The UbiB family member Cqd1 forms a novel membrane contact site in mitochondria

  1. Siavash Khosravi
  2. Xenia Chelius
  3. Ann-Katrin Unger
  4. Johanna Frickel
  5. Timo Sachsenheimer
  6. Christian Lüchtenborg
  7. Rico Schieweck
  8. Britta Brügger
  9. Benedikt Westermann
  10. Till Klecker
  11. Walter Neupert
  12. Max E. Harner

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Mitochondria are essential organelles of eukaryotic cells that are characterized by their unique and complex membrane system. They are confined from the cytosol by an envelope consisting of two membranes. Signals, metabolites, proteins and lipids have to be transferred across these membranes via proteinaceous contact sites to keep mitochondria functional. In the present study we identified a novel mitochondrial contact site that is formed by the inner membrane protein Cqd1 and the outer membrane proteins Por1 and Om14. Similar to the mitochondrial porin, Por1, Cqd1 is highly conserved, suggesting that this complex is conserved in form and function from yeast to human. Cqd1 is a member of the UbiB protein kinase-like family (also called aarF domain containing kinases). It was recently shown that Cqd1 in cooperation with Cqd2 controls the cellular distribution of coenzyme Q by a yet unknown mechanism. Our data suggest that Cqd1 in addition is involved in the homeostasis of phospholipids and contributes to the maintenance of mitochondrial morphology and architecture.

Summary statement

Here, we show that the conserved mitochondrial inner membrane protein Cqd1 interacts with the outer membrane proteins Por1 and Om14. Additionally, we provide evidence that Cqd1 is important for maintaining mitochondrial homeostasis.

Article activity feed

  1. Version published to 10.1101/2022.04.09.487722v2 on bioRxiv
  2. 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

    Reply to the reviewers

    We thank the reviewers for their constructive comments and are pleased that all reviewers share our opinion, that the present study “makes an important contribution to the molecular architecture of mitochondria”, is in addition “an important advancement in our understanding of the mechanism by which Cqd1 regulates CoQ distribution” and will “thereby appealing to the broad readership of the journals”. We are convinced that addressing the important points raised by the reviewers will further strengthen the manuscript and result in additional significant insights in the molecular function of Cqd1.

    Reviewer #1:

    The major concerns affecting the conclusions are:

    1. Experimental evidence is lacking on the contribution of contact site formation by Cqd1 to the effects on mitochondrial architecture and respiration-dependent growth. Determining the effects of the overexpression of the kinase-dead mutant on mitochondrial morphology and contact site formation with Por1-Om14 can address that.

    We thank reviewer #1 for raising these important points. Indeed, the various functions of Cqd1 might be independent from each other and so far we cannot distinguish between them. As suggested by the reviewer we will analyze the effect of overexpression of CQD1 in the Dups1 deletion mutant and make use of the point mutant in the conserved ATP binding domain which cannot complement the phenotype of the Dups1 Dcqd1 double deletion mutant. We generated a yeast mutant strain expressing Om14-3xHA in the absence of wild type Cqd1. Expression of the cqd1(E330A) mutant in the Om14-3xHA background and subsequent immunoprecipitation will allow us to test whether ATP binding is also essential for contact site formation. Preliminary experiments showed that the overexpression of cqd1(E330A) in the Dcqd1 deletion background results in a growth defect comparable to that caused by overexpression of CQD1 WT. Therefore, we think it might be more promising to analyze the interaction of Om14 and Cqd1 E330A at wild type level in order to avoid pleiotropic effects.

    In addition, we will further characterize the cqd1(E330A) mutant by analyzing the effect of its overexpression on mitochondrial morphology, cell growth and assembly of MICOS and F1FO ATP synthase in the Dcqd1 deletion background.

    1. Related to point #1, Cqd1 overexpression in deltaUsp1 cells could have addressed whether the role of Cqd1 in contact sites and mitochondrial architecture is independent of its role on CoQ distribution and phospholipid metabolism. Further characterization of the kinase-dead Cqd1 mutant on CoQ distribution, contact sites, mitochondrial archictecture and phsophsolipid metabolism might help discerning how these activities can be separated.

    We agree that the related points 1) and 2) raised by reviewer #1 are important and addressed our plans in the response on point 1).

    1. It is unclear how both Cqd1 overexpression and deletion induce mitochondrial fragmentation. Performing live cell imaging with a mitochondrial-phoactivatable GFP to measure mitochondrial fusion rates could help discerning the causes for fragmentation. It is a possibility that overexpression induced fragmentation by activating fission without changing fusion, while deletion induced fragmentation by blocking fusion.

    We thank reviewer #1 for bringing up this point. Perhaps our explanation in this respect was too short. Fig. 4E shows that deletion of CQD1 does not result in altered mitochondrial morphology, however, deletion of CQD1 in the Dups1 background leads to virtual complete fragmentation of the mitochondrial network. This is likely due to inhibition of mitochondrial fusion through disturbed processing of the fusion protein Mgm1 (see Fig. 4D). In contrast, overexpression of CQD1 does NOT result in formation of small mitochondrial fragments, but in formation of huge mitochondrial clusters which in addition contain a large proportion of ER membranes. So, we don’t think that this phenotype is related to either enhanced fission or reduced fusion. We will clarify this point in text of the revised manuscript.

    Minor comment:

    1. Figure 4 claims that mitochondrial function is impaired by ups1 deletion, which Cqd1 deletion exacerbates. However, no respiration data is shown in figure 1, only measurements of mitochondrial architecture are shown. Thus, oxygen consumption measurements are needed to claim effects on mitochondrial function.

    We did not want to claim that mitochondria lose respiratory competence upon simultaneous deletion of CQD1 and UPS1. Actually, our results indicate that the Dups1 Dcqd1 double deletion mutant grows like wild type on complete medium containing glycerol. Therefore, respiration is not impaired in this mutant. However, mitochondrial function is not restricted to ATP production by oxidative phosphorylation. The reviewer probably refers to Figure 4 where we show that mitochondrial biogenesis and dynamics are impaired in the Dups1 Dcqd1 double deletion mutant – the heading of the legend summarizes this as "mitochondrial function". We will be more precise in the revised version on this point and add a panel showing growth of the mutant strain on non-fermentable carbon source to avoid any further confusion.

    1. Some Western blots lack quantifications and statistical analyses of independent experiments.

    It is correct that some quantification and the respective statistics were missing in the initially submitted manuscript. We will add the requested information in the revised version of the manuscript.

    Reviewer #2:

    I have the following concerns for the authors to consider. (1) Although biochemical evidence shows that Cqd1 is likely a factor that forms CS structures in mitochondria, it would make the manuscript stronger if the authors can observe uneven distribution of Cqd1 in the mitochondrial membranes (assessed by fluorescent microscopy or ideally high-resolution microscopy) and the presence of Cqd1 in the region of close apposition of the OM and IM by immunogold labeling for electron microscopy.

    Two independent lines of evidence show that Cqd1 is a novel contact site protein: (i) it is found in the contact site fraction in density gradients (Fig. 6A), and (ii) it can be co-immunoprecipitated with outer membrane proteins (Fig. 6G, H, I). Furthermore, the co-IP is supported by cross-links of expected size (Fig. 6F). In sum, we feel that this is solid evidence to support our claim that Cqd1 is present in mitochondrial contact sites. However, it still might be interesting to check an uneven distribution of Cqd1 in mitochondria, as suggested by the reviewer. We will do this by 3D deconvolution fluorescence microscopy.

    (2) Since the structural characterization of Cqd1 is important to understand its interactions with the OM proteins and other UbiB protein kinase-like family proteins, Coq8 and Cqd2, take different orientations, the membrane topology of Cqd1 should be experimentally analyzed. The authors state, "two hydrophobic stretches can be identified in the Cqd1 sequence, of which the first one (amino acids 125-142) might be a bona fide transmembrane segment" (lines 97-100); then is Cqd1 a single membrane spanning protein or two-membrane spanning protein?

    Unfortunately, it was not possible to test the location of the N terminus experimentally because an N-terminally tagged variant of Cqd1 (tag inserted between presequence and mature part) turned out to be unstable. We consider it very unlikely that the second hydrophobic stretch is a transmembrane domain as it is rather short (only 11 amino acids). Furthermore, several Cqd1 homologs in other fungi, including Yarrowia lipolytica, Aspergillus niger and Schizosaccharomyces pombe, are lacking the second hydrophobic stretch. Therefore, we propose that the major part of Cqd1 including the protein kinase-like domain is exposed to the intermembrane space. We will point out this more clearly in the revised manuscript.

    (3) The authors state, "conserved GxxxG dimerization motif (amino acids 504‐508)" (Fig. 1A caption), but this description needs a reference. The GxxxG motif was proposed to mediate transmembrane helix-helix association (https://doi.org/10.1006/jmbi.1999.3489), which is not consistent with the membrane topology proposed by the authors.

    We thank reviewer #2 for this comment. It is correct that GxxxG motifs are usually present in transmembrane a-helices. However, there is information available indicating that these motifs may also be present in soluble proteins and are stabilizing dimeric interactions for instance in the homodimeric Holliday-junction protein resolvase (Kleiger et al., 2002; doi: 10.1021/bi0200763.). However, as this point is not critical for our conclusions we will remove the discussion of the GxxxG motif from the revised manuscript.

    (4) What is the role of the kinase activity of Cqd1 in the CS formation? The effects of overexpression of Cqd1 (Fig. 7) should be tested for its E330A mutant.

    We also thank reviewer #2 for raising this important point similar to reviewer #1. Please see our response to point 1) of reviewer #1.

    (5) Is there stoichiometric as well as quantitative information on the 400 kD complex consisting of Cqd1, Por1 and Om14? Does the stoichiometry and amount of the complex depend on the growth condition? Does the complex contain other Por1 interacting IM proteins like Mdm31?

    We appreciate that reviewer #2 points out this important aspect. It might well be that the amount of the Cqd1 containing complex depends on growth conditions since its presence might be important for phospholipid homeostasis, CoQ distribution and mitochondrial architecture and morphology which for sure strongly depend on growth conditions. Therefore, we will try to analyze the amount of the Cqd1 complex present in mitochondria isolated from yeast cells grown on different media by BN-PAGE. So far we do not have any information on the stoichiometry of this complex and we feel that an analysis would go beyond the scope of this study. We agree with reviewer #2 that Mdm31 is an obvious candidate for an interaction partner of Cqd1. We actually tested this by co-immunoprecipitation using Cqd1-3xHA or Mdm31-3xHA. However, none of these approaches resulted in successful co-isolation of the potential interaction partner. We will mention this result in the revised manuscript.

    (6) For Fig. 7E, the authors state, "consistently, we observed dramatically increased mitochondria‐ER interactions Cqd1 overexpression", but this observation could be due to secondary effects because overexpression of Cqd1 itself already caused abnormal morphology of mitochondria.

    We thank reviewer #2 for bringing up this important point. To check whether the increased mitochondria‐ER interactions are a secondary effect due to altered mitochondrial morphology we will analyze the mitochondria‐ER interactions in other mitochondrial morphology mutants by fluorescence microscopy. This will reveal whether abnormal mitochondrial morphology generally leads to disturbed ER structure.

    (7) Since the antagonistic role of Cqd2 to Cqd1 was proposed, the results of the experiments for Cqd1 can be compared with those for Cqd2. For example, what will become of overexpression of Cqd2 instead of Cqd1 for Fig. 7? What is the lipid composition of the cqd1Dcqd2D double deletion mutant cells (the decreased PA level is recovered?)? Lines 424-425: In summary, overexpression of Cqd1 causes severe phenotypes on growth, formation of mitochondrial structural elements, and mitochondrial architecture and morphology. Is this phenotype affected by overexpression of Cqd2?

    This point raised by reviewer #2 is very interesting. Our preliminary experiments and previously published data (Tan et al., 2013) indicate that overexpression of Cqd2 is also toxic and results in the formation of huge mitochondrial clusters. Therefore, we will extend our study and analyze the effect of overexpression of CQD2, either alone or in combination with overexpression of CQD1.

    Reviewer #3:

    1. The central point of the paper is that Cqd1 is part of a novel contact site between the inner and the outer membrane. Om14 and Por1 were identified as outer membrane components of this contact site by immunoprecipitation. The data look convincing but they were generated from targeted experiments to test the involvement of suspected proteins. Ideally, one would like to see a cross-linking mass spectrometry (XL-MS) experiment that identifies the physical interactions of Cqd1 without bias.

    We thank reviewer #3 for acknowledging the presented data as convincing. Considering the significant amount of experiments planned for the revised version of the manuscript, we hope that reviewer #3 agrees that this point is not essential.

    1. Could an analogous blot of the MICOS complex be added to Figure 6D?

    Of course, we are happy to include BN-PAGE analysis showing the running behavior of MICOS next to the Cqd1 containing complex in Fig. 6D.

    1. In the Introduction, a host of contact sites is mentioned, which are partly from older papers. I'm not sure whether this is the accepted view of the field. Also, newer data suggest that the permeability transition pore is derived from complex V rather than ANT, CK, and VDAC. The authors should double check in order to represent the current state of the art

    We thank reviewer #3 for this comment. We will update this part according to the more recent literature.

  3. 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

    Khosravi et al present a comprehensive characterization of the mitochondrial protein Cqd1. They show that Cqd1 is an integral inner membrane protein that affects the mitochondrial lipid composition and that Cqd1 deletion exacerbates the delta-ups phenotype, which is also related to abnormalities in lipids. Importantly, Cqd1 is part of a large protein complex that behaves like an inner-outer membrane contact site upon sucrose density gradient centrifugation. The outer membrane proteins Por1 and Om14 were identified as likely interaction partners of Cqd1. The authors demonstrate clearly that the complex is distinct from MICOS. The data are logically presented and the paper is well organized. The results are interesting and offer a new prospective on the function of Cqd1. Although the potential involvement in lipid metabolism is not developed from the mechanistic point of view, the discovery of a new contact site between the two mitochondrial membranes is important.

    Minor critique

    1. The central point of the paper is that Cqd1 is part of a novel contact site between the inner and the outer membrane. Om14 and Por1 were identified as outer membrane components of this contact site by immunoprecipitation. The data look convincing but they were generated from targeted experiments to test the involvement of suspected proteins. Ideally, one would like to see a cross-linking mass spectrometry (XL-MS) experiment that identifies the physical interactions of Cqd1 without bias.
    2. Could an analogous blot of the MICOS complex be added to Figure 6D?
    3. In the Introduction, a host of contact sites is mentioned, which are partly from older papers. I'm not sure whether this is the accepted view of the field. Also, newer data suggest that the permeability transition pore is derived from complex V rather than ANT, CK, and VDAC. The authors should double check in order to represent the current state of the art.

    Significance

    The paper makes an important contribution to the molecular architecture of mitochondria.

    My expertise is mainly in mitochondrial lipids

  4. 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

    I have the following concerns for the authors to consider.

    1. Although biochemical evidence shows that Cqd1 is likely a factor that forms CS structures in mitochondria, it would make the manuscript stronger if the authors can observe uneven distribution of Cqd1 in the mitochondrial membranes (assessed by fluorescent microscopy or ideally high-resolution microscopy) and the presence of Cqd1 in the region of close apposition of the OM and IM by immunogold labeling for electron microscopy.
    2. Since the structural characterization of Cqd1 is important to understand its interactions with the OM proteins and other UbiB protein kinase-like family proteins, Coq8 and Cqd2, take different orientations, the membrane topology of Cqd1 should be experimentally analyzed. The authors state, "two hydrophobic stretches can be identified in the Cqd1 sequence, of which the first one (amino acids 125-142) might be a bona fide transmembrane segment" (lines 97-100); then is Cqd1 a single membrane spanning protein or two-membrane spanning protein?　　
    3. The authors state, "conserved GxxxG dimerization motif (amino acids 504‐508)" (Fig. 1A caption), but this description needs a reference. The GxxxG motif was proposed to mediate transmembrane helix-helix association (https://doi.org/10.1006/jmbi.1999.3489), which is not consistent with the membrane topology proposed by the authors.
    4. What is the role of the kinase activity of Cqd1 in the CS formation? The effects of overexpression of Cqd1 (Fig. 7) should be tested for its E330A mutant.
    5. Is there stoichiometric as well as quantitative information on the 400 kD complex consisting of Cqd1, Por1 and Om14? Does the stoichiometry and amount of the complex depend on the growth condition? Does the complex contain other Por1 interacting IM proteins like Mdm31?
    6. For Fig. 7E, the authors state, "consistently, we observed dramatically increased mitochondria‐ER interactions Cqd1 overexpression", but this observation could be due to secondary effects because overexpression of Cqd1 itself already caused abnormal morphology of mitochondria.
    7. Since the antagonistic role of Cqd2 to Cqd1 was proposed, the results of the experiments for Cqd1 can be compared with those for Cqd2. For example, what will become of overexpression of Cqd2 instead of Cqd1 for Fig. 7? What is the lipid composition of the cqd1Dcqd2D double deletion mutant cells (the decreased PA level is recovered?) ? Lines 424-425: In summary, overexpression of Cqd1 causes severe phenotypes on growth, formation of mitochondrial structural elements, and mitochondrial architecture and morphology. Is this phenotype affected by overexpression of Cqd2?

    Significance

    Mitochondrial functions rely on the formation of intramitochondrial contact sites (CS) between the outer membrane (OM) and inner membrane (IM). It is established that MICOS, involved in cristae junction formation, contributes to the formation of the CS through its interactions with the OM proteins including the SAM complex, TOM complex, Por1 etc. However, it is also recognized that CS can be formed independently of MICOS. Here Khosravi et al. report that Cqd1 in the IM could interact with Por1 and Om14 in the OM to form MICOS-independent CS. Cqd1 was previously reported to be involved in normal cellular CoQ distribution. Now Cqd1 was shown to be genetically and functionally related to the mitochondrial lipid biosynthetic pathway involving Ups1 and Crd1. Deletion of the CQD1 gene causes PA (phosphatidic acid) to decrease and overexpression of Cqd1 causes abnormal IM morphology. Most of the experiments were carefully performed and the results are properly interpreted. The present findings will extend our understanding of the mitochondria membrane architecture significantly, thereby appealing to the broad readership of the journals.

  5. 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

    Summary:

    Khosravi et al show that the protein cqd1, which was shown to export CoQ outside the mitochondria, forms a contact site by interacting with por1-om14. They conclude that the main role of this complex is to control mitochondrial architecture and phospholipid metabolism. The data shown to draw this conclusion are the effects of cqd1 overexpression altering mitochondrial morphology, as well as the exacerbation of the effects of usp1 deletion by Cqd1 deletion.

    The major concerns affecting the conclusions are:

    1. Experimental evidence is lacking on the contribution of contact site formation by Cqd1 to the effects on mitochondrial architecture and respiration-dependent growth. Determining the effects of the overexpression of the kinase-dead mutant on mitochondrial morphology and contact site formation with Por1-Om14 can address that.
    2. Related to point #1, Cqd1 overexpression in deltaUsp1 cells could have addressed whether the role of Cqd1 in contact sites and mitochondrial architecture is independent of its role on CoQ distribution and phospholipid metabolism. Further characterization of the kinase-dead Cqd1 mutant on CoQ distribution, contact sites, mitochondrial archictecture and phsophsolipid metabolism might help discerning how these activities can be separated.
    3. It is unclear how both Cqd1 overexpression and deletion induce mitochondrial fragmentation. Performing live cell imaging with a mitochondrial-phoactivatable GFP to measure mitochondrial fusion rates could help discerning the causes for fragmentation. It is a possibility that overexpression induced fragmentation by activating fission without changing fusion, while deletion induced fragmentation by blocking fusion.

    Minor comment:

    1. Figure 4 claims that mitochondrial function is impaired by ups1 deletion, which Cqd1 deletion exacerbates. However, no respiration data is shown in figure 1, only measurements of mitochondrial architecture are shown. Thus, oxygen consumption measurements are needed to claim effects on mitochondrial function.
    2. Some Western blots lack quantifications and statistical analyses of independent experiments.

    Significance

    The finding that Cqd1 forms new contact sites and interacts with Usp1 is significant and is an important advancement in our understanding of the mechanism by which Cqd1 regulates CoQ distribution. This work will be of high interest to researchers on the mitochondria field, CoQ biogenesis, and inter and intra-organellar communication.

    However, it is still unclear whether the effects observed on mitochondrial architecture are just secondary to disturbed CoQ distribution or whether they are a primary consequence of Cqd1 forming these contact sites (effects independent of CoQ distribution and lipid metabolism as concluded by the authors).

  6. Version published to 10.1101/2022.04.09.487722v1 on bioRxiv