Evolutionary divergence reveals the molecular basis of EMRE dependence of the human MCU

Using clues from evolution, chimeric proteins, and biochemical methods, this work identifies a new domain in mitochondrial calcium uniporter that determines its dependence on its binding partner EMRE.

This manuscript obviously benefitted from the Wang et al structure of human MCU in complex with EMRE, but it instead presents the studies as if they were conceived and conducted without these insights, and then comment that they are consistent with the structure. The manuscript should be re-written in a way that acknowledges the structure as providing the rationale for many of the specific mutations created, and indeed the for the ideas regarding the JML loop and its purported function and the model regarding an inner gate.
1. I don't really understand the title. Why is "evolutionary divergence" mentioned? Furthermore, the title is misleading as it implies that the regulation of MCU activity by MICU proteins is explored and explained. However, only one experiment with over-expression of a mutant MICU2 was performed, and the results (below) do not allow for unambiguous conclusions regarding the interactions of EMRE with MCU on the mechanism of regulation by MICU proteins. .an initial slow phase that lasts for fully 20 sec, and then a more rapid one over the next 20 sec. This is in contrast to reports from other labs that have observed nearly instantaneous rapid uptake that necessitates determining initial Ca2+ uptake rates immediately after Ca2+ bolus injection. This raises some concern about the measurement system used here and the meaning of the measured uptake rates. This is particularly important because the rates are compared between different cell lines, i.e. they have been interpreted quantitatively.
3. Uptake rates are determined in part, because of thermodynamics and regulation, by the bolus of free Ca injected, but we don't have that information regarding the free Ca2+ concentration at the beginning of the uptake in any of the studies. 4. Figure legend 1(F) and 1(G) are missing.
5. Membrane topology of EMRE was previously established. I understand why they might want to repeat those experiments here, but a rationale for doing these experiments should be presented right at the beginning so the reader knows that the authors want to validate previously reported data.
6. p. 5, line 5 from bottom. In addition to refs 38 and 40, Vais et al Cell Reports demonstrated that charge neutralization of EMRE C-terminus did not inhibit mitochondrial Ca2+ uptake or MCU Ca2+ currents. It should also be cited.
7. P7. The authors use chimeric MCU proteins to hone in on a 10-residue stretch that when transplanted into human MCU from Dictyostelium MCU confers EMRE-independent MCU activity. This domain is here termed the EMRE dependence domain (EDD). This is the same stretch or amino acids identified by Wang et al Cell 2019, which was termed by them as the JML. I wonder if it is wise at this time to use different terminology for the same region of MCU. Please justify.
8. Furthermore, Wang et al already previously demonstrated the requirement for this stretch of residues for EMRE dependence/independence. What is novel in this study? 9. P10, first sentence. The authors state that their functional data highlight the importance of MCU TM domains in the EMRE-dependence of MCU, but they only demonstrated the importance of the JML region.
10. P11. The authors claim to have established for the first time that MCU TM2 interacts with EMRE. However, Wang et al in their structure manuscript previously described an interaction of EMRE with TM2.
11. Regarding the DSP crosslinking experiments: a) Chimera 9 does not require EMRE for function, yet it is cross-linked to EMRE. Chimera 1 requires EMRE and crosslinks as well as chimera 9. Chimera 5 requires EMRE, but it doesn't crosslink. This is a non-coherent set of data. How do the authors interpret these seemly paradoxical results?
12. Regarding the copper crosslinking experiments... a) What was the rationale for using MCU-KO cells rather than double MCU/EMRE KO cells? The presence of endogenous EMRE in these experiments in cells into which cys-EMRE was transfected complicates interpretations about whether these mutant EMRE are still functional. b) Fig 1B demonstrates that EMRE and MCU can be co-immunoprecipitated without crosslinking. Are the conditions more stringent here (Fig 3A,B; Fig 4A-C) so only cross-linked interactions are preserved? If so, why is un-cross-linked EMRE sometimes observed? c) In transient transfections of EMRE, an unprocessed form is often the dominant species, revealed at doublets in western blots, whereas the western blots here show a single band. Is this the fullyprocessed form or not? d) What was the rationale for choosing the specific EMRE and MCU residues for cys-mutagenesis? If it was based on the hMCU/EMRE structure, this should be stated. e) What amino acids replaced the cys in the cys-less MCU? f) In Fig 3C,D, please state what antibody was used for the IP blots. g) In Fig 4C. The authors make conclusions about "flexibility in this region of the complex". What do the authors mean? How does cross-linking to more than one residue provide information about protein dynamics or stability? There is a conclusion on page 13 that EDD is "flexible", but what is the evidence? And that it blocks exit from the pore...what is the evidence? There is too much speculation without experimental evidence. h) P13. The authors suggest that they have demonstrated that the "N-terminus of EMRE is critical for the EMRE-EDD interaction", implying, based on the earlier part of the sentence, that it is critical for the EMRE-MCU interaction, but I don't think that this has been shown here. Furthermore, they state in the Discussion that there are likely additional extended interactions between MCU and EMRE 13. Fig 4E-G are interesting. Can the authors explain the rationale for choosing C. elegans? What is the sequence conservation/divergence in these regions that might account for these results?
14. I think that reference to the tetrameric JML in the Wang et al structure as "the Ca2+ exit site" of MCU is premature, even though Wang et al made the same claim. There is no evidence that Ca2+ needs to traverse this portal to enter the matrix, or that it even does. The structure is suggestive, but without functional evidence it is still an hypothesis. Perhaps the use of a word such as "putative" might be prudent. Furthermore, the authors state that chimera 9 has a constitutively open "Ca2+ gate distal to the pore", but it has not been shown that it is indeed a gate, nor that it is either constitutively closed or open under different conditions here. I think caution is warranted until there are experimental data.
15. The authors interpret lack of apparent inhibition by mutant MICU2 of MCU activity as evidence that MICU1/2 modulates the "gate distal to the pore", but lack of inhibition does not imply that MICU1/2 does not regulate Ca2+ access into the pore, i.e. there is no evidence to support this statement. Another way of thinking about the data is that the mutant chimera 9 channel modulates the channel structure in a way that impinges on MICU1/2 regulation. Additional experimentation would be necessary to dissect these alternatives.
17. Discussion paragraph 2. The human MCU/EMRE structure isn't mentioned at all here, whereas it's the focus of this study. Clearly, the Wang et al structure informed many of the experiments in this manuscript and the authors should acknowledge this.
18. P17. The authors state that fungal MCU channels have lower conductance than animal MCU, but what is the evidence for this...it is not present in the references cited.

Referee #2 Review
Report for Author: This manuscript reports analysis to identify the EMRE and MCU interacting domains based on the use of chimeras between human and dictyostelium MCU and partial cysteine scan. They also use one of the chimera to examine a new potential role of MICU1/2 in the regulation of the MCU pore.
Although the result are fairly clean, unfortunately, they add very little to the structure of MCU/EMRE published in May of last year. The st ruct ure provides a more det ailed informat ion and makes it quit e difficult to support publicat ion of this manuscript in this journal. The only addit ional and new findings are in Figure 5, however, they are preliminary as present ed. The lack of effect of MICU2 can be for mult iple reason such as alt ered int eract ion of the chimera wit h MICU2 (lower affinit y), alt ered int eract ion wit h MICU1 that is required for the effect of MICU2, alt ered Ca2+-dependence of Ca2+ upt ake by the chimera, among ot hers. More rigorous charact erizat ion of the chimera, including biophysical analysis of chimera pore, is needed to claim a role for MICUs on channel gat ing through int eract ion bet ween MCU-EMRE.

Referee #3 Review
Report for Author: Mitochondrial calcium uptake is critical for mitochondrial and cellular function. The uptake route relies on the channel MCU which interacts with numerous regulatory/accessory proteins. One of these is the protein EMRE which interestingly appears not to be co-conserved with MCU. Still EMRE is critical for MCU function in organisms containing EMRE. This implies that there might be features in (human) MCU that are important for its functional dependence on EMRE. In this study, the authros set out to identify these regions using chimeras of EMRE-dependent and EMREindependent MCU variants. They thereby identify a region in MCU critical for EMRE interaction and characterize it further biochemically. This is a very nice piece of high-quality biochemical work. My major concern affects the conceptual novelty of the findings: the interaction site of EMRE and MCU has already been previously identified in MCU (Wang et al) -although in a slightly different amino acid window. The current study expands on this but does not significantly address the in my opinion critical and really novel question of the evolutionary and regulatory significance of EMRE.
Further points: Fig. 2: complementation studies with MCU chimeras were performed in either MCU or EMRE single KO cells. Why did the authors not use MCU-EMRE double knockout cells to exclude e.g. interference of endogenous MCU with chimeras in EMRE KO cells? Fig. 4: Cysteine crosslinking data (interaction between far-away residues) -the authors interpreted their findings as "flexible nature" of the interaction. How can they exclude artefacts from the experiment? What is the functional relevance of this flexible interaction? Is the interaction dynamic? Figure 5: MICU1 binding to MCU does not result in closing of chimera 9. This implies that EMRE is hierarchically more important for the control of opening and closing. However, compared to MICU1, its binding to MCU appears not to be dynamically regulated? Can the authors further elucidate on this and hypothesize how the non-dynamic binding of EMRE helps in regulating MCU activity.
p.3, 2nd paragraph: "Intermembrane" should read "intermembrane space" Thank you for transferring your manuscript entitled "Evolutionary divergence reveals the molecular basis of EMRE dependence of the human mitochondrial calcium uniporter and its regulation by MICU1/2" to Life Science Alliance. Your manuscript was reviewed at another journal before, and the editors transferred those reports to us with your permission.
The reviewers who evaluated your study elsewhere noted that your data are of high quality and robustness. However, they would have expected a further reaching conceptual advance. This does not preclude publication here, and I would thus like to invite you to submit a revised version of your manuscript to us. Please provide a point-by-point response to all reviewer comments and accordingly change the data representation and discussion, also leaving room for alternative explanations where needed given the input received. No further experiments are needed, so we assume that the revision is feasible despite the current SARS-CoV-2 pandemic and the associated difficulties many labs experience.
Additionally, please pay attention to the following: -please add the number of replicates to the figure legends where currently missing -please upload all figure, including S figures, as individual files -please provide your manuscript file in docx format-please include the figure legends, including the supplementary figure legends, in the main manuscript docx file -please add the info to Fig 1F and G to the legend -please list 10 authors et al in your reference list -please fill in all mandatory fields in our submission system and make sure that the author order in our system matches the one in your manuscript file -please make sure that all corresponding authors link their ORCID iD to their profile in our submission system, they should have received an email with instructions on how to do so If you are planning a press release on your work, please inform us immediately to allow informing our production team and scheduling a release date.
To upload the final version of your manuscript, please log in to your account: https://lsa.msubmit.net/cgi-bin/main.plex You will be guided to complete the submission of your revised manuscript and to fill in all necessary information. Please get in touch in case you do not know or remember your login name.
To avoid unnecessary delays in the acceptance and publication of your paper, please read the following information carefully.

A. FINAL FILES:
These items are required for acceptance.
--An editable version of the final text (.DOC or .DOCX) is needed for copyediting (no PDFs).
--High-resolution figure, supplementary figure and video files uploaded as individual files: See our detailed guidelines for preparing your production-ready images, http://www.life-sciencealliance.org/authors --Summary blurb (enter in submission system): A short text summarizing in a single sentence the study (max. 200 characters including spaces). This text is used in conjunction with the titles of papers, hence should be informative and complementary to the title. It should describe the context and significance of the findings for a general readership; it should be written in the present tense and refer to the work in the third person. Author names should not be mentioned.

B. MANUSCRIPT ORGANIZATION AND FORMATTING:
Full guidelines are available on our Instructions for Authors page, http://www.life-sciencealliance.org/authors We encourage our authors to provide original source data, particularly uncropped/-processed electrophoretic blots and spreadsheets for the main figures of the manuscript. If you would like to add source data, we would welcome one PDF/Excel-file per figure for this information. These files will be linked online as supplementary "Source Data" files. **Submission of a paper that does not conform to Life Science Alliance guidelines will delay the acceptance of your manuscript.** **It is Life Science Alliance policy that if requested, original data images must be made available to the editors. Failure to provide original images upon request will result in unavoidable delays in publication. Please ensure that you have access to all original data images prior to final submission.** **The license to publish form must be signed before your manuscript can be sent to production. A link to the electronic license to publish form will be sent to the corresponding author only. Please take a moment to check your funder requirements.** **Reviews, decision letters, and point-by-point responses associated with peer-review at Life Science Alliance will be published online, alongside the manuscript. If you do want to opt out of having the reviewer reports and your point-by-point responses displayed, please let us know immediately.** Thank you for your attention to these final processing requirements.
Thank you for this interesting contribution, we look forward to publishing your paper in Life Science Alliance.
Sincerely, Andrea Leibfried, PhD Executive Editor Life Science Alliance Meyerhofstr. 1 69117 Heidelberg, Germany t +49 6221 8891 502 e a.leibfried@life-science-alliance.org www.life-science-alliance.org Authors' Response to Reviewers July 13, 2020 We thank the reviewers for careful reading and critique of our manuscript. We addressed all of their concerns below.

Referee #1:
This manuscript explores the interaction of EMRE with MCU to attempt to understand why animal MCU requires EMRE as an essential component of the channel complex for its function. The authors employ a chimera approach of using EMRE-dependent and independent MCUs to hone in on a region of MCU that is required for EMRE regulation of MCU activity. This stretch of residues is essentially the same as that previous described by Wang  This manuscript obviously benefitted from the Wang et al structure of human MCU in complex with EMRE, but it instead presents the studies as if they were conceived and conducted without these insights, and then comment that they are consistent with the structure. The manuscript should be re-written in a way that acknowledges the structure as providing the rationale for many of the specific mutations created, and indeed the for the ideas regarding the JML loop and its purported function and the model regarding an inner gate.

Response:
We disagree with this comment that "this manuscript obviously benefitted from the Wang et al structure of human MCU." This is not at all true --the reviewer is obviously not aware of the bioRxiv version of our manuscript (published on May 14 2019) It is notable that our functional mutagenesis and crosslinking data is largely concordant with what has been published in the structure, which is important for the field. We will note that the EDD domain we have identified shows some, but not complete, overlap with the JML, such that our functional studies here are consistent with the reported structure. We report eight crosslinking events, of which six seem plausible in the context of the structure. We acknowledge now in the revised manuscript discussion that two of these cannot be explained by the structure, and hence, future experiments will be required to reconcile them and establish their physiological relevance. They appear as robust as the other structure validated crosslinks. We acknowledge in the revised discussion that these represent new features of the channel complex we don't understand, though we can't exclude the possibility of an artifactual crosslinking event. Finally, there are numerous other experimental nuggets we have included that are important data points for the field.
In the revised version of the manuscript that has been transferred to LSA, we stressed that fact that our bioRxiv submission coincides with MCU-EMRE structure paper (Wang et al). Our manuscript is extremely valuable to the field since the functional mutagenesis and cross-linking experiments were performed in a manner not biased by the structure. Hence, it can be viewed as an independent set of experiments. Crucially the Wang et al paper utilized the same experimental expression system that our group first pioneered and proposed, the HEK-293T cell system for expressing and purifying the uniporter (Sancak Science 2013). We also explain that the structure shown by Wang et al does not provide the rationale for any of the specific mutations created: based on the structure, at least two the residues that we show to crosslink, should not crosslink. This could be due to multiple reasons: (a) structure shows only one confirmation of MCU/EMRE complex (b) resolution of the structure is low (uncertainly of side chain assignments are particularly high) in the MCU transmembrane 2 and EMRE interaction surface, or (c) as we acknowledge, there can be artifactual cross-linking.
1. I don't really understand the title. Why is "evolutionary divergence" mentioned? Furthermore, the title is misleading as it implies that the regulation of MCU activity by MICU proteins is explored and explained. However, only one experiment with over-expression of a mutant MICU2 was performed, and the results (below) do not allow for unambiguous conclusions regarding the interactions of EMRE with MCU on the mechanism of regulation by MICU proteins.
Response: We used "evolutionary divergence" in the title to highlight the fact that we used systematically investigated MCU from human (requiring EMRE) and from Dicty (not requiring EMRE), to understand EMRE dependence. This divergence enabled us to map a region in MCU protein that reveals its EMRE-dependence. In the revision we have removed the final experiment with MICU2 (which was deemed too preliminary) and have revised the title so it is congruent with our original bioRxiv paper. However, this does retain the "evolutionary divergence" which is crucial as indicated above. Fig 1A and Supp Fig 5 are unusual is demonstrating 2 phases..an initial slow phase that lasts for fully 20 sec, and then a more rapid one over the next 20 sec. This is in contrast to reports from other labs that have observed nearly instantaneous rapid uptake that necessitates determining initial Ca2+ uptake rates immediately after Ca2+ bolus injection. This raises some concern about the measurement system used here and the meaning of the measured uptake rates. This is particularly important because the rates are compared between different cell lines, i.e. they have been interpreted quantitatively.

The Ca2+ uptake curves shown in
Response: Our calcium uptake traces are linear during 20-50 sec after calcium injection. In the manuscript, we calculate calcium uptake rates using data from 20-30 second after calcium injection, and the relative calcium uptake rates are comparable whether we use 20-30 or 20-50 seconds after injection (please see the figure below).
We do not see a rapid, fast uptake due to two reasons: (a) We use a plate reader for calcium uptake experiments and there's ~3-5 sec delay between calcium injection, shaking the plate and starting the measurement again. (b) We use Oregon Green Bapta 6F in our experiments, which has ~3µM affinity for calcium. We inject ~35µM of free calcium. The dye remains "saturated with calcium" and the signal intensity does not change dramatically until free extramitochondrial calcium concentration reaches lower levels, giving the impression that there's a slow uptake at the beginning. Like every experiment, this experimental readout is a function of the reagents used. If we had used a calcium dye with high affinity for calcium (for example Fluo4 with a Kd of ~350nM for calcium), this curve would look much different, almost as if there's no calcium uptake for the first half of the experiment.
Our curves may look different from some other data, for example this data pasted below from PMID: 23101639 that also injects 50µM of calcium to reach ~30µM of free extramitochondrial calcium due to several reasons, 3 of which are listed below: In all our experiments, we report calcium uptake rates relative to wild type HEK293T cells that were processed together with the experimental samples. When we compare calcium uptake rates of wild type cells across time (see table below), we do not see a significant variation in rates (a ~10% variation between samples processed on different days). This enables comparison of calcium uptake rates between cell lines since they are reported as "relative to wild type". Consequently, our calcium measurements are robust and report uniporter function under the same experimental conditions for a set of chimeric proteins and mutants. We published several papers using this assay and all of our findings have been reproducible. In addition, several publications use similar relative calcium uptake rates, including Wang et al.

Uptake rates are determined in part, because of thermodynamics and regulation, by the bolus of free Ca injected, but we don't have that information regarding the free Ca2+ concentration at the beginning of the uptake in any of the studies.
Response: As mentioned above, we report calcium uptake rates relative to a WT control in all of the experiments and this relative rate is a unitless measurement. Free calcium concentration after 50µM calcium injection is ~ 35 µM, however, this number is irrelevant for our rate comparison.

Figure legend 1(F) and 1(G) are missing.
Response: Thank you for catching this mistake. They have been added.

Membrane topology of EMRE was previously established. I understand why they might want to repeat those experiments here, but a rationale for doing these experiments should be presented right at the beginning so the reader knows that the authors want to validate previously reported data.
Response: We added the following sentence that indicates the rationale for the experiment in the main text.

p. 5, line 5 from bottom. In addition to refs 38 and 40, Vais et al Cell Reports demonstrated that charge neutralization of EMRE C-terminus did not inhibit mitochondrial Ca2+ uptake or MCU Ca2+ currents. It should also be cited.
Response: We apologize for unintentional omission of Vais et al. This paper is now been added as refence.

PThe authors use chimeric MCU proteins to hone in on a 10-residue stretch that when transplanted into human MCU from Dictyostelium MCU confers EMRE-independent MCU activity. This domain is here termed the EMRE dependence domain (EDD). This is the same stretch or amino acids identified by Wang et al Cell 2019, which was termed by them as the JML. I wonder if it is wise at this time to use different terminology for the same region of MCU. Please justify.
Response: EDD only covers 3 of 6 amino acids of JML as shown by the alignment below. We show crosslinking between MCU and EMRE in a region of EDD that does not overlap with JML (E293 and D296). In -FuraFF with a Kd of ~6µM for calcium is used, as opposed to Oregon Green Bapta 6F, with a Kd of 3µM for calcium -The experiment is carried out in a well stirred cuvette, which allows for better mixing, oxygenation of the sample and simultaneous monitoring of signal -6 million cells are used here, we use 1 million cells in our experiments, which will lead to approximately 6 times faster reduction of extramitochondrial calcium levels.
addition, our nomenclature stems from a functional readout and is unique to MCU. JML annotation is not unique to MCU, it's a structural region that many other proteins have, and does not have a function associated with it. We would like to keep EDD nomenclature because (1) it's not identical to JML (2) it's not a structural region but encompasses two structural domains (JML and adjacent coiled coil 2 of MCU) (3) communicates the function of the domain.

Furthermore, Wang et al already previously demonstrated the requirement for this stretch of residues for EMRE dependence/independence. What is novel in this study?
Response: Again this referee is not aware of the fact that the pre-print version of our paper was published 5 days after the publication of Wang et al. They are highly complementary to each other, ours is rich with functional mutagenesis, chimera analysis and physiology, while theirs is largely structures. Novel findings in our paper is summarized here: (a) We identify EDD, which is overlapping but distinct from JML. Comparison of functional data between different chimeras (compare chimera 9 to chimeras 6 and 7) clearly show the importance of residues outside of JML for EMRE function.
(b) We show more extensive interactions between EMRE and MCU pore forming TM2 that were not detected in the structure. Wang et al claims that EMRE and MCU interact only at TM1 and around the JML in TM2 (around MCU amino acids 280-300). We find robust and reproducible interaction between MCU and EMRE at amino acid 270, which is closer to the intermembrane space. We now discuss in the manuscript that these differences can be due to technical issues or due to the dynamic nature of the complex in vivo. These results also caution the community against taking the structure as the only and ultimate confirmation for MCU-EMRE complex; argues for the presence of other, functionally important MCU-EMRE interactions in vivo; and makes a strong case for additional functional experiments to understand how this interesting calcium channel functions at the molecular level in vivo.
(c) We provide data for the first time that interactions between EMRE N-terminal domain and MCU EDD are important for both binding and function of the channel (Figures 4F and 4G). Previous studies suggested that MCU TM1 and TM2 are the determinants, but our data clearly show that even though TM domains of MCU and EMRE are wild type sequences from the same species, incompatibility at EMRE's N-terminal domain and MCU's EDD lead to loss of binding between the two.

P10, first sentence. The authors state that their functional data highlight the importance of MCU TM domains in the EMRE-dependence of MCU, but they only demonstrated the importance of the JML region.
Response: We acknowledge this assertion was made only indirectly in our manuscript, and regret it was not presented more clearly. Comparison of DdMCU and Chimera 4 functional data demonstrate that the presence of DdMCU EDD alone does not ensure EMRE independence (see figure below). Rather, based on chimera 4's EMRE dependence, the transmembrane domains also play key roles in EMRE dependence of HsMCU. We now added the sentence below in the main text to make this conclusion clearer. "Chimera 4 showed reduced, but EMRE-dependent uniporter activity ( Figure 2C), suggesting that the TM domains of HsMCU are involved in its EMRE-dependence."

P11. The authors claim to have established for the first time that MCU TM2 interacts with EMRE. However, Wang et al in their structure manuscript previously described an interaction of EMRE with TM2.
Response: We apologize we forgot to change this sentence in our manuscript submission. This claim was rightfully made when we first wrote the manuscript before the structure was published and was in the bioRxiv submission and was kept in this submission unintentionally. We now refer to Wang et al as well and add the following sentence in the main text: "Collectively, these findings confirm the interaction that was observed between MCU TM1 and EMRE previously (38), and establish that the pore-forming TM2 of MCU also interacts with EMRE, both within and outside the MCU-EMRE interaction domain defined by the MCU-EMRE structure (Wang et al)"

Regarding the DSP crosslinking experiments: a) Chimera 9 does not require EMRE for function, yet it is cross-linked to EMRE. Chimera 1 requires EMRE and crosslinks as well as chimera 9. Chimera 5 requires EMRE, but it doesn't crosslink. This is a non-coherent set of data. How do the authors interpret these seemly paradoxical results?
Response: Chimera 9 does not require EMRE for function, yet it is cross-linked to EMRE -Correct Chimera 1 requires EMRE and crosslinks as well as chimera 9-Correct Chimera 5 requires EMRE, but it doesn't crosslink-Incorrect. Chimera 5 does not require EMRE to function. It is EMRE-independent ( Figure 2C). As a result, these results are not paradoxical.
We also would like to stress the fact that EMRE-binding and EMRE-dependence are independent from each other, and binding does not necessitate dependence in function (as in Chimera 9).

Regarding the copper crosslinking experiments... a) What was the rationale for using MCU-KO cells rather than double MCU/EMRE KO cells? The presence of endogenous EMRE in these experiments in cells into which cys-EMRE was transfected complicates interpretations about whether these mutant EMRE are still functional.
Response: When we started these experiments, we did not have access to MCU/EMRE double KO cells. We agree with the reviewer that it would have been better to check function of every cys-EMRE residue expressed in EMRE KO cells. In unpublished results, we performed an alanine mutagenesis screen of EMRE amino acids that are in its transmembrane domain (EMRE aa 65-85). All of the alanine substitutions, except for amino acids 69 and 81, did not disrupt EMRE function. Although these are cysteine substitutions, we expect them to be functional similar to alanine mutations.
b) Fig 1B demonstrates that EMRE and MCU can be co-immunoprecipitated without crosslinking. Are the conditions more stringent here (Fig 3A, B; Fig 4A- Response: We apologize for leaving this important point from materials and methods. In Figure 1B, 0.2% DDM is used. This milder detergent preserves MCU-EMRE interaction however recovery of the protein complex from the membrane is not complete. In cysteine and DSP crosslinking experiments, we use 1% Triton X-100 during lysis to increase recovery of the complex from the membrane and increase yield. We clarified this in materials and methods section now.

c) In transient transfections of EMRE, an unprocessed form is often the dominant species, revealed at doublets in western blots, whereas the western blots here show a single band. Is this the fully-processed form or not?
Response: In our transient transfection experiment (Figure 3A), we do not observe an unprocessed EMRE band (see below uncropped western blots for this experiment). The presence of unprocessed band really depends on the extent of overexpression, suggesting that under our experimental conditions, EMRE overexpression is not high enough to overwhelm the mitochondrial protein processing machinery.

d) What was the rationale for choosing the specific EMRE and MCU residues for cys-mutagenesis? If it was based on the hMCU/EMRE structure, this should be stated.
Response: As we mentioned above, the majority of these crosslinking experiments were done before the structure paper came out. We apologize we did not make it clear how we decided to focus on these residues. We followed the following workflow to identify residues that crosslink, which can be briefly described as a partial cysteine-scanning strategy.
(a) After making a cysteine-free MCU construct, we selected one MCU amino acid residue at a time to be the sole cysteine residue. Several MCU residues on both transmembrane domains were experimentally tested to determine their ability to crosslink to EMRE. We confirmed that these cysteine mutations did not alter MCU protein function. (b) We "scanned" several individual EMRE residues that had been mutated to cysteine for their ability to crosslink to the one MCU cysteine residue. Instead of testing all EMRE transmembrane residues for their ability to crosslink to the one MCU cystine residue, we prioritized the EMRE residues that are likely to be in the same depth in the membrane as the selected MCU residue. We "guessed" their relative depth by the knowledge that a helical turn within the membrane is 3-4 amino acids long. We did not mutate EMRE amino acids 69 and 81 to cysteines as mutation of aa69 or aa81 to alanine caused loss of EMRE function and we predicted that mutating them to cysteine will have a similar effect.

e) What amino acids replaced the cys in the cys-less MCU?
Response: We apologize for omitting this sequence. It is now added to the manuscript. MCU has 2 cysteine residues (at amino acids 26 and 33) before the predicted MTS (ending at 56aa), these were mutated to glycine residues and did not affect protein localization. It has 3 more cysteines at aa 66, 97 and 191. C66 is an evolutionarily conserved residue. It was mutated to a Y. We decided not to mutate this residue to an S, as C66 is next to an ST sequence. C97 is an evolutionarily conserved residue. It was mutated to a G. S92 residue nearby was shown to be phosphorylated. To avoid phosphorylation of this residue by a nearby kinase we chose to mutate C97 to a G. C191 was mutated to an R due to presence of an R at this position in Xenopous. This indicated that an R at this position is likely to be tolerated. Fig 3C,D, please state what antibody was used for the IP blots.

Response:
We now indicate that these are FLAG IPs in these figures. All other experimental details are also noted in the figure legends.
g) In Fig 4C. The authors make conclusions about "flexibility in this region of the complex". What do the authors mean? How does cross-linking to more than one residue provide information about protein dynamics or stability? There is a conclusion on page 13 that EDD is "flexible", but what is the evidence? And that it blocks exit from the pore...what is the evidence? There is too much speculation without experimental evidence.
Response: As the reviewer stated, cross-linking of one residue of MCU to more than one residue of EMRE does not provide information about protein dynamics or stability overall. Such crosslinking data can be explained by two possible scenarios (using MCU E296C data as example): (1) MCU E296C changes its relative localization/ orientation depending on conditions (2) MCU E296C has two static confirmations on two different MCU peptides.
In this manuscript, we favored the first scenario, because CeMCU NMR structure and high-resolution structures of fungal MCU homologs (PMID: 29954988, 29995855, 29995857) showed that the region that coincides with EDD is disordered. We now add these fungal MCU structures with their "EDD" domains highlighted as a new supplementary figure (Supplementary Figure 7). And we change the main text to state that our flexibility claim comes from these fungal structures and our crosslinking data would be consistent with this. "To understand the significance of EMRE-EDD interaction for uniporter function, we identified the homologous EDD region in four fungal MCU homologs whose high-resolution structures have been published and highlighted EDD in these structures (Supplementary Figure 7). Surprisingly, EDD appeared flexible in these structures. Based on this observation and our crosslinking data, we posit that binding of EMRE stabilizes this otherwise flexible region and allows exit of Ca 2+ ions from the pore".

Supplementary Figure 7: The region homologous to EDD appears disordered in fungi
High-resolution structures of MCU tetramers from four different fungal species are shown. The region that is homologous to EDD or amino acids that surround the EDD in these structures are shown in red. Dotted blue lines show amino acids that are omitted in these structures, all of which overlap with EDD. In the M. acridum MCU structure, one MCU chain shows alpha helical EDD, whereas the same region in the neighboring chain is not structured. PDB IDs of these structures are as follows: M. acridum (6C5W); N. crassa (5KUJ); N. fischeri  Figures 4E, F and G, we "take away" the N terminus of HsEMRE by swapping it with that of CeEMRE, and this eliminates HsEMRE-HsMCU interaction, suggesting that this N terminus region of EMRE is required for EMRE-MCU interaction. When CeEDD is introduced to HsMCU to "put back" a binding partner for CeEMRE N terminus, this restores the lost interaction. These experiments suggest that N terminus of EMRE binds to EDD. This conclusion is corroborated by cysteine crosslinking between EMRE N terminus residues aa62 and aa59 with EDD. We re-wrote this section in the manuscript.
"Previous studies suggested that a small portion of N-terminal domain of EMRE is dispensable for MCU-EMRE interaction (38). However, in contrast, our crosslinking data show that the N-terminus of EMRE directly interacts with MCU. To determine the importance of this region for uniporter function and the stability of EMRE-MCU interaction, we generated chimeric proteins using HsMCU, HsEMRE, C. elegans MCU (CeMCU) and C. elegans EMRE (CeEMRE) as shown in Figure 4E." 13. Fig 4E-G are interesting. Can the authors explain the rationale for choosing C. elegans? What is the sequence conservation/divergence in these regions that might account for these results?

Response:
We choose C.elegans as we knew that from our own unpublished data and from Tsai et al, Elife that C.elegans MCU (CeMCU) and CeEMRE will form a functional channel when expressed in MCU KO HEK 293T cells and their cDNAs were readily available. We also knew that (unpublished data) CeMCU does not bind to HsEMRE, making this system a good model to ask whether EMRE-EDD interaction is critical for EMRE-MCU interaction.
In the revised manuscript, we explained our rationale better by adding the following section We also provide sequence alignment of CeMCU-HsMCU and CeEMRE-HsMCU and highlight the sequences that were swapped (Supplementary Figure 8).
"Previous studies suggested that a small portion of N-terminal domain of EMRE is dispensable for MCU-EMRE interaction (38). However, in contrast, our crosslinking data show that the N-terminus of EMRE directly interacts with MCU. To determine the importance of this region for uniporter function and the stability of EMRE-MCU interaction, we generated chimeric proteins using HsMCU, HsEMRE, C. elegans MCU (CeMCU) and C. elegans EMRE (CeEMRE) as shown in Figure 4E. When expressed in MCU KO cells, CeMCU and CeEMRE form a functional channel, but HsMCU and CeEMRE are not compatible (38). Sequence alignment of MCU and EMRE from human and C. elegans are shown in Supplementary Figure 8. This system allowed us to test whether EDD and N terminus of EMRE contribute to MCU-EMRE interaction and channel function. HsEMRE with CeEMRE N terminal domain (HsEMRE CeNterm ) did not bind to HsMCU. When HsEMRE CeNterm was expressed with HsMCU with CeEDD (HsMCU CeEDD ), the two proteins interacted and supported mitochondrial Ca 2+ uptake (Figures 4F and 4G). We conclude that EDD and EMRE N-terminus interactions are necessary to form a stable association between MCU and EMRE and to form a functional pore."