previous work, in which endogenous Nox1 (intestine) and Nox4 (kidney) had no changes in expression in genetically silenced EROS mice.
    6. The article is conceptually divided into two parts. However, there is no clear cross-fertilization between them and they essentially do not integrate.
    7. While the authors claim that "the loss of both ROS and P2X7 signalling leads to resistance to influenza infection", this was not in fact shown in this work. It is known that P2X7 deficiency protects against influenza infection. Thus, it follows naturally that EROS deficiency, which essentially eliminates the expression of P2X7, would have the same effect. However, the role of ROS and gp91phox, i.e. whether or not they add to this equation, remains unclear.
  8. Version published to 10.1101/2021.09.14.460103v1 on bioRxiv