Mechanism of Iron–Sulfur Cluster Assembly: In the Intimacy of Iron and Sulfur Encounter
Round 1
Reviewer 1 Report
This review focus on the biosynthesis of a 2Fe-2S cluster in the ISC, SUF and NIF machineries. The authors describe the mechanism of assembly of the [2Fe-2S] cluster detailing how each players of the machineries participates in the 2Fe-2S cluster biosynthesis. The review is well conceived and clearly written. Each step of the 2Fe-2S cluster biosynthesis is well documented and the readers can clearly have a molecular view on what is well defined in the mechanism of the 2Fe-2S cluster biosynthesis and what still needs to be investigated. I appreciated the efforts of the authors to include all the different models of 2Fe-2S cluster biosynthesis that have been proposed over the last decades until reaching the most recent one that is most physiologically relevant and well summarized in Figure 6. In conclusion, my view on this piece of work is highly positive and I have few suggestions to improve it, here reported:
1. At pg. 2, line 76, and pg. 9, line 297: The authors mention that a reductive coupling of [2Fe-2S] clusters occurs to form a [4Fe-4S] cluster. I agree with this model but the authors miss to describe the most recent literature data on this mechanism, J Am Chem Soc. 2017 Jan 18;139(2):719-730 and J Am Chem Soc. 2014 Nov 19;136(46):16240-50. In these two papers indeed a detailed mechanism of [4Fe-4S] cluster assembly was proposed.
2. Figure 1 is never mentioned in the text.
3. Section 3.2.2: the authors detailed the mechanism of cysteine desulfurization to generate persulfide on cysteine desulfurase. The authors detailed the chemical reactions occurring in the latter process. I suggest to add a figure here reporting the chemical reactions to more clearly have the message. Moreover, the authors report that L-cysteine is oxidized during desulfurization but which is the species reduced by the oxidation mechanism is not described.
4. At pg. 11, line 387: the authors describe that the transfer of persulfide is a metal-dependent process. The latter was proposed by the authors as an iron-first model (ref 23). However, in the same page at line 375, the authors described already an iron-first model previously proposed (ref 62). I do not understand how these two papers both proposing an iron-first model are related. This should be clarified.
5. At pg. 15, line 531: the authors mentioned that glutaredoxins are involved in Fe-S cluster transfer. The authors should also describe and reference the following papers which detailed the mechanism of how Grxs can receive [2Fe-2S] cluster from [2Fe-2S] ISCU and transfer it to ISCA proteins, Proc Natl Acad Sci U S A. 2014 Apr 29;111(17):6203-8 and J Am Chem Soc. 2012;134(37):15213–15216.
Author Response
This review focus on the biosynthesis of a 2Fe-2S cluster in the ISC, SUF and NIF machineries. The authors describe the mechanism of assembly of the [2Fe-2S] cluster detailing how each players of the machineries participates in the 2Fe-2S cluster biosynthesis. The review is well conceived and clearly written. Each step of the 2Fe-2S cluster biosynthesis is well documented and the readers can clearly have a molecular view on what is well defined in the mechanism of the 2Fe-2S cluster biosynthesis and what still needs to be investigated. I appreciated the efforts of the authors to include all the different models of 2Fe-2S cluster biosynthesis that have been proposed over the last decades until reaching the most recent one that is most physiologically relevant and well summarized in Figure 6. In conclusion, my view on this piece of work is highly positive and I have few suggestions to improve it, here reported:
Answer: We thank the reviewer for her/his positive comments. To facilitate the review, we refer here to the manuscript without track change for the lines that have been modified.
1. At pg. 2, line 76, and pg. 9, line 297: The authors mention that a reductive coupling of [2Fe-2S] clusters occurs to form a [4Fe-4S] cluster. I agree with this model but the authors miss to describe the most recent literature data on this mechanism, J Am Chem Soc. 2017 Jan 18;139(2):719-730 and J Am Chem Soc. 2014 Nov 19;136(46):16240-50. In these two papers indeed a detailed mechanism of [4Fe-4S] cluster assembly was proposed.
Answer: We agree with the reviewer that these references were missing, they have been added now. In addition, the most recent report by the Lill-Muhlenhoff group (Weiler PNAS 2020) involving FDX2 in the reductive coupling is cited too (line 75 and line 130).
2. Figure 1 is never mentioned in the text.
Answer: Figure 1 was indeed mentioned page 4 line 135 in the original version. Several other occurrence have been added in section 3 (lines 123, 130, 133, 134).
3. Section 3.2.2: the authors detailed the mechanism of cysteine desulfurization to generate persulfide on cysteine desulfurase. The authors detailed the chemical reactions occurring in the latter process. I suggest to add a figure here reporting the chemical reactions to more clearly have the message. Moreover, the authors report that L-cysteine is oxidized during desulfurization but which is the species reduced by the oxidation mechanism is not described.
Answer: A figure presenting the mechanism of cysteine desulfurization has been added including the oxidation state of the relevant atoms that are modified during the reaction (Fig. 4). The species that is reduced during the reaction is the carbon of the desulfurized cysteine that goes from –I in L-cysteine to –III in L-alanine. Thereby, the reaction is equilibrated with two sulfurs at –II and a carbon at –I at start, leading to two sulfurs at –I and a carbon at –III. This is now mentioned at line 386.
4. At pg. 11, line 387: the authors describe that the transfer of persulfide is a metal-dependent process. The latter was proposed by the authors as an iron-first model (ref 23). However, in the same page at line 375, the authors described already an iron-first model previously proposed (ref 62). I do not understand how these two papers both proposing an iron-first model are related. This should be clarified.
Answer: In Ref 62, Cowan and co-workers just mentioned the iron first mechanism in the text, without showing the actual experiment. Our data thus provide an experimental validation of this model. This has been clarified at line 406.
5. At pg. 15, line 531: the authors mentioned that glutaredoxins are involved in Fe-S cluster transfer. The authors should also describe and reference the following papers which detailed the mechanism of how Grxs can receive [2Fe-2S] cluster from [2Fe-2S] ISCU and transfer it to ISCA proteins, Proc Natl Acad Sci U S A. 2014 Apr 29;111(17):6203-8 and J Am Chem Soc. 2012;134(37):15213–15216.
Answer: We agree with reviewer, these references have been added line 579
Reviewer 2 Report
This is a very detailed review covering the early steps of FeS biogenesis by several systems (NIF, SUF, and ISC). Since this process is quite complex and challenging to study, there has been many pivots in the field as initial hypotheses about the roles of different proteins or the mechanism of the process have not been proven to be correct. There have also been significant recent developments in the field that make this a timely article. I am not aware of other recent comprehensive reviews that cover the early steps of FeS biogenesis in several systems which makes this article important for the field. However, I suggest that the manuscript requires significant changes to increase the clarity of the writing.
In general, the story is laid out in a chronological fashion, first discussing the initial studies and subsequent hypotheses, and then how additional studies called for revision of the mechanism or role of the protein. I am a little concerned that this strictly chronological presentation of the information will be confusing for a nonexpert in the field because it is not always foreshadowed that these initial ideas about the mechanism or function of the protein have been disfavored as new information arose. I suggest adding some of this information in the opening paragraphs of these sections that provide an overview in the shifting landscape before launching into the details.
For a lengthy manuscript, there are not many figures. Additional figures would enhance the manuscript, including the section about the cysteine desulfurase mechanism, the different models for iron vs sulfur first and the bridging 2Fe2S cluster vs one bound to each IscU etc (Lines 543-555), and for the SUF mechanism, etc.
For the figures that are present, additional information needs to be added to the captions. For Figure 1, the meaning of the dashed blue arrows and the solid black arrows needs to be explicitly described. For Figure 2, for caption, it would be helpful to work on wording, "Cys381 binds to the zinc ion via exchange with Cys35". It would be helpful to explicitly state that ligands for the zinc from Iscu, nfs1 and fxn are labeled in black, red and blue respectively, and describe explain the dashed lines in panel C - which presumably are hydrogen bonds but they do not appear in any of the other panels. Finally, the blue arrow used in Figures 3-5 should be described in the captions of the figures. Particularly since they are used differently in the figures. In figures 3 and 5, they are used to indicate the path of persulfide transfer whereas in Figure 4 they are mechanistic arrows showing how the sulfur attacks the persulfide during the transfer. Additionally, at Line 129, it would be helpful here to refer to the table and make it clear that the list of protein names refer to prokaryote, yeast and mammal, respectively.
The writing needs to be edited for clarity and grammar. In particular, the use of demonstrative pronouns (this, that, these those) was often confusing. For example, many paragraphs in the early section refer to what was known, "at that time", but it is not clear what time "that" is referring to. In addition, many of the sentences are overly complex with too many clauses. For example, sentences from Line 316-318, line 383-386, lines 401-403, lines 422-425, Line 597-600 are a few examples. There is also problems with subject verb agreement in terms of singular vs plural and problems with verb tense, in terms of what should be in present tense vs what should be in past tense that make the manuscript difficult to understand. There are also some word choice issues - such as the use of the word "edifices" in the opening line of the introduction, that need to be addressed. In that case, the word, "structures" or "metallocofactors" should be considered.
Author Response
This is a very detailed review covering the early steps of FeS biogenesis by several systems (NIF, SUF, and ISC). Since this process is quite complex and challenging to study, there has been many pivots in the field as initial hypotheses about the roles of different proteins or the mechanism of the process have not been proven to be correct. There have also been significant recent developments in the field that make this a timely article. I am not aware of other recent comprehensive reviews that cover the early steps of FeS biogenesis in several systems which makes this article important for the field. However, I suggest that the manuscript requires significant changes to increase the clarity of the writing.
Answer: We thank the reviewer for her/his very positive comments that have greatly contributed to the improvement of the manuscript. We have made several changes to enhance the clarity of the manuscript. To facilitate the review, we refer here to the manuscript without track change for the lines that have been modified.
1) In general, the story is laid out in a chronological fashion, first discussing the initial studies and subsequent hypotheses, and then how additional studies called for revision of the mechanism or role of the protein. I am a little concerned that this strictly chronological presentation of the information will be confusing for a nonexpert in the field because it is not always foreshadowed that these initial ideas about the mechanism or function of the protein have been disfavored as new information arose. I suggest adding some of this information in the opening paragraphs of these sections that provide an overview in the shifting landscape before launching into the details.
Answer: We thought such a chronological presentation could be informative for experts and non-experts as well, as the some of the controversies, especially on frataxin, are still topical questions. However, we agree with the reviewer that the chronological presentation could be confusing when controversies are not set against more recent assumptions and models. We have thus removed some of these occurrences (section 3.1.1, 3.5.1) and modified the other to introduce the shifting of the landscape at the beginning of the paragraph (sections 3.1.2.1, 3.2.3).
2) For a lengthy manuscript, there are not many figures. Additional figures would enhance the manuscript, including the section about the cysteine desulfurase mechanism, the different models for iron vs sulfur first and the bridging 2Fe2S cluster vs one bound to each IscU etc (Lines 543-555), and for the SUF mechanism, etc.
Answer: Six figures have been added. One showing the overall structure of the IscS/NFS1-IscU/ISCU complexes (Fig. 3), one depicting the catalytic cycle of cysteine desulfurases (Fig. 4), one about the dimerization mechanism with the bridging Fe-S cluster (Fig. 7) and a one on the mechanism of persulfide transfer in the SufS-SufU system (Fig. 12). The former Fig. 6 is now split into 3 figures (Fig. 9, 10 and 11) to include the hypothesis that the NFS1–ISCU complex dissociates and to describe in more details the thiol-based reaction, The models of iron- and sulfur-first were indeed embedded in Figure 3 (now Figure 4), they are now highlighted more clearly.
3) For the figures that are present, additional information needs to be added to the captions. For Figure 1, the meaning of the dashed blue arrows and the solid black arrows needs to be explicitly described. For Figure 2, for caption, it would be helpful to work on wording, "Cys381 binds to the zinc ion via exchange with Cys35". It would be helpful to explicitly state that ligands for the zinc from Iscu, nfs1 and fxn are labeled in black, red and blue respectively, and describe explain the dashed lines in panel C - which presumably are hydrogen bonds but they do not appear in any of the other panels. Finally, the blue arrow used in Figures 3-5 should be described in the captions of the figures. Particularly since they are used differently in the figures. In figures 3 and 5, they are used to indicate the path of persulfide transfer whereas in Figure 4 they are mechanistic arrows showing how the sulfur attacks the persulfide during the transfer. Additionally, at Line 129, it would be helpful here to refer to the table and make it clear that the list of protein names refer to prokaryote, yeast and mammal, respectively.
Answer: All these modifications have been added accordingly for arrows and dashed lines in Fig. 1, 2, 5, 6, 8, 9, 10, 11. The dashed line in Fig. 2 shows π-π stacking between His103 and Trp155. Colours have been specified for all amino acids appearing on figures. For nomenclature, we now use a double bacterial/mammal notation for general discussions as proposed by reviewer 3 and the specific nomenclature for the description of particular experiments.
4) The writing needs to be edited for clarity and grammar. In particular, the use of demonstrative pronouns (this, that, these those) was often confusing. For example, many paragraphs in the early section refer to what was known, "at that time", but it is not clear what time "that" is referring to. In addition, many of the sentences are overly complex with too many clauses. For example, sentences from Line 316-318, line 383-386, lines 401-403, lines 422-425, Line 597-600 are a few examples. There is also problems with subject verb agreement in terms of singular vs plural and problems with verb tense, in terms of what should be in present tense vs what should be in past tense that make the manuscript difficult to understand. There are also some word choice issues - such as the use of the word "edifices" in the opening line of the introduction, that need to be addressed. In that case, the word, "structures" or "metallocofactors" should be considered.
Answer: The grammar has been edited throughout the manuscript, the references to time have been avoided as much as possible and long sentences have been split for clarity. We have modified the tense when needed, so that the past tense is used to report experiments and the present for more general considerations, discussions and conclusions.
Reviewer 3 Report
This manuscript is a lengthy and quite exhaustive review of an important but very complex topic that is iron-sulfur cluster biogenesis. The authors have made a massive effort to cover most of the relevant steps leaving out only the cytoplasmic machine which is in fact the lest well known. The manuscript is overall sound but there are a number of aspects that should be clarified, corrected or changed. I am going to list here some in the order of appearance in the manuscript but not of relevance.
1) p. 2: 'Despite huge progress in the past 30 years to elucidate the respective function of each component of these Fe-S cluster assembly machineries, a central question is still how Fe-S clusters are assembled.' This is a tautology. One studies the single components to understand the overall process and knows the overall process from the functions of individual components. Rephrase.
2) same page: line 66 'in recent years' and 73 'complexity arises'
3) p. 3: 'thus that the SUF system in the house keeping assembly 102 machinery in most bacteria and archaea.' There is verb.
4) p. 3: 'In contrast, the CIA machinery does not contain a dedicated cysteine desulfurase and thus relies on other machineries for sulfur supply.' Perhaps the authors want to mention that Nfs1 is also in the cytoplasm.
5) p.3: nomenclature. The authors state that they will use mostly the bacterial nomenclature but in reality they mix it up all. Later on they refer to mouse numbering. This is very confusing. The nomenclature should be unified. In case, the authors could write IscS/Nfs1 and so forth but it is otherwise very hard to follow.
6) p. 4: table (Table 1 should be indicated). There is a strong misunderstanding here and in the literature. It is not BolA (there is no a BolB, BolC, etc.) but Bola which in Spanish mean ball. The protein was called this way because of the round morphology phenotype observed upon over-expression of the protein in E. coli. This review is the right place to clarify this point.
7) p. 6: 'This assembly site was thus proposed to be the 177 entry point for iron, but evidence for that have been tedious and contradictory until very recently.' Why tedious?
8) Figure 2: The pdb codes for all three structures must be indicated.
9) p. 7: 'We found that the zinc ion hinders iron binding in the assembly site, but upon removal of the zinc ion, the monomeric form of ISCU binds Fe2+ in the assembly site.'. I agree but the authors should at least comment on this. Do we really envisage that, once detached from IscS/Nfs1 IscU/Isu will be reloaded with zinc? How?
10) p. 7: 'a minor fraction of iron (15%) does not bind in the assembly site when mouse apo-ISCU was incubated with one equivalent of iron, which suggests the existence of a secondary binding site.23'. Frankly I do not understand this. IscU/Isu is such a small protein with only three cysteines and, when present, a histidine. To coordinate iron (or zinc) you need residues able to do it. How would the second site exist? If at all, it could only be a non-specific weak interaction. Is this what the authors have in mind?
11) same page: as the authors say a 10 time excess of iron is unrealistic. Line 242: I do not think IscU is prone to aggregation. Certainly it is prone to misfolding but it is a highly soluble protein.
12) p. 8: 'However, the number 278 of iron binding sites in frataxin varied from one study to another (from 1 to 7 iron binding sites),..'. There is a reason for this that is linked to point 10. The binding site on frataxin is purely electrostatic and thus one would not expect a definite stoichiometry. It should be clarified. Likewise: 'We reported that FXN is not able to remove the zinc ion of ISCU'. I bet it does not. How could possibly do it? No conserved cysteines or histidines. There might occasionally be surprises but proteins do not bind ions without a sequence signature.
13) p. 10: 'is designated as open conformation'. There should be a reference.
14) p. 10: 'in which the sulfur and oxyen atoms are at the same –I redox state.' I do not understand this. Why is -I? what does it mean?
15) p.11 (and p. 14): 'another studies' must be 'other studies' or 'another study'. 'hereby, two opposite models were proposed and no clear picture has emerged at that time.' Horrible sentence. Rephrase it.
16) p.11: the discussion on the iron-first and other model in Figure 3 is rather confusing. Could the authors try to simplify it and explain better? I am not sure why 'The X-ray structure of the human NFS1-ISD11-ACP-ISCU complex solved with zinc in ISCU shows that the catalytic cysteine of NFS1 is able to bind to the zinc centre through exchange with Cys35 of ISCU, which rather supports the second hypothesis'.
17) p. 13: 'Therefore, the prominent source of sulfide ions in thiol-based reconstitution is the persulfide of NFS1 not the one of ISCU.' Not sure what this means. Sulfur comes from NFS1 and goes to IscU.
18) p. 15: the story of the IscU dimer. 'dimerization is an obligatory step to form the dinuclear [2Fe2S] cluster'. I am not sure here. How do the authors imagine this dimer? in order to co-coordinate the cluster they should be head-to-tail. is this feasible at all in the NSF1/Isu complex? It is certainly impossible in the IscS/IscU complex. This point appears again on p. 18 ('Following reduction by FDX2, a mononuclear iron-sulfide [Fe-S] complex is 664 possibly formed and ISCU dimerizes to generate a transient inter-subunit [2Fe2S] cluster that would 665 quickly segregate on one monomer.').
19) p. 17: 'However, as discussed in the previous sections, the thiol-based reaction is not physiologically relevant, which means that the effects of FXN and CyaY are unrelated to Fe-S cluster biosynthesis in vivo.' This sentence gives the impression that the authors believe that frataxin is not involved at all in cluster biogenesis which is contradicted further down. They probably mean that frataxin acts on the desulfurase rather than on IscU.
20) p. 18: the authors do not discuss at all detachment of IscU from the enzyme. Why? also they talk about Fe(II) and this is reasonable but what keeps Fe(II) in this state? should this point not be mentioned?
21) p. 19: 'The sulfide ions then combine with iron, either in solution or in the assembly site, to form [2Fe2S] and [4Fe4S] clusters.'. Is the latter really formed on the IscS/IscU complex? later on the authors talk about a reductive coupling. It is more likely that this occurs after IscU is dissociated from IscS.
22) p. 23: 'There are now compelling evidence pointing to a role of frataxin as a kinetic activator of persulfide transfer.' No. The sentence must be rephrased to 'There are now compelling evidence pointing to a role of frataxin as a kinetic modulator of persulfide transfer.'
23) All throughout the ms: structuration and sulfurization: they sound to me Gallicisms. What about gaining structure and sulfuration?
Author Response
This manuscript is a lengthy and quite exhaustive review of an important but very complex topic that is iron-sulfur cluster biogenesis. The authors have made a massive effort to cover most of the relevant steps leaving out only the cytoplasmic machine which is in fact the lest well known. The manuscript is overall sound but there are a number of aspects that should be clarified, corrected or changed. I am going to list here some in the order of appearance in the manuscript but not of relevance.
Answer: We thank the reviewer for her/his positive comments that have greatly contributed to improve the manuscript. We have made several changes to simplify the manuscript and to address the points raised by the reviewer. To facilitate the review, we refer here to the manuscript without track change for the lines that have been modified.
1) p. 2: 'Despite huge progress in the past 30 years to elucidate the respective function of each component of these Fe-S cluster assembly machineries, a central question is still how Fe-S clusters are assembled.' This is a tautology. One studies the single components to understand the overall process and knows the overall process from the functions of individual components. Rephrase.
Answer: We have rephrased this sentence (line 61)
2) same page: line 66 'in recent years' and 73 'complexity arises'
Answer: These modifications have been implemented
3) p. 3: 'thus that the SUF system in the house keeping assembly 102 machinery in most bacteria and archaea.' There is verb.
Answer: “in” was replaced by “is”
4) p. 3: 'In contrast, the CIA machinery does not contain a dedicated cysteine desulfurase and thus relies on other machineries for sulfur supply.' Perhaps the authors want to mention that Nfs1 is also in the cytoplasm.
Answer: Cytoplasmic NFS1 is apparently not involved in Fe-S cluster biosynthesis, but we have modified the text to account for this cytoplasmic fraction (line 113).
5) p.3: nomenclature. The authors state that they will use mostly the bacterial nomenclature but in reality they mix it up all. Later on they refer to mouse numbering. This is very confusing. The nomenclature should be unified. In case, the authors could write IscS/Nfs1 and so forth but it is otherwise very hard to follow.
Answer: The nomenclature is a challenge here. We thank the reviewer for her/his proposition to use a double nomenclature, which we have applied for general considerations (line 127). However, a unified nomenclature is not suited when describing specific experiments; we thus prefer to keep a specific notation when experiments are described. To prevent confusion, we have checked that each time, the name of the organism is associated with the name of the proteins. We also prefer to keep the mouse/human numbering for amino acids, as it is the most well described system for now.
6) p. 4: table (Table 1 should be indicated). There is a strong misunderstanding here and in the literature. It is not BolA (there is no a BolB, BolC, etc.) but Bola which in Spanish mean ball. The protein was called this way because of the round morphology phenotype observed upon over-expression of the protein in E. coli. This review is the right place to clarify this point.
Answer: We thank the reviewer for this clarification on the etymology of the Bola protein. The spelling has been modified in Table 1.
7) p. 6: 'This assembly site was thus proposed to be the 177 entry point for iron, but evidence for that have been tedious and contradictory until very recently.' Why tedious?
Answer: This is an error, we meant “tenuous” instead. This has been modified
8) Figure 2: The pdb codes for all three structures must be indicated.
Answer: The PDB codes have been added to all old and new figures.
9) p. 7: 'We found that the zinc ion hinders iron binding in the assembly site, but upon removal of the zinc ion, the monomeric form of ISCU binds Fe2+ in the assembly site.'. I agree but the authors should at least comment on this. Do we really envisage that, once detached from IscS/Nfs1 IscU/Isu will be reloaded with zinc? How?
Answer: The sequence of metalation of ISCU is an interesting question, but there are indeed several questions here. First, we do not know whether the NFS1-ISCU complex dissociates at any step of the synthesis of Fe-S clusters and their transfer. Although we and other groups, were able to reconstitute “free” ISCU with a Fe-S cluster in vitro by using catalytic amounts of NFS1, we have preliminary data indicating that ISCU does not dissociate from the complex when both proteins are present in stoichiometric amounts. Dissociation occurs when ISCU is present in excess, so that iron-loaded ISCU is able to displace Fe-S loaded ISCU. Thus, Fe-S cluster formation can also occur on the complex. So the sequence of Fe-S cluster formation subsequently to reduction of the persulfide by FDX2 is unclear. Our data strongly suggest that the product of the reduction is a mononuclear [FeS] species that form a [2Fe2S] by dimerization of ISCU, but whether this happen within the complex or outside with “free” ISCU is unclear. We are thus now presenting two hypothesis for the formation of the [2Fe2S] cluster via internal (NFS1-assisted, Fig. 9) and external (as free protein, Fig. 10) dimerization of ISCU.
Likewise, we do not know whether zinc removal and iron insertion occurs within the NFS1-ISCU complex or with free ISCU. Since IscU/SCU still binds IscS/NFS1 in its apo form, there is no mechanism indicating that the complex dissociates in the course of Fe-S cluster synthesis. Moreover, we have unpublished data showing that the complex with apo-ISCU can be refilled with iron and is competent for Fe-S cluster formation. Thus, dissociation of ISCU is not mandatory for iron loading. Whether a zinc ion would come again in between is not known. If the role of the zinc ion is to stabilize the structured form of ISCU, the zinc ion is not needed when apo-IscU/ISCU is bound to IscS/NFS1 since IscU/ISCU retains its structured form in the complex (NMR data on E. coli IscU-IscS from the Pastore group and crystal structures of the human NFS1-ApoISCU and bacterial IscS-apo-IscU complexes). However, the HscA/B chaperone system may compete with IscS/NFS1 and thus promote dissociation of the complex to enable Fe-S transfer. In this case, a zinc ion may be needed to stabilize free IscU/ISCU. But this is all very speculative and out of the scope of the manuscript, as we focus on Fe-S cluster synthesis rather than Fe-S transfer. However, we have open the discussion on the hypothesis that iron insertion may occur either with “free ISCU or within the complex (sections 3.4 and 3.6).
10) p. 7: 'a minor fraction of iron (15%) does not bind in the assembly site when mouse apo-ISCU was incubated with one equivalent of iron, which suggests the existence of a secondary binding site.23'. Frankly I do not understand this. IscU/Isu is such a small protein with only three cysteines and, when present, a histidine. To coordinate iron (or zinc) you need residues able to do it. How would the second site exist? If at all, it could only be a non-specific weak interaction. Is this what the authors have in mind?
Answer: We agree with the reviewer, it is very unlikely that a secondary binding site exists since no other structurally defined metal binding site appears in the tri-dimensional structures of IscU/SCU and binding of iron to the assembly is correlated with Fe-S cluster biosynthesis. We were raising this possibility in the light of the XAS data of the Stemmler group which reported that ferrous iron does not bind to the assembly site but at another site. This is very intriguing, why they did not observed binding in the assembly site? However, a recent study of E. coli IscU from Barondeau’s group contradicts these data (Patra et al. JACS 2020). We have modified the text to emphasize that binding outside of the assembly site is most likely non-specific (line 244). We also attempt to explain their results by variability in protein preparations, since in our hands apo-ISCU tends to form higher order oligomer that do not bind iron. Alternatively, the XAS samples may contain a mixture of iron bound non-specifically and bound in the assembly site, which cannot be properly simulated, if one does not know the proportion of each species.
11) same page: as the authors say a 10 time excess of iron is unrealistic. Line 242: I do not think IscU is prone to aggregation. Certainly it is prone to misfolding but it is a highly soluble protein.
Answer: By aggregation we mean formation of higher order oligomers, but still soluble. In our hands, when the metal is absent, a significant portion of mouse ISCU eluates as high molecular weight oligomers on gel filtration. These oligomers are not able to acquire iron. Although this might be a specific feature of mouse ISCU, it is worth mentioning this behaviour to attempt explain the result reported by the Stemmler group. The sentence has been modified (line 246).
12) p. 8: 'However, the number 278 of iron binding sites in frataxin varied from one study to another (from 1 to 7 iron binding sites),..'. There is a reason for this that is linked to point 10. The binding site on frataxin is purely electrostatic and thus one would not expect a definite stoichiometry. It should be clarified. Likewise: 'We reported that FXN is not able to remove the zinc ion of ISCU'. I bet it does not. How could possibly do it? No conserved cysteines or histidines. There might occasionally be surprises but proteins do not bind ions without a sequence signature.
Answer: We fully agree that metal binding with high affinity requires a defined metal binding site, the text has been clarified accordingly (line 286). Due to the controversy on the role of frataxin, we used the reconstituted system to test several hypothesis linked to metal insertion, including the most unlikely ones such as zinc removal. However, we have removed this sentence since it does not make sense with the structure of frataxin, as spotted by the reviewer.
13) p. 10: 'is designated as open conformation'. There should be a reference.
Answer: The reference has been added
14) p. 10: 'in which the sulfur and oxyen atoms are at the same –I redox state.' I do not understand this. Why is -I? what does it mean?
Answer: This refers to formal oxidation states of atoms in organic molecules that is used to determine which atom is oxidized or reduced during a chemical reaction and how many electrons are exchanged. Each atom tries to satisfy the octet rules (8 electrons in their valence shell, or 2 in the case of hydrogen) by attracting electrons from their neighbours. However, an atom will gain or lose electrons depending on its electronegativity relative to the other atoms that are covalently attached to it. –I means that the atom is reduced with one electron. The sulfane sulfur of persulfide is recurrently ascribed a formal oxidation state of “0”, and thus the other sulfur would be at –II. However, in homonuclear bond, the oxidation states of the two atoms are equals since they have the same electronegativity and thus in a persulfide one sulfur cannot take electron from the other. Thereby, the determination of the oxidation state in a persulfide leads to a –I state for both sulfur. A new figure (Fig. 4) has been added which includes these oxidation states and the text includes more details (line 385).
15) p.11 (and p. 14): 'another studies' must be 'other studies' or 'another study'. 'hereby, two opposite models were proposed and no clear picture has emerged at that time.' Horrible sentence. Rephrase it.
Answer: This has been modified
16) p.11: the discussion on the iron-first and other model in Figure 3 is rather confusing. Could the authors try to simplify it and explain better? I am not sure why 'The X-ray structure of the human NFS1-ISD11-ACP-ISCU complex solved with zinc in ISCU shows that the catalytic cysteine of NFS1 is able to bind to the zinc centre through exchange with Cys35 of ISCU, which rather supports the second hypothesis'.
Answer: We have simplified these parts. In the main text, we were confusingly referring to the mechanism proposed for the SufS-SufU system, which will be described later in the text (see section 5.3.3) with a new figure (Fig. 12).
17) p. 13: 'Therefore, the prominent source of sulfide ions in thiol-based reconstitution is the persulfide of NFS1 not the one of ISCU.' Not sure what this means. Sulfur comes from NFS1 and goes to IscU.
Answer: This means that thiols react faster with the persulfide of NFS1 than with the one of ISCU. As pointed by the reviewer “sulfur comes from NFS1 and goes to ISCU”, but as thiols are poor reductant of the persulfide of ISCU (most likely due to the presence of the metal), virtually no sulfide is produced under these conditions. However, in reconstitutions assays, L-cysteine is in excess so that, after transfer of the persulfide to ISCU, NFS1 is ready for another catalytic cycle and another persulfide is formed. Since, thiols reacts faster with the persulfide of NFS1 than with the one of ISCU, the source of sulfide ions is NFS1 not ISCU, this is what our sentence means.
The Figure 4 (now Fig 6) has been modified to describe this process in more details. In contrast, with FDX2 that only reduces the persulfide of ISCU, the source of sulfide ions is ISCU not NFS1. The text has been modified too to clarify these points (line 480-485).
18) p. 15: the story of the IscU dimer. 'dimerization is an obligatory step to form the dinuclear [2Fe2S] cluster'. I am not sure here. How do the authors imagine this dimer? in order to co-coordinate the cluster they should be head-to-tail. is this feasible at all in the NSF1/Isu complex? It is certainly impossible in the IscS/IscU complex. This point appears again on p. 18 ('Following reduction by FDX2, a mononuclear iron-sulfide [Fe-S] complex is 664 possibly formed and ISCU dimerizes to generate a transient inter-subunit [2Fe2S] cluster that would 665 quickly segregate on one monomer.').
Answer: We agree with the reviewer, the dimer should be head-to-tail or arranged along a C2 axis of symmetry. The modelling of the ISCU proteins on the NFS1-ISCU complex shows that the ISCU proteins would be arranged along a C2 axis (see new Figure 3 and line 598). Thereby, formation of a dinuclear centre is theoretically possible. However, this model is highly speculative. It is also unclear whether it could happen with IscS-IscU. A discussion is now included on this matter (line 624). We thus now propose two hypothesis for the dimerization of ISCU as highlighted in Fig. 9 and Fig. 10 (line 755-757): either it is assisted by NFS1, either ISCU dissociates from the complex and dimerizes as a free protein.
19) p. 17: 'However, as discussed in the previous sections, the thiol-based reaction is not physiologically relevant, which means that the effects of FXN and CyaY are unrelated to Fe-S cluster biosynthesis in vivo.' This sentence gives the impression that the authors believe that frataxin is not involved at all in cluster biogenesis which is contradicted further down. They probably mean that frataxin acts on the desulfurase rather than on IscU.
Answer: The reviewer is right we meant that in this case, frataxin acts on the desulfurase rather than on ISCU and that the effect of frataxin on the desulfurase is not related to physiologically relevant Fe-S cluster biosynthesis. The sentence has been modified (line 698).
20) p. 18: the authors do not discuss at all detachment of IscU from the enzyme. Why? also they talk about Fe(II) and this is reasonable but what keeps Fe(II) in this state? should this point not be mentioned?
Answer: As discussed at the points 9 and 18, whether or not the NFS1-ISCU and IscS-IscU complexes dissociates at any step of the reaction is unclear at the moment, but we have now open the discussion on this matter at sections 3.4 (line 624) and 3.6 (Fig. 9 and 10 and line 755-757).
21) p. 19: 'The sulfide ions then combine with iron, either in solution or in the assembly site, to form [2Fe2S] and [4Fe4S] clusters.'. Is the latter really formed on the IscS/IscU complex? later on the authors talk about a reductive coupling. It is more likely that this occurs after IscU is dissociated from IscS.
Answer: This paragraph is a summary that refers to the thiol-based reaction described in more details at section 3.3.2. Formation of [4Fe4S] clusters within the NFS1-ISCU complex has been reported by the Barondeau’s and Ollagnier’s groups in DTT-based reconstitution assays (Colin JACS 2013, Fox Biochemistry 2015). This is apparently a specific feature of DTT compared to other thiols such as cysteine. DTT increases the rate of sulfide production but is generally poorly efficient to reduce Fe-S clusters. We thus favour the hypothesis that DTT stimulates the formation of [4Fe4S] clusters by increasing the rate of sulfide production rather than reductive coupling. Moreover, this is consistent with the mechanism of Fe-S cluster assembly by sequential accumulation of iron and sulfide ions that is proposed at section 3.4 for the thiol-based reaction (line 612-620).
22) p. 23: 'There are now compelling evidence pointing to a role of frataxin as a kinetic activator of persulfide transfer.' No. The sentence must be rephrased to 'There are now compelling evidence pointing to a role of frataxin as a kinetic modulator of persulfide transfer.'
Answer: The sentence has been modified
23) All throughout the ms: structuration and sulfurization: they sound to me Gallicisms. What about gaining structure and sulfuration?
Answer: We found only one occurrence of structuration at page 12, but this paragraph has been deleted for clarity of the manuscript. Desulfurization is the original term used in this field to describe the cysteine desulfurase reaction, it has been introduced by Denis Dean in 1994 (see ref 193 in the revised version), we thus prefer to keep this wording.
