The giant mimivirus 1.2 Mb genome is elegantly organized into a 30-nm diameter helical protein shield

Article activity feed

1. 

   Version published to 10.7554/elife.77607 on eLife

   Jul 28, 2022
2. 

   eLife

   Jun 8, 2022

   Author Response

   Reviewer #2 (Public Review):

   In this manuscript, Villalta, Schmitt, Estrozi and colleagues report their results on genome compaction in one of the most complex known viruses, the Mimivirus. This work will be of interest to a broad readership, and particularly to virologists and structural biologists. The authors describe a novel mechanism used by mimivirus to compact and package its 1.2 Mb dsDNA genome. In particular, the mimivirus genome is shown to be packed into magnificent cylinder-like assemblies composed of GMC-type oxidoreductases, presenting yet another remarkable case of enzyme exaptation. By using cryo-electron microscopy (cryo-EM) and cryo-electron tomography (cryo-ET), the authors determined the structures of such fibers in several relaxation states, which presumably represent different stages of …

   Author Response

   Reviewer #2 (Public Review):

   In this manuscript, Villalta, Schmitt, Estrozi and colleagues report their results on genome compaction in one of the most complex known viruses, the Mimivirus. This work will be of interest to a broad readership, and particularly to virologists and structural biologists. The authors describe a novel mechanism used by mimivirus to compact and package its 1.2 Mb dsDNA genome. In particular, the mimivirus genome is shown to be packed into magnificent cylinder-like assemblies composed of GMC-type oxidoreductases, presenting yet another remarkable case of enzyme exaptation. By using cryo-electron microscopy (cryo-EM) and cryo-electron tomography (cryo-ET), the authors determined the structures of such fibers in several relaxation states, which presumably represent different stages of nucleoprotein unpacking upon delivery into host cytoplasm. The authors also suggest (although do not directly visualize) that the lumen of the genomic fibers contains several viral enzymes, most notably, DNA-dependent RNA polymerase, which is necessary for cytoplasmic replication of the mimivirus. Overall, this is an important discovery, which further expands our appreciation of the "inventiveness" of viruses.

   We thank this reviewer for the positive and constructive comments. We provide now some additional data corresponding to unpublished follow up studies, we hope will help all reviewers assessing the quality and reliability of our work.

   I am not an expert on helical reconstructions and cannot evaluate the validity of the models. Thus, my specific comments will focus on aspects of the work with which I am more familiar.

   1. In light of the presented results, it is reasonable to assume that GMC-type oxidoreductases of the mimivirus are very important for the formation of functional virions. However, in a previous study (PMID: 21646533), it has been shown that the genes encoding GMC-type oxidoreductases can be deleted from the virus genome (M4 mutant) without the loss of infectivity. The M4 virions were devoid of the external fibers decorating the icosahedral capsid, but the genome was still packaged. How do the authors reconcile these results with those presented in the present manuscript? This should be addressed in the Discussion section.

   In fact, like the reviewers, we initially assumed that the GMC-oxidoreductases were essential. Now, we believe it might be premature to assume that GMC-type oxidoreductases are the only type of proteins that can be involved in the scaffolding of the Mimiviridae genomic fibers. We managed to extract the genomic fiber of M4 (the isolate without GMC oxidoreductases). The fiber also has a rod-shaped structure but protein composition analysis of the purified fiber shows that different proteins are involved in its assembly.

   We hope the reviewers will accept to reserve our finding for a following publication.

   1. The authors state that mimivirus encodes two GMC-type oxidoreductases (qu_946 and qu_143) and that both could be fitted into the electron densities. However, I could not understand whether the authors think that the fibers are heteroassemblies of both oxidoreductases or different fibers are composed of different proteins, or only one is used for fiber formation. Please clarify. In case you are not able to distinguish between the two homologs (e.g., due to limited resolution), state so explicitly.

   We cannot discriminate between the two GMC-oxidoreductases due to their close identity (69% identity, 81% similarity) and the resolution of the map. Yet we think that in most cells the qu_946 GMC-oxidoreductase is the most abundant at the time of genome packaging (from our proteomic study, between 2 and 9 times). Yet, in some cells the second GMCoxidoreductase could become the most abundant and, in that case, the genomic fiber is built using qu_143.

   1. I am slightly puzzled by the observed "ball of yarn". It is hard for me to imagine that a cylindrical container/fiber containing a continuous dsDNA genome could be bent or fragmented into bundles because this would break the protein-protein interactions holding the fiber together. In Figures 1C and S1, are these parts of the same fiber or multiple fibers coming out of one capsid? Related to this question - is there evidence (e.g., from qPCR) that Mimivirus carries a single copy of genomic dsDNA per capsid?

   We believe this reviewer should think in terms of packaging. The folded genome is packaged through two lipid membranes (the one lining the capsid interior and the one in the nucleoid) concomitantly with its wrapping by the protein shell ribbon. Thus, there is plenty of space in the nucleoid at the beginning of the packaging and the genomic fiber is gently folded inside. But as more genome needs to be packaged, this compresses the flexible fiber into the nucleoid until it is totally encased in the nucleoid. That also defines the size of the nucleoid in the icosahedral capsid. This tight packaging is exemplified in Fig 1A for instance or the AFM images of the nucleoid enclosed in P3 of this file.

   We provide a more general answer in the answers requested by the editor.

   We think that the entire genome can only be packaged in the capsid through its assembly within the protein shell. We also think the genomic fiber is progressively built on the genomic DNA while it progresses into the capsid, most likely by an energy driven packaging machinery. This process can be compared to bacterial pili assembly, except that pili are built on the surface of the cell, while the genomic fiber is built into a compartment, the nucleoid, forcing it to fold in this compartment, which is only possible due to the high flexibility of the genomic fiber. Thus, the entire genome corresponds to ~40 µm of genomic fiber, which when folded as a ball can entirely fit into the nucleoid. The organization of the genome in a large “tubular structure” and its folding inside the nucleoid compartment has been previously reported by AFM studies of the mimivirus particles (Kuznetsov, Y. G. et al. Virology 2010; Kuznetsov YG et al. J. Virol. 2013, Fig 15), which the authors refer to as “highly condensed nucleoprotein masses about 350 nm in diameter within the inner membrane sacs of virions”, with the presence of tubular structure they refer to as “thick cables of the nucleic acid” (image P3 herein).

   1. The authors describe the interactions between the monomers in the dimer of qu_946 as well as between qu_946 and DNA. I would also like to see a brief description of protein-protein interactions between subunits within the same helical strand as well as between helical strands, which hold the whole assembly together (i.e., what are the contacts between green subunits as well as between green and yellow subunits shown in Fig 2C). The authors suggest that the shell "would guide the folding of the dsDNA strands into the structure" (L310). To support this statement, the authors could show the lumen of the fiber rendered by electrostatic potential.

   We thank this reviewer for these suggestions. An additional supplementary Table (Table S4) is now provided listing the various contacting residues in each genomic fiber map and for each GMC-oxidoreductase. The number of contacts obviously decrease in the relaxed structure, but even in the compact forms, we noticed there are relatively few contacts intra and inter-strands, which may also explain the flexibility of the structure. We now provide a new figure 3 in which the lumen of the fiber is rendered by electrostatic potential for the Cl1a map and each of the two GMC-oxidoreductases.

   1. Please provide some background information on the distribution of GMC-type oxidoreductases in other families of giant viruses, so that it is clearer whether the described packaging mechanism is specific to mimiviruses or is more widespread.

   This is a central point, also linked to the question about M4. In fact, like the reviewers, we initially assumed that the GMC-oxidoreductases were essential. Now, we believe it might be premature to assume that GMC-type oxidoreductases are the only type of proteins that can be involved in the scaffolding of the Mimiviridae genomic fibers.

   If this reviewer still thinks this is essential to this manuscript we can provide a multiple alignment of the GMC-oxidoreductases of members of each clade upon request.

   Reviewer #3 (Public Review):

   Since it was presented to the scientific community as a viral entity, mimivirus has the unlimited capacity to cause surprise and admiration. In this manuscript, Villalta, Schmitt, Estrozi, et al. and Abergel present how the mimivirus gigantic genome is organized into the virion. The authors succeeded in developing a protocol to trigger virus genome uncoating followed by genome-associated proteins purification. The presented data indicates that a helical shield composed of two GMC-type oxidoreductases is associated with the mimivirus genome, named genomic fiber. By cryo-EM, and cryo-tomography different forms and stages of the genomic fiber were detailed described, indicating the dynamics of fibers conformational changes, likely related to genome packing and uncoating during the virus replication cycle. In-depth analysis of a substantial number of individual virus fibers revealed that the mimivirus genome is folded and organized inside the aforementioned helical shield, which seems to be novel among giant icosahedral viruses. Proteomics in association with image analysis indicates that mimivirus packed genome forms a channel, which accommodates key enzymes related to early phases of the replication cycle, especially RNA polymerase subunits.

   I must disclose that I am not an expert on structural virology and proteomic analysis. Therefore, I don't feel I can contribute to the improvement of this kind of analysis. That said, I congratulate the authors for their efforts to make the manuscript story understandable to nonexperts.

   We are grateful to this reviewer for these positive comments.

   I have a few suggestions and comments:

   1. Please consider the "nucleocapsid" concept during genomic fiber presentation. I believe it fits in;

   We fully agree and this was why we referred to APBV-1. Obviously, it was not clear and we now explicitly use the word “nucleocapsid” in the text.

   1. The "ball of yarn" analogy is nice, but fig 1C shows several fibers unconnected (free) in one of their ends. I am wondering if it means that the genomic fiber is not a long-single structure covering the whole genome, but a bunch of several independent helical structures covering the whole genome and attached in such "ball of yarn". Like several threads connected. Could the authors clarify that please?

   In the “ball of yarn” structures, there are clearly breaks that give the impression of multiple fibers. Yet, these breaks are due to the multiple steps of the extraction, enrichment and purification treatment. The genomic fiber is built as a long (~40 µm) single structure folded in the nucleoid while it is loaded. As a result, it is tightly packed into the nucleoid and broken into fragments upon release due to the fragilizing treatment. As exemplified in the CryoEM image provided above (P9) on freshly opened capsids, these breaks appear to depend on the treatment. This reviewer could also look at the answer we provided to Reviewer 2 point 3 as this could help clarify how it is possible to package the genomic fiber and subsequently fold it into the nucleoid to the point where it is tightly packed and under pressure.

   1. Considering previously published data on proteomics of viral factories and transcriptomics of mimivirus: is there any temporal association between GMC-type oxidoreductases' peak of expression and genome replication during the viral cycle? what about RNA pol subunits? Are all those proteins highly expressed during the late cycle? or do they reach the peak concomitantly with genome replication? This information can support the discussion on the genome-fibers assembly during the cycle.

   We thank this reviewer for these suggestions. We now added time of expression of the proteins involved in the genomic fiber composition along the manuscript. We added explicit sentences in the main text both for the GMC-oxidoreductases and RNA polymerase subunits. The RNA polymerase as well as proteins involved in mRNA maturation are in the virion (Table S2 B) and studies by others demonstrate early transcription takes place in the nucleoid once transferred in the host cytoplasm (Reference 24). We also provided a link to the reviewers where to find the expression data for the different mimivirus genes. http://www.igs.cnrs-mrs.fr/mimivirus/

   1. Taken together, data seem convincing to demonstrate that the virus genome is located inside the helical shield. However, I believe that the authors could better explain why we only see 20 kb fragments in the gel, including in the control (in Fig S2).

   We hope our answers to this comment will convince this reviewer.

   Fig S2 corresponds to a regular 1% agarose gel and not to a PFGE gel. This gel was simply to show there is DNA associated with the genomic fiber and not to show the size of the DNA as the genomic fiber has been broken into pieces and we thus do not expect to have very high molecular weight. I must point out that when extracting the DNA form Mimivirus capsids using standard kits and pipetting, it also migrates at the top of the gel (Lane 1 in Fig. S2) while it would likely appear as a smear above 20 kb on a PFGE. By contrast when the viral particles are put into plugs prior lysis, the genomic DNA migrates at the proper size, as shown in the publication from Boyer et al. 2011 (reference 31), showing the genome of Mimivirus is a linear genome migrating around 1.37 Mb (Fig 1, Panel B, Lane M1). In P9 of this letter, an image of a long (> 6 µm) and flexible fiber is presented.

   Reviewer #4 (Public Review):

   In the manuscript "The giant Mimivirus 1.2 Mb genome is elegantly organized into a 30 nm helical protein shield", the authors show that, when subjected to low pH stress, the Mimivirus particle releases 30nm-diameter filamentous assemblies. These filaments consist of a protein shell that envelopes the Mimivirus genomic DNA. The protein shell is composed of two GMC-oxidoreductases, the same protein that forms the long fibers emanating from the capsid of the Mimivirus.

   Overall, despite being interested in the subject, this scientist was left confused about several aspects of the paper described below. The presentation of the material is also confusing.

   We hope the answers and images we provide to all Reviewers in page 2 to 12 herein will clarify the various points raised by this reviewer.

   1. The presented data do not allow the estimation of the amount of mimivirus genome organized into 30 nm diameter filaments. Hence, the title of the paper is misleading.

   The entire genome should be packaged in the genomic fiber. That was already observed by other and we now provide an image of the nucleoid imaged by AFM that was published. The image was extracted from Kuznetsov et al. J. Virol. 2013. See p9 of this letter.

   2)The filamentous structures are a result of extremely harsh treatment of the virus particle, which starts with a 1.5 hour-long incubation at pH 2. Do the filaments actually exist inside the virus particle as the title of the paper implies?

   The 1 h incubation at 30°C and pH 2 was only applied to recover the nucleoids (see material and method section “Nucleoid extraction”) presented in Fig S1A. Acidic treatment was never applied to produce the genomic fiber as we noticed it is sensitive to both temperature and acidic treatment. All steps of the extraction protocol were performed at pH 7.5 (section: “Extraction and purification of the mimivirus genomic fiber”). We must emphasize that the release of the genomic fiber can be seen at the very first step of the extraction protocol (protease treatment). The sample was also controlled at each step of the protocol by negative staining TEM to assess the status of the genomic fiber. We had to optimize the protocol as using a too soft proteolytic treatment led to too few opened particles but with mostly a compact genomic fiber released, if it was too harsh, all particles were opened but the genomic fiber was mostly in the ribbon state. We had to compromise to get a decent amount of compact and relaxing structures to be able to perform the present work. We would like to stress out that we could reproducibly obtain the genomic fiber from many preparations and that we could observe them with different virions (including M4), even using different protocols (only the one with the better yield is reported in the manuscript).

   In the Figure 1B the genomic fiber can be seen inside a virion and is still encased in the membrane compartment. These structures were not reported in previous cryo-EM analyses of the virions. As said above, they were only reported by AFM studies of the mimivirus particles (Kuznetsov, Y. G. et al. Virology 2010; Kuznetsov YG et al. J. Virol. 2013, Fig 15). See p9.

   Or [might] these filaments [form during] host take over?

   Or [perhaps] these filaments [result from a harsh in vitro treatment] and have nothing to do with either?"

   The first two questions can be answered with the help of cryoFIB tomography, which might be beyond the scope of a "paper revision". However, the properties of the two GMCoxidoreductases in the presence and in the absence of genomic DNA must be examined in greater detail. Can these proteins, by themselves, form similar hollow filaments (or any filaments) when subjected to the same treatment as the virus?

   I personally have difficulties to imagine that such a complex structure could be the result of an artefact due to the treatment for several reasons:

   • It is unlikely that by simply putting the GMC-oxidoreductases with DNA would result in a helical structure where the DNA is folded 5 times and internally lining the protein shell (extended data video1 of one tomogram). It would be like crystallizing the proteins (in a heterogeneous sample) onto the folded DNA to form a helix with a hollow lumen. The crystallographic data obtained by others by on the mimivirus GMC-oxidoreductase did not produce tubular structures either and they reported 3 crystal forms. They overexpressed the proteins in E. coli and did not report such structures bound to DNA either.

   • Given the presence of compact and relaxed forms, once relaxed the helix cannot go back to a compact state passively by simply rewinding suggesting the relaxed forms are the result of decompaction of a constrained structure. This is also supported by the loss of DNA in the relaxed state Cl3. Last steps of unfolding correspond to the loss of one ribbon strand after the other.

   • The contacts between chains intra and inter strand are also scarce supporting an active assembly of the structure. We now provide an additional supplementary Table S4 with the different contacts for the different states of the genomic fiber.

   1. Although the assignment of the qu_946 oxidoreductase to the corresponding cryo-EM density is correct (as the resolution is high enough), I am confused about the other oxidoreductase (qu_143). Where does it fit to? Which structure does it form?

   We cannot discriminate between the two GMC-oxidoreductases due to their close identity (69% identity, 81% similarity) and the resolution of the map. Yet we think that in most cells the qu_946 GMC-oxidoreductase is the most abundant at the time of genome packaging (from our proteomic study, between 2 and 9 times). Yet, in some cells the second GMCoxidoreductase could become the most abundant and, in that case, the genomic fiber is built using qu_143.

   Equally important, what is going on with the N-terminal 50-residue domain of qu_946? Is there a space for it in the cryoEM map? Is it disordered?

   The N-terminal domain is only present in the fibrils decorating the capsids. As illustrated in Fig S12, when analyzed by MS-based proteomics, the comparison of the peptide coverage of the GMC-oxidoreductases whether they compose the fibrils or the genomic fiber is not the same. The N-terminal domain is clearly covered when the fibrils (data not shown) or intact virions are analyzed and not covered when the analysis is performed on the genomic fiber. That is why we propose this N-terminal domain could be an addressing signal (see main text) and that a protease could be cleaving it in the case of the genomic fiber assembly.

   Main text: The proteomic analyses provided different sequence coverages for the GMCoxidoreductases depending on whether samples were virions or the purified genomic fiber preparations, with substantial under-representation of the N-terminal domain in the genomic fiber (Fig. S12). Accordingly, the maturation of the GMC-oxidoreductases involved in genome packaging must be mediated by one of the many proteases encoded by the virus or the host cell.

   Indeed, there is no space to accommodate this domain as it would prevent the interaction between the protein shell and the DNA or/and induce an increase of the genomic fiber diameter that would be too big to be accommodated into the nucleoid.

   1. The bubblegram analysis is not very convincing. The bubbles appear to correlate with the length or thickness of the structure - the long or overlapped structures form bubbles. The bubbles may not be due to the presence of DNA.

   The point is, as demonstrated by our structural studies, that the relaxed structure lost the DNA. This is why bubble cannot be seen in the relaxed broken fibers. On long fibers still in compact form, the DNA is visible in the structure and bubble can be seen. Yet the evidence for the presence of DNA in the structure is also provided by the agarose gel of the purified genomic fiber and the cryo-EM structures. Bubblegrams are just one additional analysis which was provided.

   Read the original source
3. 

   eLife

   Jun 8, 2022

   Evaluation Summary:

   Giant dsDNA viruses, with genomes in excess of 1Mb that encode more than a thousand genes, were only recently discovered and their study offers new opportunities to understand life's evolved mechanisms. In this manuscript, Villalta and colleagues report results on one of the most complex known viruses, the Mimivirus. Its genome is compacted into magnificent fibers comprising apparently repurposed GMC-type oxidoreductase paralogs assembled as a helical coat around genomic dsDNA. Cryo-EM and cryo-ET image analysis yielded a structural model of the fiber in multiple states. The authors also provide some evidence that additional viral enzymes, including RNA polymerases, exist within the fiber assemblies. Pending the resolution of certain issues that emerged in peer review, the study will be of broad interest to biologists.

   Evaluation Summary:

   Giant dsDNA viruses, with genomes in excess of 1Mb that encode more than a thousand genes, were only recently discovered and their study offers new opportunities to understand life's evolved mechanisms. In this manuscript, Villalta and colleagues report results on one of the most complex known viruses, the Mimivirus. Its genome is compacted into magnificent fibers comprising apparently repurposed GMC-type oxidoreductase paralogs assembled as a helical coat around genomic dsDNA. Cryo-EM and cryo-ET image analysis yielded a structural model of the fiber in multiple states. The authors also provide some evidence that additional viral enzymes, including RNA polymerases, exist within the fiber assemblies. Pending the resolution of certain issues that emerged in peer review, the study will be of broad interest to biologists.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #3 agreed to share their name with the authors.)

   Read the original source
4. 

   eLife

   Jun 8, 2022

   Reviewer #1 (Public Review):

   Giant dsDNA viruses, with genomes in excess of 1Mb encoding more than one thousand genes, were only recently discovered and their study offers new opportunities to probe life's evolved mechanisms. Little is known how these "organisms" protect and organize their genomes. This fascinating study reveals a helical protein casing comprised of oxidoreductase-family proteins, which assemble 5- or 6-start helices genomic DNA lining the lumen of the helix. The remarkable evolutionary strategy for packaging the genome appears to be a convergent solution by comparison with distant thermotolerant viruses.

   Read the original source
5. 

   eLife

   Jun 8, 2022

   Reviewer #2 (Public Review):

   In this manuscript, Villalta, Schmitt, Estrozi and colleagues report their results on genome compaction in one of the most complex known viruses, the Mimivirus. This work will be of interest to a broad readership, and particularly to virologists and structural biologists. The authors describe a novel mechanism used by mimivirus to compact and package its 1.2 Mb dsDNA genome. In particular, the mimivirus genome is shown to be packed into magnificent cylinder-like assemblies composed of GMC-type oxidoreductases, presenting yet another remarkable case of enzyme exaptation. By using cryo-electron microscopy (cryo-EM) and cryo-electron tomography (cryo-ET), the authors determined the structures of such fibers in several relaxation states, which presumably represent different stages of nucleoprotein unpacking upon …

   Reviewer #2 (Public Review):

   In this manuscript, Villalta, Schmitt, Estrozi and colleagues report their results on genome compaction in one of the most complex known viruses, the Mimivirus. This work will be of interest to a broad readership, and particularly to virologists and structural biologists. The authors describe a novel mechanism used by mimivirus to compact and package its 1.2 Mb dsDNA genome. In particular, the mimivirus genome is shown to be packed into magnificent cylinder-like assemblies composed of GMC-type oxidoreductases, presenting yet another remarkable case of enzyme exaptation. By using cryo-electron microscopy (cryo-EM) and cryo-electron tomography (cryo-ET), the authors determined the structures of such fibers in several relaxation states, which presumably represent different stages of nucleoprotein unpacking upon delivery into host cytoplasm. The authors also suggest (although do not directly visualize) that the lumen of the genomic fibers contains several viral enzymes, most notably, DNA-dependent RNA polymerase, which is necessary for cytoplasmic replication of the mimivirus. Overall, this is an important discovery, which further expands our appreciation of the "inventiveness" of viruses.

   I am not an expert on helical reconstructions and cannot evaluate the validity of the models. Thus, my specific comments will focus on aspects of the work with which I am more familiar.

   1. In light of the presented results, it is reasonable to assume that GMC-type oxidoreductases of the mimivirus are very important for the formation of functional virions. However, in a previous study (PMID: 21646533), it has been shown that the genes encoding GMC-type oxidoreductases can be deleted from the virus genome (M4 mutant) without the loss of infectivity. The M4 virions were devoid of the external fibers decorating the icosahedral capsid, but the genome was still packaged. How do the authors reconcile these results with those presented in the present manuscript? This should be addressed in the Discussion section.

   2. The authors state that mimivirus encodes two GMC-type oxidoreductases (qu_946 and qu_143) and that both could be fitted into the electron densities. However, I could not understand whether the authors think that the fibers are heteroassemblies of both oxidoreductases or different fibers are composed of different proteins, or only one is used for fiber formation. Please clarify. In case you are not able to distinguish between the two homologs (e.g., due to limited resolution), state so explicitly.

   3. I am slightly puzzled by the observed "ball of yarn". It is hard for me to imagine that a cylindrical container/fiber containing a continuous dsDNA genome could be bent or fragmented into bundles because this would break the protein-protein interactions holding the fiber together. In Figures 1C and S1, are these parts of the same fiber or multiple fibers coming out of one capsid? Related to this question - is there evidence (e.g., from qPCR) that Mimivirus carries a single copy of genomic dsDNA per capsid?

   4. The authors describe the interactions between the monomers in the dimer of qu_946 as well as between qu_946 and DNA. I would also like to see a brief description of protein-protein interactions between subunits within the same helical strand as well as between helical strands, which hold the whole assembly together (i.e., what are the contacts between green subunits as well as between green and yellow subunits shown in Fig 2C). The authors suggest that the shell "would guide the folding of the dsDNA strands into the structure" (L310). To support this statement, the authors could show the lumen of the fiber rendered by electrostatic potential.

   5. Please provide some background information on the distribution of GMC-type oxidoreductases in other families of giant viruses, so that it is clearer whether the described packaging mechanism is specific to mimiviruses or is more widespread.

   Read the original source
6. 

   eLife

   Jun 8, 2022

   Reviewer #3 (Public Review):

   Since it was presented to the scientific community as a viral entity, mimivirus has the unlimited capacity to cause surprise and admiration. In this manuscript, Villalta, Schmitt, Estrozi, et al. and Abergel present how the mimivirus gigantic genome is organized into the virion. The authors succeeded in developing a protocol to trigger virus genome uncoating followed by genome-associated proteins purification. The presented data indicates that a helical shield composed of two GMC-type oxidoreductases is associated with the mimivirus genome, named genomic fiber. By cryo-EM, and cryo-tomography different forms and stages of the genomic fiber were detailed described, indicating the dynamics of fibers conformational changes, likely related to genome packing and uncoating during the virus replication cycle. …

   Reviewer #3 (Public Review):

   Since it was presented to the scientific community as a viral entity, mimivirus has the unlimited capacity to cause surprise and admiration. In this manuscript, Villalta, Schmitt, Estrozi, et al. and Abergel present how the mimivirus gigantic genome is organized into the virion. The authors succeeded in developing a protocol to trigger virus genome uncoating followed by genome-associated proteins purification. The presented data indicates that a helical shield composed of two GMC-type oxidoreductases is associated with the mimivirus genome, named genomic fiber. By cryo-EM, and cryo-tomography different forms and stages of the genomic fiber were detailed described, indicating the dynamics of fibers conformational changes, likely related to genome packing and uncoating during the virus replication cycle. In-depth analysis of a substantial number of individual virus fibers revealed that the mimivirus genome is folded and organized inside the aforementioned helical shield, which seems to be novel among giant icosahedral viruses. Proteomics in association with image analysis indicates that mimivirus packed genome forms a channel, which accommodates key enzymes related to early phases of the replication cycle, especially RNA polymerase subunits.

   I must disclose that I am not an expert on structural virology and proteomic analysis. Therefore, I don't feel I can contribute to the improvement of this kind of analysis. That said, I congratulate the authors for their efforts to make the manuscript story understandable to non-experts.

   I have a few suggestions and comments:

   1. Please consider the "nucleocapsid" concept during genomic fiber presentation. I believe it fits in;

   2. The "ball of yarn" analogy is nice, but fig 1C shows several fibers unconnected (free) in one of their ends. I am wondering if it means that the genomic fiber is not a long-single structure covering the whole genome, but a bunch of several independent helical structures covering the whole genome and attached in such "ball of yarn". Like several threads connected. Could the authors clarify that please?

   3. Considering previously published data on proteomics of viral factories and transcriptomics of mimivirus: is there any temporal association between GMC-type oxidoreductases' peak of expression and genome replication during the viral cycle? what about RNA pol subunits? Are all those proteins highly expressed during the late cycle? or do they reach the peak concomitantly with genome replication? This information can support the discussion on the genome-fibers assembly during the cycle.

   4- Taken together, data seem convincing to demonstrate that the virus genome is located inside the helical shield. However, I believe that the authors could better explain why we only see 20 kb fragments in the gel, including in the control (in Fig S2).

   Read the original source
7. 

   eLife

   Jun 8, 2022

   Reviewer #4 (Public Review):

   In the manuscript "The giant Mimivirus 1.2 Mb genome is elegantly organized into a 30 nm helical protein shield", the authors show that, when subjected to low pH stress, the Mimivirus particle releases 30nm-diameter filamentous assemblies. These filaments consist of a protein shell that envelopes the Mimivirus genomic DNA. The protein shell is composed of two GMC-oxidoreductases, the same protein that forms the long fibers emanating from the capsid of the Mimivirus.

   Overall, despite being interested in the subject, this scientist was left confused about several aspects of the paper described below. The presentation of the material is also confusing.

   1. The presented data do not allow the estimation of the amount of mimivirus genome organized into 30 nm diameter filaments. Hence, the title of the paper is …

   Reviewer #4 (Public Review):

   In the manuscript "The giant Mimivirus 1.2 Mb genome is elegantly organized into a 30 nm helical protein shield", the authors show that, when subjected to low pH stress, the Mimivirus particle releases 30nm-diameter filamentous assemblies. These filaments consist of a protein shell that envelopes the Mimivirus genomic DNA. The protein shell is composed of two GMC-oxidoreductases, the same protein that forms the long fibers emanating from the capsid of the Mimivirus.

   Overall, despite being interested in the subject, this scientist was left confused about several aspects of the paper described below. The presentation of the material is also confusing.

   1. The presented data do not allow the estimation of the amount of mimivirus genome organized into 30 nm diameter filaments. Hence, the title of the paper is misleading.

   2. The filamentous structures are a result of extremely harsh treatment of the virus particle, which starts with a 1.5 hour-long incubation at pH 2. Do the filaments actually exist inside the virus particle as the title of the paper implies? Or these filaments are formed in the process of host take over? Or these filaments are a result of the harsh treatment of the particle and have nothing to do with either? The first two questions can be answered with the help of cryoFIB tomography, which might be beyond the scope of a "paper revision". However, the properties of the two GMC-oxidoreductases in the presence and in the absence of genomic DNA must be examined in greater detail. Can these proteins, by themselves, form similar hollow filaments (or any filaments) when subjected to the same treatment as the virus?

   3. Although the assignment of the qu_946 oxidoreductase to the corresponding cryo-EM density is correct (as the resolution is high enough), I am confused about the other oxidoreductase (qu_143). Where does it fit to? Which structure does it form? Equally important, what is going on with the N-terminal 50-residue domain of qu_946? Is there a space for it in the cryoEM map? Is it disordered?

   4. The bubblegram analysis is not very convincing. The bubbles appear to correlate with the length or thickness of the structure - the long or overlapped structures form bubbles. The bubbles may not be due to the presence of DNA.

   Read the original source
8. 

   Version published to 10.1101/2022.02.17.480895v1 on bioRxiv

   Feb 18, 2022