1. Microbiology and Infectious Disease
  2. Structural Biology and Molecular Biophysics

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

  1. Alejandro Villalta
  2. Alain Schmitt
  3. Leandro F Estrozi
  4. Emmanuelle RJ Quemin
  5. Jean-Marie Alempic
  6. Audrey Lartigue
  7. Vojtěch Pražák
  8. Lucid Belmudes
  9. Daven Vasishtan
  10. Agathe MG Colmant
  11. Flora A Honoré
  12. Yohann Couté
  13. Kay Grünewald
  14. Chantal Abergel  Is a corresponding author
  1. Aix–Marseille University, Centre National de la Recherche Scientifique, Information Génomique & Structurale, Unité Mixte de Recherche 7256 (Institut de Microbiologie de la Méditerranée, FR3479, IM2B), France
  2. Univ. Grenoble Alpes, CNRS, CEA, Institut de Biologie Structurale (IBS), France
  3. Centre for Structural Systems Biology, Leibniz Institute for Experimental Virology (HPI), University of Hamburg, Germany
  4. Division of Structural Biology, Wellcome Centre for Human Genetics, University of Oxford, United Kingdom
  5. Univ. Grenoble Alpes, CEA, INSERM, IRIG, BGE, France
https://doi.org/10.7554/eLife.77607
Share this article
Cite this article
  1. Alejandro Villalta
  2. Alain Schmitt
  3. Leandro F Estrozi
  4. Emmanuelle RJ Quemin
  5. Jean-Marie Alempic
  6. Audrey Lartigue
  7. Vojtěch Pražák
  8. Lucid Belmudes
  9. Daven Vasishtan
  10. Agathe MG Colmant
  11. Flora A Honoré
  12. Yohann Couté
  13. Kay Grünewald
  14. Chantal Abergel
(2022)
The giant mimivirus 1.2 Mb genome is elegantly organized into a 30-nm diameter helical protein shield
eLife 11:e77607.
https://doi.org/10.7554/eLife.77607
  2. Download BibTeX
  3. Download .RIS

Abstract

Mimivirus is the prototype of the Mimiviridae family of giant dsDNA viruses. Little is known about the organization of the 1.2 Mb genome inside the membrane-limited nucleoid filling the ~0.5 µm icosahedral capsids. Cryo-electron microscopy, cryo-electron tomography, and proteomics revealed that it is encased into a ~30-nm diameter helical protein shell surprisingly composed of two GMC-type oxidoreductases, which also form the glycosylated fibrils decorating the capsid. The genome is arranged in 5- or 6-start left-handed super-helices, with each DNA-strand lining the central channel. This luminal channel of the nucleoprotein fiber is wide enough to accommodate oxidative stress proteins and RNA polymerase subunits identified by proteomics. Such elegant supramolecular organization would represent a remarkable evolutionary strategy for packaging and protecting the genome, in a state ready for immediate transcription upon unwinding in the host cytoplasm. The parsimonious use of the same protein in two unrelated substructures of the virion is unexpected for a giant virus with thousand genes at its disposal.

Editor's evaluation

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 mechanisms. Little is known how these "organisms" protect and organize their genomes. This fascinating study reveals a helical protein assembly comprised of oxidoreductase-family proteins, which assemble into multi-start helical fibers, with genomic DNA lining the lumen of the fiber.

https://doi.org/10.7554/eLife.77607.sa0

Introduction

Acanthamoeba infecting giant viruses were discovered with the isolation of mimivirus (La Scola et al., 2003; Raoult et al., 2004). Giant viruses now represent a highly diverse group of dsDNA viruses infecting unicellular eukaryotes (Abergel et al., 2015) which play important roles in the environment (Schulz et al., 2020; Moniruzzaman et al., 2020; Kaneko et al., 2021). They also challenge the canonical definitions of viruses (Forterre, 2010; Claverie and Abergel, 2010) as they can encode central translation components (Raoult et al., 2004; Abergel et al., 2007) as well as a complete glycosylation machinery (Piacente et al., 2017; Notaro et al., 2021) among other unique features.

Mimivirus has been the most extensively studied giant virus infecting Acanthamoeba (Colson et al., 2017). The virions are 0.75 µm wide and consist of icosahedral capsids of 0.45 µm diameter surrounded by a dense layer of radially arranged fibrils (Raoult et al., 2004). Structural analyses of the virions have provided some insights into the capsid structure (Kuznetsov et al., 2013; Xiao et al., 2009; Klose et al., 2010; Ekeberg et al., 2015; Schrad et al., 2020) but given the size of the icosahedral particles (and hence the sample thickness), accessing the internal organization of the core of the virions remains challenging. Consequently, little is known about the packaging of the 1.2 Mb dsDNA genome (Chelikani et al., 2014). Inside the capsids, a lipid membrane delineates an internal compartment (~340 nm in diameter, Figure 1A, Figure 1—figure supplement 1), referred to as the nucleoid, which contains the viral genome, together with all proteins necessary to initiate the replicative cycle within the host cytoplasm (Schrad et al., 2020; Kuznetsov et al., 2010; Claverie et al., 2009; Arslan et al., 2011). Acanthamoeba cells engulf mimivirus particles, fooled by their bacteria-like size and the heavily glycosylated decorating fibrils (La Scola et al., 2003; Raoult et al., 2004; Notaro et al., 2021). Once in the phagosome, the Stargate portal located at one specific vertex of the icosahedron opens up (Zauberman et al., 2008), enabling the viral membrane to fuse with that of the host vacuole to deliver the nucleoid into the host cytoplasm (Schrad et al., 2020; Claverie et al., 2009). EM studies have shown that next the nucleoid gradually loses its electron dense appearance, transcription begins and the early viral factory is formed (Arslan et al., 2011; Suzan-Monti et al., 2007; Mutsafi et al., 2014). Previous atomic force microscopy studies of the mimivirus infectious cycle suggested that the DNA forms a highly condensed nucleoprotein complex enclosed within the nucleoid (Kuznetsov et al., 2013). Here, we show that opening of the large icosahedral capsid in vitro led to the release of rod-shaped structures of about 30-nm width. These structures were further purified and the various conformations characterized using cryo-electron microscopy (cryo-EM), tomography, and mass spectrometry (MS)-based proteomics.

Figure 1 with 10 supplements see all
Download asset Open asset
The mimivirus genomic fiber.

(A) Micrograph of an ultrathin section of resin-embedded infected cells showing the DNA tightly packed inside mimivirus capsids (C) with electron dense material inside the nucleoid (N). The string-like features, most likely enhanced by the dehydration caused by the fixation and embedding protocol, correspond to the genomic fiber (F) packed into the nucleoid. (B) Micrograph of negative stained mimivirus capsid (C) opened in vitro with the genomic fiber (F) still being encased into the membrane-limited nucleoid (N). (C) Multiple strands of the flexible genomic fiber (F) are released from the capsid (C) upon proteolytic treatment. (D) Micrograph of negative stained purified mimivirus genomic fibers showing two conformations the right fiber resembling the one in (B) and free DNA strands (white arrowheads). (E) Slices through two electron cryo tomograms of the isolated helical protein shell of the purified genomic fibers in compact (top) or relaxed conformation (bottom) in the process of losing one protein strand (Figure 1—figure supplement 2, Figure 1—videos 1, 4). (F) Different slices through the two tomograms shown in (E) reveal DNA strands lining the helical protein shell of the purified genomic fibers in compact (top) or relaxed conformation (bottom). Examples of DNA strands extending out at the breaking points of the genomic fiber are marked by black arrowheads. Note in the top panel, individual DNA strands coated by proteins (red arc). The slicing planes at which the mimivirus genomic fibers were viewed are indicated on diagrams on the top right corner as blue dashed lines and the internal colored segments correspond to DNA strands lining the protein shell. The thickness of the tomographic slices is 1.1 nm and the distance between tomographic slices in panels (E) and (F) is 4.4 nm. Scale bars as indicated.

Results

Capsid opening induces the release of a ~30-nm-wide rod-shaped structure that contains the dsDNA genome

We developed an in vitro protocol for particle opening that led to the release of ~30 nm-wide rod-shaped fibers of several microns in length (Figure 1, Figure 1—figure supplement 1). We coined this structure the mimivirus genomic fiber. Complete expulsion of the opened nucleoid content produced bundled fibers resembling a ‘ball of yarn’ (Kuznetsov et al., 2013; Figure 1C). The capsid opening procedure involves limited proteolysis and avoids harsh conditions, as we found that the structure becomes completely denatured by heat (95°C) and is also sensitive to acidic treatment, thus preventing its detection in such conditions (Schrad et al., 2020). Various conformations of the genomic fiber were observed, sometimes even on the same fiber (Figure 1D), ranging from the most compact rod-shaped structures (Figure 1D [left], and Figure 1E, F [top]) to more relaxed structures where DNA strands begin to dissociate (Figure 1D [right], and Figure 1E, F [bottom]). After optimizing the in vitro extraction on different strains of group-A mimiviruses, we focused on an isolate from La Réunion Island (mimivirus reunion), as more capsids were opened by our protocol, leading to higher yields of genomic fibers that were subsequently purified on sucrose gradient. All opened capsids released genomic fibers (Figure 1C, Figure 1—figure supplement 1).

The first confirmation of the presence of DNA in the genomic fiber was obtained by agarose gel electrophoresis (Figure 1—figure supplement 3). Cryo-EM bubblegram analysis (Wu et al., 2012; Cheng et al., 2014) gave a further indication that the nucleic acid is located in the fiber lumen. Alike other nucleoprotein complexes, fibers are expected to be more susceptible to radiation damage then pure proteinaceous structures. Surprisingly, the specimen could sustain higher electron irradiation before the appearance of bubbles compared to other studies (Mishyna et al., 2017): 600 e−/Ų for relaxed helices and up to 900 e−/Ų for long compact ones, while no bubbles could be detected in unfolded ribbons (Figure 1—figure supplement 4). For comparison, bacteriophage capsids containing free DNA, that is not in the form of nucleoproteins, show bubbling for doses of ~30–40 e−/Ų (Mishyna et al., 2017).

Cryo-EM single-particle analysis of the different compaction states of the mimivirus genomic fiber

In order to shed light on the mimivirus genome packaging strategy and to determine the structure of the purified genomic fibers, we performed cryo-EM single-particle analysis. The different conformations of the genomic fiber initially observed by negative staining (Figure 1) and cryo-EM resulted in a highly heterogeneous dataset for single-particle analysis. In order to separate different conformations, in silico sorting through 2D classification (using Relion, He and Scheres, 2017; Scheres, 2012) was performed. Next, we carried out cluster analysis relying on the widths of the helical segments and correlations (real and reciprocal spaces) between the experimental 2D-class patterns (Figure 1—figure supplements 5 and 6 ). Three independent clusters (Cl) could be distinguished, corresponding to the compact (Cl1), intermediate (Cl2), and relaxed (Cl3) fiber conformations (Figure 1—figure supplement 6) with the latter being the widest. For each cluster, we determined their helical symmetry parameters by image power spectra analyses and performed structure determination and refinement (Figure 2, Figure 2—figure supplements 1–3).

Figure 2 with 5 supplements see all
Download asset Open asset
Structures of the mimivirus genomic fiber for Cl1a (A–C), Cl3a (D–F), and Cl2 (G–I): Electron microscopy (EM) maps of Cl1a (A) and Cl3a (D) are shown with each monomer of one GMC-oxidoreductase dimer colored in green and orange and three adjacent dimers in yellow, to indicate the large conformational change taking place between the two fiber states.

The transition from Cl1a to Cl3a (5-start helix) corresponds to a rotation of each individual unit (corresponding to a GMC-oxidoreductase dimer) by ~−10° relative to the fiber longitudinal axis and a change in the steepness of the helical rise by ~−11°. Scale bars, 50 Å. Compared to Cl1a, the Cl2 6-start helix (G) shows a difference of ~2° relative to the fiber longitudinal axis and ~2° in the steepness of the helical rise. Scale bars, 50 Å. Cross-sectional (bottom) and longitudinal (top) sections through the middle of final Cl1a (B), Cl3a (E), and Cl2 (H) EM maps. Scale bars, 50 Å. Longitudinal (top) and orthogonal (bottom) views of final Cl1a (C), Cl3a (F), and Cl2 (I) EM maps color coded according to each start of the 5-start helix. Densities for some asymmetric units in the front have been removed on the side view map to show the five DNA strands lining the protein shell interior. Scale bars 50 Å.

Figure 3 with 4 supplements see all
Download asset Open asset
Maps of the compact (Cl1a and Cl2) genomic fiber structures.

(A) Cl1a EM map corresponding to the protein shell prior focused refinement is shown as a transparent surface and the five DNA strands as solid surface. One protein dimer strand is shown yellow except for one asymmetric unit (transparent yellow) to illustrate the dimer fit. The position of a second dimer (green) from the adjacent dimer strand is shown to emphasize that the DNA strand (gold) is lining the interface between two dimers. (B) Cartoon representation of GMC-oxidoreductase qu_946 dimers fitted into one of the Cl1a 5-start helix strands in the 3.7-Å resolution map. The map is shown at a threshold highlighting the periodicity of contacts between the dsDNA and the protein shell. Charged distribution on surface representation of Cl1a protein shell made of qu_946 dimers (C) or qu_143 dimers (D). Cartoon representation of qu_946 (E) and qu_143 (F) fitted into the Cl1a cryo-electron microscopy (cryo-EM) maps highlighting the interacting residues (given as stick models) between each monomer and one dsDNA strand. The isosurface threshold chosen allows visualization of density for the manually built N-terminal residues, including terminal cysteines (stick model), of two neighboring monomers that could form a terminal disulfide bridge. (G) Zoom into the 3.3 Å resolution focused refined Cl1a map illustrating the fit of the side chains and the FAD ligand (Figure 3—video 1). (H) Cartoon representation of the DNA fitted in the focused refined DNA only map (Figure 3—figure supplement 2). (I) Focused refined Cl1a map colored by monomer, next to a cartoon representation of the qu_946 dimer (α-helices in red, β-strands in blue, and coils in yellow). Secondary structure elements are annotated in both representations (H: helix, B: beta-strand).

For both Cl1a and Cl3a conformations, after 3D refinement, we obtained helical structures of 3.7 Å resolution (Fourrier shell correlation [FSC] threshold 0.5, masked), corresponding to 5-start left-handed helices made of a ~8-nm-thick proteinaceous external shell (Figure 2—figure supplements 1–2). For the most compact conformation (Cl1a) five dsDNA strands were lining the interior of the protein shell leaving a ~9-nm-wide central channel (Figures 2 and 3, Supplementary file 1). The dsDNA strands appear as curved cylinders in the helical structure, the characteristic shape of the DNA (minor and major groove) becoming only visible after focused refinement of a single strand of dsDNA (Figure 3, Figure 3—figure supplements 1 and 2). In the relaxed subcluster Cl3a, the DNA strands at the interface to the ~17-nm-wide central channel are not clearly recognizable (Figure 2, Supplementary file 1, and Figure 2—figure supplement 2), most likely because they are at least partially detached inside the broken expanded fiber. The breaks after relaxation of the helix might be the result of the extraction and purification treatment, while DNA will remain in the central channel, at least in the early phase of Acanthamoeba infection.

Finally, the 4-Å resolution Cl2 map obtained after 3D refinement (Figure 2—figure supplement 3) corresponds to a 6-start left-handed helix made of a~8-nm-thick proteinaceous external shell, with six dsDNA strands lining the shell interior and leaving a ~12-nm-wide inner channel (Figure 2, Supplementary file 1).

The most abundant proteins in the genomic fiber are GMC oxidoreductases, the same that compose the fibrils decorating mimivirus capsid

MS-based proteomic analyses performed on three biological replicates identified two GMC oxidoreductases as the main components of the purified genomic fiber (qu_946 and qu_143 in mimivirus reunion corresponding to L894/93 and R135, respectively, in mimivirus prototype) (Supplementary file 2). The two mimivirus reunion proteins share 69% identity (81% similarity). The available mimivirus R135 GMC-oxidoreductase dimeric structure (Klose et al., 2015) (PDB 4Z24, lacking the 50 amino acid long cysteine-rich N-terminal domain) was fitted into the EM maps (Figures 2 and 3). This is quite unexpected, since GMC oxidoreductases are already known to compose the fibrils surrounding mimivirus capsids (Notaro et al., 2021; Boyer et al., 2011). The corresponding genes are highly expressed during the late phase of the infection cycle at the time of virion assembly. Notably, the proteomic analyses provided different sequence coverages for the GMC oxidoreductases depending on whether samples were intact virions or purified genomic fiber preparations, with substantial under-representation of the N-terminal domain in the genomic fiber (Figure 2—figure supplement 4). 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. Interestingly, mimivirus M4 (Boyer et al., 2011), a laboratory strain having lost the genes responsible for the synthesis of the two polysaccharides decorating mimivirus fibrils (Notaro et al., 2021) also lacks the GMC-oxidoreductase genes. Additional studies on this specific variant will be key to establish if it exhibits a similar genomic fiber, and if yes, which proteins are composing it.

Analysis of the genomic fiber structure

The EM maps and FSC curves of Cl1a are shown in Figure 2—figure supplement 1. An additional step of refinement focused on the asymmetric unit further improved the local resolution to 3.3 Å as indicated by the corresponding FSC (Figure 3 and Figure 2—figure supplement 1). After fitting the most abundant GMC-oxidoreductase qu_946 (SWISS-MODEL model Waterhouse et al., 2018) in the final map of Cl1a, five additional N-terminal residues in each monomer were manually built using the uninterpreted density available. This strikingly brings the cysteines of each monomer (C51 in qu_946) close enough to allow a disulfide bridge, directly after the 50 amino acids domain not covered in the proteomic analysis of the genomic fiber (Figure 3G, Figure 2—figure supplement 4). The N-terminal chain, being more disordered than the rest of the structure, it is absent in the focused refined map, and also absent in the Cl3a map of the relaxed helix, suggesting that a break of the disulfide bridge could be involved in the observed unwinding process. Models of the three helical assemblies and asymmetric unit were further refined using the real-space refinement program in PHENIX 1.18.2 (Liebschner et al., 2019). In the 3.3-Å resolution map of the asymmetric unit, most side chains and notably the FAD cofactor are accommodated by density suggesting that the oxidoreductase enzyme could be active (Figure 3D). Density that can be attributed to the FAD cofactor is also present in the Cl2 and Cl3a maps. The atomic models of Cl1a and Cl3a dimers are superimposable with a core root-mean-square deviation (RMSD) of 0.68 Å based on Cα atoms.

Inspection of individual genomic fibers in the tomograms confirmed the coexistence of both 5- and 6-start left-handed helices containing DNA (Figure 1—figure supplement 2 and Figure 3—animation 1, Figure 1—video 1). Further, some intermediate and relaxed structures were also observed in which the DNA segments appeared detached from the protein shell and sometimes completely absent from the central channel of the broken fibers. Both GMC oxidoreductases (qu_946 and qu_143) can be fitted in the 5- and 6-start maps.

In relaxed or broken fibers, large electron dense structures that might correspond to proteins inside the lumen were sometimes visible (Figure 1—figure supplement 2C, E and Figure 1—video 3) as well as dissociating DNA fragments, either in the central channel (Figure 1—figure supplement 2B and Figure 1—video 2) or at the breakage points of the fibers in its periphery (Figure 1—figure supplement 2D and Figure 1—video 4). Densities corresponding to the dimer subunits composing the protein shell were also commonly observed on dissociated DNA strands (Figure 1E, F, Figure 1—figure supplement 2, and Figure 1—video 1; Figure 1—video 2; Figure 1—video 3; Figure 1—video 4).

In the Cl1a and Cl2 helices, the monomers in each dimer are interacting with two different dsDNA strands. As a result, the DNA strands are interspersed between two dimers, each also corresponding to a different strand of the protein shell helix (Figure 3). Based on the periodic contacts between protein shell and DNA strands, these interactions might involve, in the case of the Cl1a helix, one aspartate (D82 relative to the N-terminal methionine in qu_946), one glutamate (E321), two lysines (K344, K685), one arginine (R324), and a histidine (H343) or one asparagine (N80), two lysines (K319, K342), one arginine (R322), and one tyrosine (Y687), in the case of the qu_143 (Figure 3E, F, Supplementary file 3). Intra- and interstrands contacts between each dimer are presented in Supplementary file 3 for qu_143 and qu_946 in Cl1a, Cl2, and Cl3 maps.

Despite the conformational heterogeneity and the flexibility of the rod-shaped structure, we were able to build three atomic models of the mimivirus genomic fiber, in compact (5- and 6-start) and relaxed (5-start) states. Higher resolution data would still be needed to determine the precise structure of the dsDNA corresponding to the viral genome (Figure 3B, H), however, the lower resolution for this part of the map even in focused refinement runs (Figure 3H) might also mean that the DNA does not always bind in the same orientation.

Rough estimation of genome compaction to fit into the nucleoid

Since there is a mixture of five and six strands of DNA in the genomic fiber, this could correspond to five or six genomes per fiber or to a single folded genome. Assuming that the length of DNA in B-form is ~34 Å for 10 bp, the mimivirus linear genome of 1.2 × 106 bp would extend over ~400 µm and occupy a volume of 1.3 × 106 nm3 (~300 µm and ~1 × 106 nm3 if in A-form) (Li et al., 2019). The volume of the nucleoid (Kuznetsov et al., 2010) (~340 nm in diameter) is approximately 2.1 × 107 nm3 and could accommodate over 12 copies of viral genomes in a naked state, but only 40 µm of the ~30-nm-wide flexible genomic fiber. Obviously, the mimivirus genome cannot be simply arranged linearly in the genomic fiber and must undergo further compaction to accommodate the 1.2 Mb genome in a ~40-µm-long genomic fiber. As a result, the complete mimivirus genome, folded at least five times, fits into the helical shell. This structure surprisingly resembles a nucleocapsid, such as the archaea infecting APBV1 nucleocapsid (Ptchelkine et al., 2017).

Additional proteins, including RNA polymerase subunits, are enriched in the genomic fiber

The proteomic analysis of fiber preparations revealed the presence of additional proteins including several RNA polymerase subunits: Rpb1 and Rpb2 (qu_530/532 and qu_261/259/257/255), Rpb3/11 (qu_493), Rpb5 (qu_245), RpbN (qu_379), and Rpb9 (qu_219), in addition to a kinesin (qu_313), a regulator of chromosome condensation (qu_366), a helicase (qu_572), to be possibly associated with the genome (Supplementary file 2). In addition to the two GMC oxidoreductases, at least three oxidative stress proteins were also identified together with hypothetical proteins (Supplementary file 2). RNA polymerase subunits start being expressed 1hr postinfection with a peak after 5 hr and are expressed until the end the infection cycle. GMC oxidoreductases, kinesin, regulator of chromosome condensation are all expressed after 5 hr of infection until the end of the cycle.

As expected, the core protein (qu_431) composing the nucleoid and the major capsid proteins (MCP, qu_446) were significantly decreased in the genomic fiber proteome compared to intact virions. In fact, qu_431 and qu_446 represent, respectively, 4.5% and 9.4% of the total protein abundance in virions whereas they only account for 0.4% and 0.7% of the total protein abundance in the genomic fiber, suggesting that they could be contaminants in this preparation. On the contrary, we calculated enrichment factors of more than 500 (qu_946) and 26 (qu_143) in the genomic fiber samples compared to the intact virion. Finally, the most abundant RNA polymerase subunit (qu_245) is increased by a factor of eight in the genomic fiber compared to intact virion (if the six different subunits identified are used, the global enrichment is sevenfold). Furthermore, upon inspection of the negative staining micrographs, macromolecules strikingly resembling the characteristic structure of the poxviruses RNA polymerase (Grimm et al., 2019) were frequently observed scattered around the unwinding fiber and sometimes sitting on DNA strands near broken fibers (Figure 4). Together with the tomograms showing large electron dense structures in the lumen, some RNA polymerase units could occupy the center of the genomic fiber.

Figure 4
Download asset Open asset
RNA polymerase could be associated to the genomic fiber.

(A) Micrograph of negative stained fiber with released DNA still being connected to a relaxed and broken fiber and adjacent scattered macromolecular complexes that might resemble RNA polymerases. (B) Strikingly, some of them (black arrows) appear to sit on a DNA strand (white arrow). (C) E, particle extracted from the NS-TEM image; P, projections of vaccinia virus RNA polymerase (6RIC and 6RUI, Grimm et al., 2019) structure in preferred orientation; CE, clean extraction (see Material and methods). White and black boxed corresponds to images with white and black asterisk (CE), respectively. Scale bar, 50 Å. Negative staining imaging may dehydrate the objects and change macromolecules volumes.

Discussion

Several DNA compaction solutions have been described. For instance, the DNA of filamentous viruses infecting archaea is wrapped by proteins to form a ribbon which in turn folds into a helical rod forming a cavity in its lumen (DiMaio et al., 2015; Wang et al., 2020). In contrast, the chromatin of cellular eukaryotes consists of DNA wrapped around histone complexes (Robinson et al., 2006). It was recently shown that the virally encoded histone doublets of the Marseilleviridae can form nucleosomes (Liu et al., 2021; Valencia-Sánchez et al., 2021) and such organization would be consistent with previous evolutionary hypotheses linking giant DNA viruses with the emergence of the eukaryotic nucleus (Bell, 2001; Bell, 2020; Chaikeeratisak et al., 2017; Claverie, 2006; Takemura, 2001). Herpesviruses (Gong et al., 2019; Liu et al., 2019), bacteriophages (Sun et al., 2015; Rao and Feiss, 2015), and APBV1 archaeal virus (Ptchelkine et al., 2017) package their dsDNA genome as naked helices or coils. Yet, APBV1 nucleocapsid structure strikingly resemble the mimivirus genomic fiber with a proteinacious shell enclosing the folded dsDNA genome. Consequently, mimivirus genomic fiber is a nucleocapsid further bundled as a ball of yarn into the nucleoid, itself encased in the large icosahedral capsids. The structure of the mimivirus genomic fiber described herein supports a complex assembly process where the DNA must be folded into five or six strands prior to or concomitant with packaging, a step that may involve the repeat containing regulator of chromosome condensation (qu_366) identified in the proteomic analysis of the genomic fiber. The proteinaceous shell, via contacting residues between the dsDNA and the GMC oxidoreductases, would guide the folding of the dsDNA strands into the structure prior loading into the nucleoid. The lumen of the fiber being large enough to accommodate the mimivirus RNA polymerase, we hypothesize that it could be sitting on the highly conserved promoter sequence of early genes (Suhre et al., 2005). This central position would support the involvement of the RNA polymerase in genome packaging into the nucleoid and could determine the channel width via its anchoring on the genome (Figure 4D). According to this scenario, the available space (although tight) for the RNA polymerase inside the genomic fiber lumen suggests it could be sterically locked inside the compact form of the genomic fiber and could start moving and transcribing upon helix relaxation, initiating the replicative cycle and the establishment of the cytoplasmic viral factory. The genome and the transcription machinery would thus be compacted together into a proteinaceous shield, ready for transcription upon relaxation (Figure 2—video 1). This organization would represent a remarkable evolutionary strategy for packaging and protecting the viral genome, in a state ready for immediate transcription upon unwinding in the host cytoplasm. This is conceptually reminiscent of icosahedral and filamentous dsRNA viruses which pack and protect their genomes together with the replicative RNA polymerase into an inner core (Toriyama, 1986; Collier et al., 2016; Ding et al., 2019). As a result, replication and transcription take place within the protein shield and viral genomes remain protected during their entire infectious cycle. In the case of dsDNA viruses however, the double helix must additionally open up to allow transcription to proceed, possibly involving the helicase identified in our proteomic study (Supplementary file 2). Finally, in addition to their structural roles, the FAD containing GMC oxidoreductases making the proteinaceous shield, together with other oxidative stress proteins (Supplementary file 2), could alleviate the oxidative stress to which the virions are exposed while entering the cell by phagocytosis.

Mimivirus virion thus appears as a Russian doll, with its icosahedral capsids covered with heavily glycosylated fibrils, two internal membranes, one lining the capsid shell, the other encasing the nucleoid, in which the genomic fiber is finally folded. To our knowledge, the structure of the genomic fiber used by mimivirus to package and protect its genome in the nucleoid represents the first description of the genome organization of a giant virus. Since the genomic fiber appears to be expelled from the nucleoid as a flexible and subsequently straight structure starting decompaction upon release, we suspect that an active, energy-dependent, process is required to bundle it into the nucleoid during virion assembly. Such an efficient structure is most likely shared by other members of the Mimiviridae family infecting Acanthamoeba and could be used by other dsDNA viruses relying on exclusively cytoplasmic replication like poxviruses to immediately express early genes upon entry into the infected cell (Malkin et al., 2003; Blanc-Mathieu et al., 2021). Finally, the parsimonious use of moonlighting GMC oxidoreductases playing a central role in two functionally unrelated substructures of the mimivirus particle: (1) as a component of the heavily glycosylated peripheral fibril layer and (2) as a proteinaceous shield to package the dsDNA into the genomic fibers questions the evolutionary incentive leading to such an organization for a virus encoding close to a thousand proteins.

Materials and methods

Nucleoid extraction

Request a detailed protocol

Mimivirus reunion virions defibrillated, as described previously (Notaro et al., 2021; Kuznetsov et al., 2010), were centrifuged at 10,000 × g for 10 min, resuspended in 40 mM TES pH 2 and incubated for 1 hr at 30°C to extract the nucleoid from the opened capsids.

Extraction and purification of the mimivirus genomic fiber

Request a detailed protocol

The genomic fiber was extracted from 12 ml of purified mimivirus reunion virions at 1.5 × 1010 particles/ml, split into 12 × 1 ml samples processed in parallel. Trypsin (Sigma T8003) in 40 mM Tris–HCl pH 7.5 buffer was added at a final concentration of 50 µg/ml and the virus-enzyme mix was incubated for 2 hr at 30°C in a heating dry block (Grant Bio PCH-1). Dithiothreitol (DTT) was then added at a final concentration of 10 mM and incubated at 30°C for 16 hr. Finally, 0.001% Triton X-100 was added to the mix and incubated for 4 hr at 30°C. Each tube was vortexed for 20 s with 1.5-mm diameter stainless steel beads (CIMAP) to separate the fibers from the viral particles and centrifuged at 5000 × g for 15 min to pellet the opened capsids. The supernatant was recovered, and the fibers were concentrated by centrifugation at 15,000 × g for 4 hr at 4°C. Most of the supernatant was discarded leaving 12 × ~200 µl of concentrated fibers that were pooled and layered on top of ultracentrifuge tubes of 4 ml (polypropylene centrifuge tubes, Beckman Coulter) containing a discontinuous sucrose gradient (40%, 50%, 60%, 70% [wt/vol] in 40 mM Tris–HCl pH 7.5 buffer). The gradients were centrifuged at 200,000 × g for 16 hr at 4°C. Since no visible band was observed, successive 0.5 ml fractions were recovered from the bottom of the tube, the first one supposedly corresponding to 70% sucrose. Each fraction was dialyzed using 20 kDa Slide-A-Lyzers (ThermoFisher) against 40 mM Tris–HCl pH 7.5 to remove the sucrose. These fractions were further concentrated by centrifugation at 15,000 × g, at 4°C for 4 hr and most of the supernatant was removed, leaving ~100 µl of sample at the bottom of each tube. At each step of the extraction procedure the sample was imaged by negative staining transmission electron microscopy (TEM) to assess the integrity of the genomic fiber (Figure 1—figure supplement 1). Each fraction of the gradient was finally controlled by negative staining TEM. For proteomic analysis, an additional step of concentration was performed by speedvac (Savant SPD131DDA, Thermo Scientific).

Negative stain TEM

Request a detailed protocol

300-mesh ultra-thin carbon-coated copper grids (Electron Microscopy Sciences, EMS) were prepared for negative staining by adsorbing 4–7 µl of the sample for 3 min, followed by two washes with water before staining for 2 min in 2% uranyl acetate. The grids were imaged either on a FEI Tecnai G2 microscope operated at 200 keV and equipped with an Olympus Veleta 2k camera (IBDM microscopy platform, Marseille, France); a FEI Tecnai G2 microscope operated at 200 keV and equipped with a Gatan OneView camera (IMM, microscopy platform, France) or a FEI Talos L120c operated at 120 keV and equipped with a Ceta 16M camera (CSSB multi-user cryo-EM facility, Germany, Figure 1, Figure 1—figure supplement 1).

Cryo-electron tomography

Sample preparation

Request a detailed protocol

For cryo-electron tomography (cryo-ET) of the mimivirus genomic fiber, samples were prepared as described above for single-particle analysis except that 5 nm colloidal gold fiducial markers (UMC, Utrecht) were added to the sample right before plunge freezing at a ratio of 1:2 (sample:fiducial markers).

Data acquisition

Request a detailed protocol

Tilt series were acquired using SerialEM (Mastronarde, 2005) on a Titan Krios (Thermo Scientific) microscope operated at 300 keV and equipped with a K3 direct electron detector and a GIF BioQuantum energy filter (Gatan). We used the dose-symmetric tilt-scheme (Hagen et al., 2017) starting at 0° with a 3° increment to ±60° at a nominal magnification of ×64,000, a pixel size of 1.4 Å and a total fluence of 150 e−/Ų over the 41 tilts, that is, ~3.7 e−/Ų/tilt for an exposure time of 0.8 s fractionated into 0.2 s frames (Supplementary file 4).

Data processing

Request a detailed protocol

Tilt series were aligned and reconstructed using the IMOD (Kremer et al., 1996). For visualization purposes, we applied a binning of 8 and SIRT-like filtering from IMOD (Kremer et al., 1996) as well as a bandpass filter bsoft (Heymann and Belnap, 2007). The tomograms have been deposited on EMPIAR, accession number 1131 and videos were prepared with Fiji (Figure 1E, F, Figure 1—figure supplement 2, and Figure 1—video 1; Figure 1—video 2; Figure 1—video 3; Figure 1—video 4).

Agarose gel electrophoresis and DNA dosage to assess the presence of DNA into the fiber

Request a detailed protocol

Genomic DNA was extracted from 1010 virus particles using the PureLink TM Genomic DNA mini kit (Invitrogen) according to the manufacturer’s protocol. Purified genomic fiber was obtained following the method described above. The purified fiber was treated by adding proteinase K (PK) (Takara ST 0341) to 20 µl of sample (200 ng as estimated by dsDNA Qubit fluorometric quantification) at a final concentration of 1 mg/ml and incubating the reaction mix at 55°C for 30 min. DNase treatment was done by adding DNase (Sigma 10104159001) and MgCl2 to a final concentration of 0.18 mg/ml and 5 mM, respectively, in 20 µl of sample and incubated at 37°C for 30 min prior to PK treatment. For RNase treatment, RNase (Sigma SLBW2866) was added to 20 µl (200 ng) of sample solution to a final concentration of 1 mg/ml and incubated at 37°C for 30 min prior to PK treatment. All the samples were then loaded on a 1% agarose gel and stained with ethidium bromide after migration. The bands above the 20 kbp marker correspond to the stacked dsDNA fragments of various lengths compatible with the negative staining images of the broken fibers where long DNA fragment are still attached to the helical structure. The mimivirus purified genomic DNA used as a control migrates at the same position (Figure 1—figure supplement 3).

Cryo-EM bubblegram analysis

Request a detailed protocol

Samples were prepared as described for single-particle analysis. Dose series were acquired on a Titan Krios (Thermo Scientific) microscope operated at 300 keV and equipped with a K3 direct electron detector and a GIF BioQuantum (Gatan) energy filter. Micrographs were recorded using SerialEM (Mastronarde, 2005) at a nominal magnification of ×81,000, a pixel size of 1.09 Å, and a rate of 15 e−/pixel/s (Figure 1—figure supplement 4 and Supplementary file 4). Dose series were acquired by successive exposures of 6 s, resulting in an irradiation of 75 e−/Ų per exposure. Micrographs were acquired with 0.1 s frames and aligned in SerialEM (Mastronarde, 2005). In a typical bubblegram experiment, 12–15 successive exposures were acquired in an area of interest with cumulative irradiations of 900–1125 e−/Ų total (Figure 1—figure supplement 4).

Single-particle analysis by cryo-EM

Sample preparation

Request a detailed protocol

For single-particle analysis, 3 µl of the purified sample were applied to glow-discharged Quantifoil R 2/1 Cu grids, blotted for 2 s using a Vitrobot Mk IV (Thermo Scientific) and applying the following parameters: 4°C, 100% humidity, blotting force 0, and plunge frozen in liquid ethane/propane cooled to liquid nitrogen temperature.

Data acquisition

Request a detailed protocol

Grids were imaged using a Titan Krios (Thermo Scientific) microscope operated at 300 keV and equipped with a K2 direct electron detector and a GIF BioQuantum energy filter (Gatan). 7656 movie frames were collected using the EPU software (Thermo Scientific) at a nominal magnification of ×130,000 with a pixel size of 1.09 Å and a defocus range of −1 to −3 μm. Micrographs were acquired using EPU (Thermo Scientific) with 8-s exposure time, fractionated into 40 frames and 7.5 e−/pixel/s (total fluence of 50.5 e−/Ų) (Supplementary file 4).

2D classification and clustering of 2D classes

Request a detailed protocol

All movie frames were aligned using MotionCor2 (Zheng et al., 2017) and used for contrast transfer function (CTF) estimation with CTFFIND-4.1 (Rohou and Grigorieff, 2015). Helical segments of the purified genomic fibers, manually picked with Relion 3.0 (He and Scheres, 2017; Scheres, 2012), were initially extracted with different box sizes, 400 pixels for 3D reconstructions, 500 pixels for initial 2D classifications and clustering, and 700 pixels to estimate the initial values of the helical parameters. Particles were subjected to reference-free 2D classification in Relion 3.1.0 (He and Scheres, 2017; Scheres, 2012), where multiple conformations of the fiber were identified (Figure 1—figure supplements 5 and 6).

We then performed additional cluster analysis of the 194 initial 2D classes provided by Relion (Figure 1—figure supplement 5) to aim for more homogeneous clusters, eventually corresponding to different states (Figure 1—figure supplement 6). A custom two-step clustering script was written in python with the use of Numpy (Harris et al., 2020) and Scikit-learn (Pedregosa et al., 2011) libraries. First, a few main clusters were identified by applying a DBSCAN (Hahsler, 2019) clustering algorithm on the previously estimated fiber external width values (W1). The widths values, estimated by adjusting a parameterized cross-section model on each 2D stack, range from roughly 280 to 340 Å. The cross-section model fitting process is based on adjusting a section profile (S) described in the equation bellow on the longitudinally integrated 2D-class profile (Figure 1—figure supplement 6A). The cross-section model is composed of a positive cosine component (parameterized by the center position μ, its width σ1 and amplitude a1) associated with the fiber external shell, a negative cosine component (parameterized by the center position μ, its width σ2 and amplitude a2) associated with the central hollow lumen (W2), and a constant p0, accounting for the background level, as.

S(μ, σ1,a1,σ2,a2,p0,x)=a1cos⁡x−μσ1−a2cos⁡x−μσ2+p0

Then, as a second step, each main cluster was subdivided into several subclusters by applying a KMEANS (Mannor et al., 2011) clustering algorithm on a pairwise similarity matrix. This similarity metric was based on a 2D image cross-correlation scheme, invariant to image shifts, and mirroring (Guizar-Sicairos et al., 2008). The number of subclusters was manually chosen by visual inspection. For the most populated 2D classes corresponding to the most compact conformations, the number of subclusters was small: n = 2 subclusters, Cl1a and Cl1b, and n = 1 subcluster for the intermediate class, Cl2. However, the number of subclusters was higher (n = 5) for the relaxed conformation (Cl3 from green to purple in Figure 1—figure supplement 6), highlighting the overall heterogeneity of our dataset with compact, intermediate, relaxed states and even loss of one protein strands (Cl3b) and unwound ribbons.

Identification of candidate helical parameters

Request a detailed protocol

Fourier transform analysis methods have been used to identify helical parameters candidates (Coudray et al., 2016; Diaz et al., 2010; Sachse, 2015) for the Cl1a, Cl2, and Cl3a clusters. For each cluster, we first estimated the repeat distance by applying the method consisting in a longitudinal autocorrelation with a windowed segment of the real-space 2D class of fixed size (100 Å) (Diaz et al., 2010). Then, a precise identification of the power spectrum maxima could be achieved on a high signal-to-noise ratio power spectrum, obtained by averaging all the constituting segments in the Fourier domain, which helps lowering the noise, and fill in the CTF zeros regions. The best candidates were validated with Helixplorer (http://rico.ibs.fr/helixplorer/).

For the Cl1a cluster, the parameters of a 1-start helix have been identified with a rise of 7.937 Å and a twist of 221.05°. For the Cl2 cluster, the candidate parameters are a rise of 20.383 Å and a twist of 49.49° and C3 cyclic symmetry. For the Cl3a cluster, the candidate parameters correspond to a rise of 31.159 Å and a twist of 24° and D5 symmetry (Figure 1—figure supplement 6C–E).

Cryo-EM data processing and 3D reconstruction

Request a detailed protocol
- Cl1a–Cl3a
Request a detailed protocol

After helical parameters determination, two last 2D classifications were performed on segments extracted with a box size of 400 pixels (decimated to 200 pixels) using the proper rises for the most compact Cl1a (7.93 Å, 113,026 segments, overlap ~98.2%) and the relaxed Cl3a (31.16 Å, 16,831 segments, overlap ~92.9%). Values of the helical parameters (rise and twist) were then used for Relion 3D classification (He and Scheres, 2017; Scheres, 2012), with a ±10% freedom search range, using a featureless cylinder as initial reference (diameter of 300 Å for the compact particles Cl1a and 340 Å for the relaxed particles Cl3a). The superimposable 3D classes (same helical parameters, same helix orientation) were then selected, reducing the dataset to 95,722 segments for the compact fiber (Cl1a) and to 15,289 segments for the relaxed fiber (Cl3a). After re-extraction of the selected segments without scaling, further 3D refinement was performed with a 3D classification output low-pass filtered to 15 Å as reference. With this, the maps were resolved enough (Cl1a: 4.4 Å, Cl3a: 4.8; FSC threshold 0.5) to identify secondary structure elements (with visible cylinders corresponding to the helices) (Figure 2—figure supplements 1–2).

- Cl2
Request a detailed protocol

The 12 2D classes corresponding to segments of the Cl2 conformation, were extracted with a box size of 400 pixels (decimated to 200 pixels, rise 20.4 Å, 5,775 segments, overlap ~94.9%). They were used in Relion for 3D refinement with the helical parameters values identified previously, with a ±10% freedom search range, using a 330 Å large featureless cylinder as initial reference and resulted in a 7.1-Å map (FSC threshold 0.5) and C3 cyclic symmetry (Figure 2—figure supplement 3).

Focused refinement of a single oxidoreductase dimer

Request a detailed protocol

The Cl1a 3D map was used to make a mask corresponding to a single dimer of oxidoreductases through the segmentation module (Pintilie et al., 2010) of the program Chimera (Pettersen et al., 2004) (this mask was deposited as part of the EMDB D_1292117739). With Relion and this mask we performed partial signal subtraction (Bai et al., 2015) to remove the information of the other dimers from the experimental images generating a new stack of subtracted images. This was followed by focused refinement with an initial model created (command "relion_reconstruct") from the subtracted dataset and the corresponding orientation/centering parameters from the partial signal subtraction. CTF parameters refinement was performed followed by a last 3D refinement step and postprocessing (B-factor applied −45). This led to the best resolved 3.3-Å 3D map (FSC threshold 0.5, masked, Figure 2—figure supplement 1) that was used to build the atomic model of the GMC oxidoreductase and its ligands (Supplementary file 1, Figure 3). As expected, densities corresponding to DNA were still visible in the final 3D map because the 3D density used during partial signal subtraction contained no high-resolution information about the DNA structure. A Relion project directory summarizing all the steps to perform the focused refinement of the asymmetric unit (oxidoreductase dimer) is available under the link: https://src.koda.cnrs.fr/igs/genfiber_cl1a_focusrefine_relion_pipeline.

DNA focused refinement

Request a detailed protocol

The 3D EM maps of Cl1a and Cl2 were used to make masks for each DNA strand separately with ChimeraX (Goddard et al., 2018). Each strand was used in a partial signal subtraction removing information from all proteins keeping only information near the presumed DNA regions on the corresponding experimental images. The subtracted (DNA only) datasets were merged and subjected to 2D classification to allow visual assessment of the quality of the subtracted images (Figure 3—figure supplement 1). 3D refinement was performed followed by 3D classification and final refinement of the best 3D classes (Figure 3E, Figure 3—figure supplement 2).

For the Cl3a compaction state (where some barely visible densities could correspond to remaining DNA) a cylinder with a diameter corresponding to the lumen of the relaxed structure (140 Å) was used as mask for partial signal subtraction in an attempt to enhance the information from the presumed DNA regions. However, the signal remained too weak to be recognized as such.

Automatic picking of the three different conformations (Cl1a, Cl2, and Cl3a)

Request a detailed protocol

Projections from the different helical maps were used as new input for automatic picking of each Cl1a, Cl2, and Cl3a clusters to get more homogeneous datasets for each conformation (Estrozi and Navaza, 2008). For each dataset a final round of extraction (box size Cl1a: 380 pixels, 121,429 segments; Cl2: 400 pixels, 8479 segments; Cl3a: 400 pixels, 11,958 segments) and 3D refinement with solvent flattening (Central z length 30%) was performed using the appropriate helical parameters and additional symmetries (none for Cl1a, C3 for Cl2, and D5 for Cl3a). This led to the improved maps presented in Figures 2 and 3, Figure 2—figure supplements 1–3 (Cl1a: 3.7 Å; Cl2: 4.0 Å; Cl3a: 3.7 Å; FSC threshold 0.5, masked). Postprocessing was performed with B-factor −80 (Supplementary file 1, Figure 2—figure supplements 1–3).

Structures refinement

Request a detailed protocol

The resolution of the EM map enabled to fit the R135 dimeric structure (Klose et al., 2015) (PDB 4Z24) into the maps using UCSF Chimera 1.13.1 (Pettersen et al., 2004). The qu_143 and qu_946 models were obtained using SWISS-MODEL (Waterhouse et al., 2018) (closest PDB homolog: 4Z24). It is only at that stage that the best fitted qu_946 model was manually inspected and additional N-terminal residues built using the extra density available in the cryo-EM map of the 5-start compact reconstruction (Figure 3). The entire protein shell built using the corresponding helical parameters were finally fitted into the Cl1a and Cl3a maps and were further refined against the map using the real-space refinement program in PHENIX 1.18.2 (Liebschner et al., 2019; Figure 2—figure supplements 1 and 2). The qu_143 and qu_946 models were also used to build the entire shell using the Cl2 helical parameters and symmetry and were further refined using the real-space refinement program in PHENIX 1.18.2 (Liebschner et al., 2019; Figure 2—figure supplement 3). Validations were also performed into PHENIX 1.18.2 (Liebschner et al., 2019) using the comprehensive validation program and statistics in Supplementary file 1 correspond to the qu_143 model. The qu_946 model was manually corrected and ultimately refined and validated into PHENIX 1.18.2 (Liebschner et al., 2019) using the highest resolution focused refined Cl1a map. In that map the two first amino acid are disordered including the cysteine.

MS-based proteomic analysis of mimivirus virion and genomic fiber

Request a detailed protocol

Proteins extracted from total virions and purified fiber were solubilized with Laemmli buffer (4 volumes of sample with 1 volume of Laemmli 5× – 125 mM Tris–HCl pH 6.8, 10% sodium dodecyl-sulfate (SDS), 20% glycerol, 25% β-mercaptoethanol, and traces of bromophenol blue) and heated for 10 min at 95°C. Three independent infections using three different batches of virions were performed and the genomic fiber was extracted from the resulting viral particles to analyze three biological replicates. Extracted proteins were stacked in the top of an SDS–polyacrylamide gel electrophoresis PAGE gel (4–12% NuPAGE, Life Technologies), stained with Coomassie blue R-250 (BioRad) before in-gel digestion using modified trypsin (Promega, sequencing grade) as previously described Casabona et al., 2013. Resulting peptides were analyzed by online nanoliquid chromatography coupled to tandem MS (UltiMate 3000 RSLCnano and Q-Exactive Plus, Thermo Scientific). Peptides were sampled on a 300 µm × 5 mm PepMap C18 precolumn and separated on a 75 µm × 250 mm C18 column (Reprosil-Pur 120 C18-AQ, 1.9 μm, Dr. Maisch) using a 60 min gradient for fiber preparations and a 140 min gradient for virion. MS and MS/MS data were acquired using Xcalibur (Thermo Scientific). Peptides and proteins were identified using Mascot (version 2.7.0) through concomitant searches against mimivirus reunion, classical contaminant databases (homemade), and the corresponding reversed databases. The Proline software (Bouyssié et al., 2020) was used to filter the results: conservation of rank 1 peptides, peptide score ≥25, peptide length ≥6, peptide-spectrum-match identification false discovery rate <1% as calculated on scores by employing the reverse database strategy, and minimum of 1 specific peptide per identified protein group. Proline was then used to perform a compilation and MS1-based quantification of the identified protein groups. Intensity-based absolute quantification (iBAQ) (Schwanhäusser et al., 2011) values were calculated from MS intensities of identified peptides. The viral proteins detected in a minimum of two replicates are reported with their molecular weight, number of identified peptides, sequence coverage, and iBAQ values in each replicate. The number of copies per fiber for each protein was calculated according to their iBAQ value based on the GMC-oxidoreductase iBAQ value (see below, Supplementary file 2). Mapping of the identified peptides for the GMC oxidoreductases are presented in Figure 2—figure supplement 4.

Protein/DNA ratio validation

Request a detailed protocol

To compare the theoretical composition of the genomic fiber accommodating a complete genome with the experimental concentrations in protein and DNA of the sample, we performed DNA and protein quantification (Qubit fluorometric quantification, Thermo Fischer Scientific) on two independent samples of the purified genomic fiber. This returned a concentration of 1.5 ng/µl for the dsDNA and 24 ng/µl for the proteins in one sample and 10 ng/µl for the dsDNA and 150 ng/µl for the protein in the second sample (deposited on agarose gel, Figure 1—figure supplement 3). Considering that all proteins correspond to the GMC-oxidoreductase subunits (~71 kDa), in the sample there is a molecular ratio of 7.2 dsDNA basepairs per GMC-oxidoreductase subunit in the first sample and 7.68 in the second. Based on our model, a 5-start genomic fiber containing the complete mimivirus genome (1,196,989 bp) should be composed of ~95,000 GMC-oxidoreductases subunits (~93,000 for a 6-start). This gives a molecular ratio of 12.5 dsDNA basepairs per GMC-oxidoreductase subunit, which would be more than experimentally measured. However, in the cryo-EM dataset, there are at most 50% of genomic fibers containing the DNA genome (mostly Cl1), while some DNA strands can be observed attached to the relaxed genomic fiber Cl3 but the vast majority of released DNA was lost during the purification of the genomic fiber. Applying an estimated loss of 50% to the total DNA compared to the measured values, we obtain a ratio, which is in the same order of magnitude of the ones measured in the purified genomic fiber samples.

Model visualization

Request a detailed protocol

Molecular graphics and analyses were performed with UCSF Chimera 1.13.1 (Pettersen et al., 2004) and UCSF ChimeraX 1.1 (Goddard et al., 2018), developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, with support from National Institutes of Health R01-GM129325 and the Office of Cyber Infrastructure and Computational Biology, National Institute of Allergy and Infectious Diseases.

Analysis of macromolecules in NS-TEM images of mimivirus genomic fiber

Request a detailed protocol

Manually picked positions of the particles of interest have been used to automatically extract 193-Å square area from the micrographs (Figure 4C (E)). Then, manual clipping of the particle from its noisy background has been achieved using GIMP’s intelligent scissors and a smooth transparent to black mask, producing the clean extraction image (Figure 4C (CE)). This clipping step was applied in order to improve the semi-automatic identification of the closest orientation in all RNA polymerase projections. The projection views of the different RNA polymerase were produced by converting the PDB model into a volume density through EMAN2’s e2pdb2mrc software, and converting the volume into a 2D projection using a dedicated python script, ultimately applying a Gaussian blur filter (σ = 2.9 Å) in order to roughly simulate the whole imaging transfer function. The projections dataset of all orientations of the vaccinia virus RNA polymerase structure was produced with a 5° rotation step in all angles (PDB: 6RIC, equivalent subunits identified by our MS-based proteomics of purified mimivirus genomic fibers). Preferred projections (Figure 4C (P)) orientations were manually assessed.

Data availability

Mimivirus reunion genome has been deposited under the following accession number: BankIt2382307 Seq1 MW004169. 3D reconstruction maps and the corresponding PDB have been deposited to EMDB (Deposition number Cl1a: 7YX4, EMD-14354; Cl1a focused refined: D_1292117739; Cl3a: 7YX5, EMD-14355; Cl2: 7YX3, EMD-14353). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD021585 and 10.6019/PXD021585. The tomograms have been deposited in EMPIAR under accession number 1131 and tomograms video is provided with the article.

The following data sets were generated
    1. Brun V
    2. Coute Y
    (2021) PRIDE
    ID PXD021585. Proteomic analysis of virion and genomic fibre of Mimivirus reunion.
    https://www.ebi.ac.uk/pride/archive/projects/PXD021585

References

    1. Goddard TD
    2. Huang CC
    3. Meng EC
    4. Pettersen EF
    5. Couch GS
    6. Morris JH
    7. Ferrin TE
    (2018)
    Ucsf chimerax: Meeting modern challenges in visualization and analysis: Ucsf chimerax visualization system
    Protein Science: A Publication of the Protein Society 27:14–25.
  47. Book
    1. Mannor S
    2. Jin X
    3. Han J
    4. Jin X
    5. Han J
    6. Jin X
    7. Han J
    8. Zhang X
    (2011)
    K-Means Clustering
    Springer.
    1. Pedregosa F
    2. Varoquaux G
    3. Gramfort A
    4. Michel V
    5. Thirion B
    6. Grisel O
    7. Blondel M
    8. Prettenhofer P
    9. Weiss R
    10. Dubourg V
    11. Vanderplas J
    12. Passos A
    13. Cournapeau D
    14. Brucher M
    15. Perrot M
    16. Duchesnay É
    (2011)
    Scikit-learn: machine learning in python
    Journal of Machine Learning Research: JMLR 12:2825–2830.

Decision letter

  1. Adam Frost
    Reviewing Editor; University of California, San Francisco (Adjunct), United States
  2. Sara L Sawyer
    Senior Editor; University of Colorado Boulder, United States
  3. Adam Frost
    Reviewer; University of California, San Francisco (Adjunct), United States
  4. Jonatas S Abrahão
    Reviewer; Universidade Federal de Minas Gerais, Brazil

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "The giant Mimivirus 1.2 Mb genome is elegantly organized into a 30 nm helical protein shield" for consideration by eLife. Your article has been reviewed by 4 peer reviewers, including Adam Frost as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by a Reviewing Editor and Sara Sawyer as the Senior Editor. The following individual involved in the review of your submission has agreed to reveal their identity: Jonatas S Abrahão (Reviewer #3).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1. Several reviewers expressed uncertainty about the significance of the structural observations for our understanding of viral behavior. Critically, one reviewer noted: "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." Another review notes that the answer to this question should address "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?" Another reviewer noted "The presented data do not [enable us] to estimate the amount of [the] mimivirus genome organized into 30 nm diameter filaments. Hence, the title of the paper is [overreaching]" and "The filamentous structures [result from] an 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 [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?" Addressing these comments is essential.

2. The reviewers requested a more comprehensive description of the GMC oxidoreductase paralogs and the properties of the helical coat. One reviewer noted "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 Figure 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." And multiple reviewers requested a more detailed description of whether the helically averaged density map is simply unable to distinguish between qu_946 and qu_143, including a description of the amino acid percent identity versus similarity, especially for the contact-forming surface residues that govern the protein-protein and protein-nucleic acid contacts?

Another reviewer noted "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." A reviewer wrote concerning the properties of the coat "[the] authors could better explain why we only see 20 kb fragments in the gel, including in the control (in Figure S2)." Another reviewer wrote "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?" Another wrote, "Rough estimation of genome compaction to fit into the nucleoid" section. Even though these back-of-the-envelope-type calculations appear to be reasonable, the last sentence "As a result, the Mimivirus genome is probably organized…" throws a monkey wrench into all of this. I see no resemblance in the organizations of the APBV1 genome and the DNA in mimivirus fibers. Finally, an EM reviewer wrote "I am not yet convinced the authors have resolved the putative disulfide bridge between protomers. Please make a figure of the density around this bond at a more stringent map threshold so that the reader can appreciate the strength of the EM evidence for the disulfide." Addressing these comments is essential.

There were also less critical but still valuable requests to consider:

3. Several reviewers were unsure how to think about the "balls of yarn." One reviewer wrote "slightly puzzled by the observed "ball of yarn". It is hard for me to imagine that a cylindrical container/fiber-containing 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?" Another reviewer wrote "The "ball of yarn" analogy is nice, but Figure 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 [a] "ball of yarn". Like several threads connected. Could [the] authors clarify?" Another reviewer wrote "wondering if authors attempted to get [scanning or transmission] electron microscopy images of mimivirus with exposed ball(s) of yarn?"

4. The methods would benefit from more detailed descriptions. Multiple reviewers agree that the methods section on focused refinements for both the dimer of oxidoreductases and the DNA within needs much more explanation for other groups to reproduce. Please enhance the methods description and consider including example scripts. Also, please include the output from Helixplorer (http://rico.ibs.fr/helixplorer/).

5. The reviewers suggest the following modifications to which and where certain figures are included. At least two EM reviewers agree that "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." and would favor minimizing arguments that depend on these data. Another reviewer wrote "Panels C and D of Figure 4 are too speculative to be included in the main text. Panels A and B are fascinating and pose a hypothesis worthy of further investigation, but please omit panels C and D." This request was modified with a great suggestion by another reviewer "All images of the putative RNA polymerases should be boxed out from the panel A (from the whole photograph, that is) and a collage of these boxes should be shown on the same scale as the Poxvirus RNAP. The latter should be shown in several orientations. Perhaps, projection views instead of the molecular surface (which is shown in Figure 4C) would match the negatively stained images better (but maybe not). Panel 4B should be shown on the same scale as panel A and the putative RNAP can be placed (in color) into the channel of the fiber."

6. There were a number of suggestions to improve the readability and generality of the text.

– L140: "single DNA strand" but the authors probably mean single molecule of dsDNA or single double-helix. Please revise to avoid ambiguity.

– L178: Please soften "must be" – no evidence is presented that any of the mimivirus proteases are involved in the processing of oxidoreductases. The involvement of cellular proteases cannot be a priori excluded.

– L44 Please consider the flexibility of "giant virus" concept. The term was first used in 1999 to refer to Paramecium bursaria chlorella virus, which is substantially smaller than mimivirus. As the term was inaugurated in the context of chloroviruses, would be friendly if the authors state that "Giant viruses OF AMOEBAS were discovered with the isolation of mimivirus".

– Mimivirus or mimivirus? Please see https://talk.ictvonline.org/information/w/faq/386/how-to-write-virus-species-and-other-taxa-names.

– Figure S12: fiber or fibre?

– Medusavirus genome harbored genes for all five types of histones (H1, H2A, H2B, H3, and H4). Please consider adding this to the discussion topic.

L. 303-304. APBV1 capsid structure appears strikingly similar to the Mimivirus genomic fiber but Mimivirus… I am sorry but I do not see any resemblance.

L. 334-338. This philosophical excursion might be correct, but I do not see any data supporting these hypotheses.

The grammar needs to be corrected in many places starting with the title " The giant Mimivirus 1.2 Mb genome is elegantly organized into a 30 nm-DIAMETER helical protein shield".

L. 79. "… resulted in the observation…" <- simplify by removing unnecessary words.

L. 80. "The used capsid opening procedure involves…" <- remove "used".

L. 83-84. "a same" -> THE same.

Reviewer #1 (Recommendations for the authors):

1. The methods section on focused refinements for both the dimer of oxidoreductases and the DNA within needs much more explanation for other groups to reproduce. Please enhance the methods description and consider including example scripts.

2. I am not yet convinced the authors have resolved the putative disulfide bridge between protomers. Please make a figure of the density around this bond at a more stringent map threshold so that the reader can appreciate the strength of the EM evidence for the disulfide.

3. Panels C and D of Figure 4 are too speculative to be included in the main text. Panels A and B are fascinating and pose a hypothesis worthy of further investigation, but please omit panels C and D.

4. Please include the outputs Helixplorer (http://rico.ibs.fr/helixplorer/)

Reviewer #2 (Recommendations for the authors):

L140: "single DNA strand" but the authors probably mean single molecule of dsDNA or single double-helix. Please revise to avoid ambiguity.

L178: Please soften "must be" – no evidence is presented that any of the mimivirus proteases are involved in the processing of oxidoreductases. The involvement of cellular proteases cannot be a priori excluded.

Reviewer #3 (Recommendations for the authors):

– In addition to information that is in public preview, I recommend that authors consider the flexibility of "giant virus" concept. The term was first used in 1999 to refer to Paramecium bursaria chlorella virus, which is substantially smaller than mimivirus. As the term was inaugurated in the context of chloroviruses, would be friendly if the authors state that "Giant viruses OF AMOEBAS were discovered with the isolation of mimivirus" (eg line 44);

– Mimivirus or mimivirus? please see https://talk.ictvonline.org/information/w/faq/386/how-to-write-virus-species-and-other-taxa-names.

– Figure S12: fiber or fibre?

– Please consider promoting Figure S5 as Figure 5.

– I am wondering if the authors attempted to get scanning electron microscopy images of mimivirus with an exposed ball of yarn. Sounds good.

– Medusavirus genome harbored genes for all five types of histones (H1, H2A, H2B, H3, and H4). Please consider adding this to the discussion topic.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "The giant mimivirus 1.2 Mb genome is elegantly organized into a 30 nm diameter helical protein shield" for further consideration by eLife. Your revised article has been evaluated by Sara Sawyer (Senior Editor) and Reviewing Editor Adam Frost. The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

First, we find the uncertainty and controversy around the role and identification of the GMC oxidoreductases confusing. Please upload the cryo-EM maps and fitted atomic coordinates for the reviewers. Second, while we understand your decision to describe the genetic system and the genomic fibers of the M4 mutant in a separate publication, we feel that an additional explanation regarding the M4 mutant is necessary for this manuscript. The observation that the genomic fibers are constructed from different proteins in other mimiviruses, and potentially in the M4 mutant which lacks qu_946 and qu_143 poses quite a puzzle and raises concerns about the validity of the structural assignment.

https://doi.org/10.7554/eLife.77607.sa1

Author response

Essential revisions (for the authors):

1. Several reviewers expressed uncertainty about the significance of the structural observations for our understanding of viral behavior. Critically, one reviewer noted: "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.”

These reviewers are totally right and this point also bothered us. We thus managed to extract the genomic fiber of M4 (the isolate without GMC oxidoreductases). The fiber also has a rodshaped structure but protein composition analysis of the purified fiber shows that different proteins are involved in its assembly.

We do not think GMC-oxidoreductases are essential. The packaging machinery appears to be able to use a variety of the most abundant proteins available at the stage of genome packaging. This corresponds to the GMC-oxidoreductases for mimivirus but is another protein in the case of M4. Follow up studies we are currently conducting clearly show the GMC-oxidoreductase is not essential for the genomic fiber formation.

The focus of the current manuscript is to report the structure of mimivirus genome organization in the genomic fiber, which is already a first piece of work. We consider the study of M4 genomic fiber, and the testing of the hypotheses this finding raises, as follow up studies. We would prefer not to discuss this in the current manuscript and hope the reviewers and the editor will believe our preliminary data and will agree on keeping this information for a subsequent publication.

Another review notes that the answer to this question should address "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?"

The transcriptomic data of mimivirirus prototype infectious cycle have been published and corresponding data are publicly available under this link: http://www.igs.cnrsmrs.fr/mimivirus/ (clicking on mimivirus Genome Browser). They report both 454 and Solid data.

Author response image 1 shows the table of the genes with the keyword “Polymerase” in the answers to the reviewers for Mimivirus prototype (pink) and Mimivirus reunion (blue) to help them identify the equivalence between the different subunits.

Author response image 1
Download asset Open asset

The GMC-oxidoreductase genes are expressed after DNA replication, at the time of virion assembly, starting at 5h Post Infection (PI) until the end of the cycle. For reference, the DNA polymerase starts being expressed after 1h until the end, with a peak at 3 hours PI. The RNA polymerase subunits start also being expressed after 1 h PI until the end of the cycle. The RNA polymerase subunits are loaded in the virions as well as the mRNA maturation machinery (PolyA polymerase, mRNA capping enzyme, etc…) (supplementary file 2B).We now added a sentence in the main text reporting the late expression of the GM-Coxidoreductases and the fact that they start being expressed after DNA replication (p5, main text). In the section “Additional proteins, including RNA polymerase subunits, are enriched in the genomic fiber” we also provide a sentence on RNA polymerase subunits and GM-Coxidoreductases, kinesin, regulator of chromosome condensation expression, all expressed until the end of the cycle (main text, p9).

Another reviewer noted "The presented data do not [enable us] to estimate the amount of [the] mimivirus genome organized into 30 nm diameter filaments.

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 into this compartment. This is only possible due to the high flexibility of the genomic fiber. So, the entire genome corresponds to ~40 µm of genomic fiber, which when packaged as a ball can entirely fit into the nucleoid (~350 nm diameter).

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, Figure 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”.

We believe the Reviewers 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 and that also defines the size of the nucleoid in the icosahedral capsid. This tight packaging is exemplified in Figure 1A for instance or in the AFM images of the nucleoid in Kuznetsov, Y. G. et al. Virology 2010; Kuznetsov YG et al. J. Virol. 2013, Figure 15.

Hence, the title of the paper is [overreaching]" and "The filamentous structures [result from] an extremely harsh treatment of the virus particle, which starts with a 1.5 hour-long incubation at pH 2.

There is a misunderstanding here. 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 Figure 1—figure supplement 1A. Acidic treatment was never applied to produce the genomic fiber as we noticed it is sensitive to temperature and acidic pH. 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).

Do the filaments actually exist inside the virus particle as the title of the paper implies?

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, Figure 15).

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?"

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:

  1. 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 (Figure 3-animation 1 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.

  2. 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 strands of the ribbon, one after the other.

  3. The contacts between chains intra and inter strand are also scarce supporting an active assembly of the structure. We now provide an additional supplementary file 3 with the different contacts for the different states of the genomic fiber. There are very few interstart contacts.

2. The reviewers requested a more comprehensive description of the GMC oxidoreductase paralogs and the properties of the helical coat. One reviewer noted "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 Figure 2C).

A new Table (Supplementary file 3) has been provided with contacting residues intra and inter strands and their conservation. We also changed Cl2 by Cl1a maps in Figure 3F to show the qu_143 contacts with DNA in the same map as for qu_946 (Figure 3E). Finally, to comply with another reviewer’s question these contacting residues are shown for both qu_946 and qu_143 in the Cl1a, Cl2 and Cl3a maps. Conserved and divergent residues are highlighted in green and red, respectively, in Supplementary file 3. Inter- and intra-strands contacts are clearly loosening between the two compact Cl1a and Cl2 structures, compared to the relaxed Cl3, supporting the unfolding process and the use of an energy driven machinery to build and package the genomic fiber into the capsids.

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." And multiple reviewers requested a more detailed description of whether the helically averaged density map is simply unable to distinguish between qu_946 and qu_143, including a description of the amino acid percent identity versus similarity, especially for the contact-forming surface residues that govern the protein-protein and protein-nucleic acid contacts?

We think this is an excellent suggestion and we thus replaced in Figure 3 previous panels by the corresponding electrostatic views of the protein shell with the DNA (Figure 3C-D). We believe it is interesting to note that the electrostatics are different between qu_946 or qu_143, showing that despite the high level of identity (69%) between the two proteins, there are differences and that this does not impact the packaging of the DNA. Along with the finding that M4 use a different protein to package its DNA, we believe this supports even further an energy driven packaging machinery.

We now added a sentence in the main text p5:” The two mimivirus reunion proteins share 69% identity (81% similarity).”

Another reviewer noted "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."

If reviewers still think it would be useful, we can provide a multiple alignment of the GM-Coxidoreductases of prototype members of the different clades of the family as a supplementary figure.

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 fiber.

Author response image 2
Download asset Open asset

A reviewer wrote concerning the properties of the coat "[the] authors could better explain why we only see 20 kb fragments in the gel, including in the control (in Figure S2)."

Figure 1—figure supplement 3 (old Figure 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 by the extraction protocol, pipetted several times before being loaded on the gel and we thus did 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 Figure 1—figure supplement 3) 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 (Figure 1, Panel B, Lane M1).

Another reviewer wrote "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 Figure 2—figure supplement 4, when analyzed by MS-based proteomics, the comparison of the peptide coverage of the GMC-oxidoreductases depending if they compose the fibrils or the genomic fiber is not the same. The N-terminal domain is clearly covered when the fibrils or intact virions are analyzed and not covered when the analysis is performed on the purified 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 prior to genomic fiber assembly.

Main text: “The proteomic analyses provided different sequence coverages for the GMC-oxidoreductases depending on whether samples were intact virions or purified genomic fiber preparations, with substantial under-representation of the N-terminal domain in the genomic fiber (Figure 2—figure supplement 4). 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 induce an increase of the genomic fiber diameter that would be too big to be accommodated into the nucleoid.

Another wrote, "Rough estimation of genome compaction to fit into the nucleoid" section. Even though these back-of-the-envelope-type calculations appear to be reasonable, the last sentence "As a result, the Mimivirus genome is probably organized…" throws a monkey wrench into all of this. I see no resemblance in the organizations of the APBV1 genome and the DNA in mimivirus fibers.

The APBV1 nucleocapsid is also a 30 nm diameter helical structure made of proteins with the dsDNA genome lining the interior of the shell. The genomic fiber also looks like a nucleocapsid, but the DNA is circular for APBV1 and linear for mimivirus. The most important difference being that the mimivirus genomic fiber is actually folded in the nucleoid, itself encased in the icosahedral capsid, while APBV1 nucleocapsid is the virion. If this reviewer thinks this is confusing, we could remove the reference to this archaeal virus. We tried to clarify by proposing the following sentence: “As a result, the complete mimivirus genome folded at least 5 times fits into the helical shell. This structure surprisingly resembles a nucleocapsid, such as the archaea infecting APBV1 nucleocapsid.”

Finally, an EM reviewer wrote "I am not yet convinced the authors have resolved the putative disulfide bridge between protomers. Please make a figure of the density around this bond at a more stringent map threshold so that the reader can appreciate the strength of the EM evidence for the disulfide."

We agree with this reviewer that the density does not permit to totally confidently conclude that there is a disulfide bridge. Our rational was as follow: building additional residues into the uninterpreted density brought the cysteines of the two chains close enough to allow a disulfide bridge and fully filled the available density. Clearly in the fully relaxed structure the density is weaker, suggesting there is some additional disorder at this location that led us to think that a disulfide bridge break could fragilize the helix and initiate its decompaction. In Author response image 3A (zoomed in C) is the Cl1a compact map at the same threshold than in the relaxed Cl3 in Author response image 3B (zoomed in D). If the reviewers decide this is convincing, we can add this figure as a supplementary figure.

Author response image 3
Download asset Open asset

We finally thought that since the helical structure is in a decompaction process, the density at that location is an average between locked and unlocked structures explaining why it was less defined than the rest of the structure and thus absent of the focused refined map. In the main text we only suggest a disulfide bridge break could be involved in the unwinding process. We now added “could” in the Figure 3 legend p7: “The isosurface threshold chosen allows visualization of density for the manually built N-terminal residues, including terminal cysteines (stick model), of two neighboring monomers that could form a terminal disulfide bridge.”

There were also less critical but still valuable requests to consider:

3. Several reviewers were unsure how to think about the "balls of yarn." One reviewer wrote "slightly puzzled by the observed "ball of yarn". It is hard for me to imagine that a cylindrical container/fiber-containing 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?

It is the same fiber but the treatments are not exactly the same to produce the relaxing fiber coming out of the capsids and the “ball of yarn” structure, for the latter, there are clearly breakages that give the impression of multiple fibers. The genomic fiber is highly flexible and rolled up in the nucleoid.

In the enclosed cryo-EM micrograph of the expelling virions (Author response image 4) the capsids were opened by a simple proteinase K treatment with 1 mM DTT. The sample was frozen and observed immediately after treatment.

There is no scale bar but the capsids are roughly half a micrometer diameter. Reviewers can also recognize the nucleoid inside most capsids, but with different sizes reflecting the release process progress. The flexibility of the fiber is also visible.

Author response image 4
Download asset Open asset

In the NS-TEM image (Author response image 5), the Reviewers can also see how long and flexible the genomic fiber can be with no breaks.

Author response image 5
Download asset Open asset

Related to this question – is there evidence (e.g., from qPCR) that mimivirus carries a single copy of genomic dsDNA per capsid?"

It would be possible to accommodate many copies of Mimivirus genome as naked DNA but only one copy can be fitted into the helical structure as ~40 µm are needed to fit the folded genome. These 40 µm also correspond to the volume of the nucleoid.

Another reviewer wrote "The "ball of yarn" analogy is nice, but Figure 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 [a] "ball of yarn". Like several threads connected. Could [the] authors clarify?"

At that stage the genomic fiber has been broken due to the multiple steps of extraction, enrichment and purification, which does not happen in vivo. As can be seen in the Cryo-EM picture above, with only one step of treatment, it appears intact and is not broken into fragments.

Another reviewer wrote "wondering if authors attempted to get [scanning or transmission] electron microscopy images of mimivirus with exposed ball(s) of yarn?"

This is a good suggestion but we did not. They are globally 350 nm diameter, as for the nucleoid. The point of this manuscript was to report the structure of the mimivirus genome organization into this helical structure as well as its composition.

4. The methods would benefit from more detailed descriptions. Multiple reviewers agree that the methods section on focused refinements for both the dimer of oxidoreductases and the DNA within needs much more explanation for other groups to reproduce. Please enhance the methods description and consider including example scripts. Also, please include the output from Helixplorer (http://rico.ibs.fr/helixplorer/).

We now expanded the focused refinement procedure in the Material and method section and enclose the output of Helixplorer as supplementary figure (Figure 1—figure supplement 6, C-E).

We also prepared a Relion project directory recapitulating the different steps applied to perform the focused refinement of the asymmetric unit (oxidoreductase dimer). This project directory was made after cleaning (directory filled by different jobs) and renumbering the initial project directory. As we recovered the motion corrected micrographs from the Hamburg platform, the “Import” task of the raw micrographs is not available, but a corresponding task was added to the project directory in order to summarize and have coherence for all steps of the procedure. The different “.mrc” or “.mrcs” files were replaced by empties “.zip” to reduce the disk usage as they are not needed to summarize the processes applied. The directory can be found under this link: https://src.koda.cnrs.fr/igs/genfiber_cl1a_focusrefine_relion_pipeline

5. The reviewers suggest the following modifications to which and where certain figures are included. At least two EM reviewers agree that "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." and would favor minimizing arguments that depend on these data.

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 were just one additional analysis.

Another reviewer wrote "Panels C and D of Figure 4 are too speculative to be included in the main text. Panels A and B are fascinating and pose a hypothesis worthy of further investigation, but please omit panels C and D." This request was modified with a great suggestion by another reviewer "All images of the putative RNA polymerases should be boxed out from the panel A (from the whole photograph, that is) and a collage of these boxes should be shown on the same scale as the Poxvirus RNAP. The latter should be shown in several orientations. Perhaps, projection views instead of the molecular surface (which is shown in Figure 4C) would match the negatively stained images better (but maybe not). Panel 4B should be shown on the same scale as panel A and the putative RNAP can be placed (in color) into the channel of the fiber."

We now changed Figure 4 accordingly and increased the size of Figure 4B. We now present projections of six orientations of the vaccinia virus RNAP corresponding to the best extracted boxes. We must mention that negative staining imaging may dehydrate the objects and change macromolecules volumes. This is also mentioned in Figure 4 legend.

Concerning the last sentence: “Panel 4B should be shown on the same scale as panel A and the putative RNAP can be placed (in color) into the channel of the fiber." We did change the scale of Panel 4B but it is unclear to us if we can leave the colored RNAP fitted in the channel?

6. There were a number of suggestions to improve the readability and generality of the text.

– L140: "single DNA strand" but the authors probably mean single molecule of dsDNA or single double-helix. Please revise to avoid ambiguity.

We modified the text accordingly:

“The dsDNA strands appear as curved cylinders in the helical structure, the characteristic shape of the DNA (minor and major groove) becoming only visible after focused refinement of a single strand of dsDNA (Figure 3, Figure 3—figure supplement 1-2).”

– L178: Please soften "must be" – no evidence is presented that any of the mimivirus proteases are involved in the processing of oxidoreductases. The involvement of cellular proteases cannot be a priori excluded.

We now changed the sentence to: “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.”

– L44 Please consider the flexibility of "giant virus" concept. The term was first used in 1999 to refer to Paramecium bursaria chlorella virus, which is substantially smaller than mimivirus. As the term was inaugurated in the context of chloroviruses, would be friendly if the authors state that "Giant viruses OF AMOEBAS were discovered with the isolation of mimivirus".

This is true and it is now modified in the introduction section.

– Mimivirus or mimivirus? Please see https://talk.ictvonline.org/information/w/faq/386/how-to-write-virus-species-and-other-taxa-names.

We changed Mimivirus by mimivirus all along the manuscript to comply with ICTV rule, as requested by this reviewer.

– Figure S12: fiber or fibre?

This has now been corrected.

– Medusavirus genome harbored genes for all five types of histones (H1, H2A, H2B, H3, and H4). Please consider adding this to the discussion topic.

We did not talk about Medusavirus because it was not proved these histones could organize into nucleosomes.

L. 303-304. APBV1 capsid structure appears strikingly similar to the Mimivirus genomic fiber but Mimivirus… I am sorry but I do not see any resemblance.

The APBV1 nucleocapsid is also a 30 nm diameter helical structure made of proteins with the dsDNA genome lining the interior of the shell. The genomic fiber also looks like a nucleocapsid, but the DNA is circular for APBV1 and linear for mimivirus. The most important difference being that the mimivirus genomic fiber is actually folded in the nucleoid, itself encased in the icosahedral capsid, while APBV1 nucleocapsid is the virion. If this reviewer thinks this is confusing, we could remove the reference to this archaeal virus.

However, to clarify this point we change the sentence in the Result section to:

“As a result, the complete mimivirus genome, folded at least 5 times, fits into the helical shell. This structure surprisingly resembles a nucleocapsid, such as the archaea infecting APBV1 nucleocapsid.”

And in the Discussion section to:

“Yet, APBV1 nucleocapsid structure strikingly resemble the mimivirus genomic fiber with a proteinacious shell enclosing the folded dsDNA genome. Consequently, mimivirus genomic fiber is a nucleocapsid further bundled as a ball of yarn into the nucleoid, itself encased in the large icosahedral capsids.”

L. 334-338. This philosophical excursion might be correct, but I do not see any data supporting these hypotheses.

We now removed this paragraph.

The grammar needs to be corrected in many places starting with the title " The giant Mimivirus 1.2 Mb genome is elegantly organized into a 30 nm-DIAMETER helical protein shield".

L. 79. "… resulted in the observation…" <- simplify by removing unnecessary words.

We replace it by: “produced bundled fibers…”

L. 80. "The used capsid opening procedure involves…" <- remove "used".

Done

L. 83-84. "a same" -> THE same.

This is now corrected.

Reviewer #1 (Recommendations for the authors):

1. The methods section on focused refinements for both the dimer of oxidoreductases and the DNA within needs much more explanation for other groups to reproduce. Please enhance the methods description and consider including example scripts.

2. I am not yet convinced the authors have resolved the putative disulfide bridge between protomers. Please make a figure of the density around this bond at a more stringent map threshold so that the reader can appreciate the strength of the EM evidence for the disulfide.

3. Panels C and D of Figure 4 are too speculative to be included in the main text. Panels A and B are fascinating and pose a hypothesis worthy of further investigation, but please omit panels C and D.

4. Please include the outputs Helixplorer (http://rico.ibs.fr/helixplorer/)

All this reviewer comments have been addressed and answered above.

Helixplorer output are now provided in the supplementary material (Figure 1—figure supplement 6C-E).

Reviewer #2 (Recommendations for the authors):

L140: "single DNA strand" but the authors probably mean single molecule of dsDNA or single double-helix. Please revise to avoid ambiguity.

Thanks for pointing this out, this is now clarified.

L178: Please soften "must be" – no evidence is presented that any of the mimivirus proteases are involved in the processing of oxidoreductases. The involvement of cellular proteases cannot be a priori excluded.

The sentence has been modified accordingly.

Reviewer #3 (Recommendations for the authors):

– In addition to information that is in public preview, I recommend that authors consider the flexibility of "giant virus" concept. The term was first used in 1999 to refer to Paramecium bursaria chlorella virus, which is substantially smaller than mimivirus. As the term was inaugurated in the context of chloroviruses, would be friendly if the authors state that "Giant viruses OF AMOEBAS were discovered with the isolation of mimivirus" (eg line 44);

This is totally true and this has now been corrected.

– Mimivirus or mimivirus? please see https://talk.ictvonline.org/information/w/faq/386/how-to-write-virus-species-and-other-taxa-names.

This reviewer is right but it is always difficult to change old habits! This has now been changed all along the manuscript.

– Figure S12: fiber or fibre?

Thank you for pointing this out, it is now corrected. –

– Please consider promoting Figure S5 as Figure 5.

We can if requested by the editors but we think it is a bit technical.

– I am wondering if the authors attempted to get scanning electron microscopy images of mimivirus with an exposed ball of yarn. Sounds good.

No but this is something we should consider in additional studies.

– Medusavirus genome harbored genes for all five types of histones (H1, H2A, H2B, H3, and H4). Please consider adding this to the discussion topic.

As it was never demonstrated they can form nucleosomes we prefer not to speculate.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

First, we find the uncertainty and controversy around the role and identification of the GMC oxidoreductases confusing. Please upload the cryo-EM maps and fitted atomic coordinates for the reviewers. Second, while we understand your decision to describe the genetic system and the genomic fibers of the M4 mutant in a separate publication, we feel that an additional explanation regarding the M4 mutant is necessary for this manuscript. The observation that the genomic fibers are constructed from different proteins in other mimiviruses, and potentially in the M4 mutant which lacks qu_946 and qu_143 poses quite a puzzle and raises concerns about the validity of the structural assignment.

I do not understand the Reviewers additional request as all maps and coordinates files were already provided for review purposes.

They are also available on the eLife site as Supporting Zip Documents that we thought were made available to the reviewers and editors:

  1. Cl1a : compact 5-start map and coordinates

  2. Cl1a : focus refine map and coordinates

  3. Cl1a focus refine DNA map

  4. Cl3 : relaxed 5-start genomic fiber map and coordinates

  5. Cl2 : compact 6-start map and coordinates

From the very beginning the reviewers could thus assess the quality of the reconstruction and the fit of the GMC-oxidoreductases in the various maps, including the focus refined map at 3.3 A resolution showing the FAD co-facteur into density.

Reviewers could go back to the 5 maps and coordinates of the GMC-oxidoreductase whenever they wish. They can visualize these data using a variety of software tool (coot, pymol, chimera, chimeraX…) and activate the symmetry to visualize the entire helices.

A video of the focus refined map with the fitted pdb structure was also provided as supplementary video (Video_S2.mp4) with a final zoom on the FAD co-factor showing it was in density. I wonder if the reviewers were aware of all this supplementary material.

The presence of the GMC-oxidoreductase as a main scaffolding protein in the Mimivirus genomic fiber is an experimental fact strongly supported by MS-based proteomics and by the 4 maps. Again, reviewers could assess the unambiguous fit of the protein into the 4 maps. The fact that this result is unexpected (i.e. "confusing"), and thus constitutes an important discovery, should not be used to deny its validity a priori.

As for M4, I really don't understand why it has become the focus of all attentions, as it is not the subject of our study, the purpose of which is to establish the first high resolution structure of a Mimiviridae chromosome.

Microbiology is full of examples of genes and structures once thought to be essential, then subsequently discovered lacking in related microorganisms. Considering that the GMC-oxidoreducase cannot compose Mimivirus genomic fiber simply because the homologous genes are not present in M4 is a mere opinion, not a scientific argument.

I did include in the revised version of this manuscript a sentence explicitly referring to M4:

“Interestingly, mimivirus M4 (31), a laboratory strain having lost the genes responsible for the synthesis of the two polysaccharides decorating mimivirus fibrils (11) also lacks the GMC-oxidoreductase genes. Additional studies on this specific variant will be key to establish if it exhibits a similar genomic fiber, and if yes, which proteins are composing it.”

We hope the reviewers will now have a careful look to the fit of the structure in the different maps and will get convinced on the validity of our results. I would recommend to use Coot or Pymol to be able to zoom easily into the structure and thus assess the fit of the secondary structures, the side chains as well as the one of the co-factor.

https://doi.org/10.7554/eLife.77607.sa2

Article and author information

Author details

  1. Alejandro Villalta

    Aix–Marseille University, Centre National de la Recherche Scientifique, Information Génomique & Structurale, Unité Mixte de Recherche 7256 (Institut de Microbiologie de la Méditerranée, FR3479, IM2B), Marseille, France
    Contribution
    Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – review and editing
    Contributed equally with
    Alain Schmitt and Leandro F Estrozi
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7857-7067

  2. Alain Schmitt

    Aix–Marseille University, Centre National de la Recherche Scientifique, Information Génomique & Structurale, Unité Mixte de Recherche 7256 (Institut de Microbiologie de la Méditerranée, FR3479, IM2B), Marseille, France
    Contribution
    Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – review and editing
    Contributed equally with
    Alejandro Villalta and Leandro F Estrozi
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3565-8692

  3. Leandro F Estrozi

    Univ. Grenoble Alpes, CNRS, CEA, Institut de Biologie Structurale (IBS), Grenoble, France
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – review and editing
    Contributed equally with
    Alejandro Villalta and Alain Schmitt
    Competing interests
    No competing interests declared

  4. Emmanuelle RJ Quemin

    Centre for Structural Systems Biology, Leibniz Institute for Experimental Virology (HPI), University of Hamburg, Hamburg, Germany
    Present address
    Department of Virology, Institute for Integrative Biology of the Cell (I2BC), Centre National de la Recherche Scientifique UMR9198, Gif-sur-Yvette Cedex, France
    Contribution
    Resources, Data curation, Funding acquisition, Visualization, Writing – review and editing
    Competing interests
    No competing interests declared

  5. Jean-Marie Alempic

    Aix–Marseille University, Centre National de la Recherche Scientifique, Information Génomique & Structurale, Unité Mixte de Recherche 7256 (Institut de Microbiologie de la Méditerranée, FR3479, IM2B), Marseille, France
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared

  6. Audrey Lartigue

    Aix–Marseille University, Centre National de la Recherche Scientifique, Information Génomique & Structurale, Unité Mixte de Recherche 7256 (Institut de Microbiologie de la Méditerranée, FR3479, IM2B), Marseille, France
    Contribution
    Formal analysis, Investigation, Methodology, Writing – review and editing
    Competing interests
    No competing interests declared

  7. Vojtěch Pražák

    1. Centre for Structural Systems Biology, Leibniz Institute for Experimental Virology (HPI), University of Hamburg, Hamburg, Germany
    2. Division of Structural Biology, Wellcome Centre for Human Genetics, University of Oxford, Oxford, United Kingdom
    Contribution
    Validation, Investigation, Visualization, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8149-218X

  8. Lucid Belmudes

    Univ. Grenoble Alpes, CEA, INSERM, IRIG, BGE, Grenoble, France
    Contribution
    Formal analysis, Investigation
    Competing interests
    No competing interests declared

  9. Daven Vasishtan

    1. Centre for Structural Systems Biology, Leibniz Institute for Experimental Virology (HPI), University of Hamburg, Hamburg, Germany
    2. Division of Structural Biology, Wellcome Centre for Human Genetics, University of Oxford, Oxford, United Kingdom
    Contribution
    Validation, Investigation, Visualization
    Competing interests
    No competing interests declared

  10. Agathe MG Colmant

    Aix–Marseille University, Centre National de la Recherche Scientifique, Information Génomique & Structurale, Unité Mixte de Recherche 7256 (Institut de Microbiologie de la Méditerranée, FR3479, IM2B), Marseille, France
    Contribution
    Investigation
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-2004-4073

  11. Flora A Honoré

    Aix–Marseille University, Centre National de la Recherche Scientifique, Information Génomique & Structurale, Unité Mixte de Recherche 7256 (Institut de Microbiologie de la Méditerranée, FR3479, IM2B), Marseille, France
    Contribution
    Investigation
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0390-8730

  12. Yohann Couté

    Univ. Grenoble Alpes, CEA, INSERM, IRIG, BGE, Grenoble, France
    Contribution
    Data curation, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3896-6196

  13. Kay Grünewald

    1. Centre for Structural Systems Biology, Leibniz Institute for Experimental Virology (HPI), University of Hamburg, Hamburg, Germany
    2. Division of Structural Biology, Wellcome Centre for Human Genetics, University of Oxford, Oxford, United Kingdom
    Contribution
    Resources, Supervision, Funding acquisition, Visualization, Writing – review and editing
    Competing interests
    No competing interests declared

  14. Chantal Abergel

    Aix–Marseille University, Centre National de la Recherche Scientifique, Information Génomique & Structurale, Unité Mixte de Recherche 7256 (Institut de Microbiologie de la Méditerranée, FR3479, IM2B), Marseille, France
    Contribution
    Conceptualization, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing – review and editing
    For correspondence
    Chantal.Abergel@igs.cnrs-mrs.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-1875-4049

Funding

European Research Council (832601)

  • Chantal Abergel

Agence Nationale de la Recherche (ANR-16-CE11-0033-01)

  • Chantal Abergel

Agence Nationale de la Recherche (ANR-10-INBS-08-01)

  • Yohann Couté

Agence Nationale de la Recherche (ANR-17-EURE-0003)

  • Yohann Couté

Wellcome Trust (107806/Z/15/Z)

  • Kay Grünewald

152/774-1 (INST 152/772-1)

  • Kay Grünewald

Alexander von Humboldt-Stiftung (FRA 1200789 HFST-P)

  • Emmanuelle RJ Quemin

Agence Nationale de la Recherche (ANR-10-INBS-04)

  • Chantal Abergel

The funders had no role in study design, data collection, and interpretation, or the decision to submit the work for publication. For the purpose of Open Access, the authors have applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission.

Acknowledgements

The cryo-EM work was performed at the multi-user Cryo-EM Facility at CSSB. We thank Jean-Michel Claverie for his comments on the manuscript and discussions all along the project. We thank Irina Gutsche, Ambroise Desfosses, Eric Durand, and Juan Reguera for their helpful support on structural work. We thank Carolin Seuring for support and technical help. Processing was performed on the DESY Maxwell cluster, at IBS and IGS. We thank Wolfgang Lugmayr for assistance and Sebastien Santini and Guy Schoehn for support. The preliminary electron microscopy experiments were performed on the PiCSL-FBI core facility (Nicolas Brouilly, Fabrice Richard, and Aïcha Aouane, IBDM, AMU-Marseille), member of the France-BioImaging national research infrastructure and on the IMM imaging platform (Artemis Kosta). This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement no. 832601). This work was also partially supported by the French National Research Agency ANR-16-CE11-0033-01. Proteomic experiments were partly supported by ProFI (ANR-10-INBS-08-01) and GRAL, a program from the Chemistry Biology Health (CBH) Graduate School of University Grenoble Alpes (ANR-17-EURE-0003). Cryo-EM data collection was supported by DFG grants (INST 152/772-1|152/774-1|152/775-1|152/776-1). Work in the laboratory of Kay Grünewald is funded by the Wellcome Trust (107806/Z/15/Z), the Leibniz Society, the City of Hamburg, and the BMBF (05K18BHA). Emmanuelle Quemin received support from the Alexander von Humboldt foundation (individual research fellowship no. FRA 1200789 HFST-P). Chantal Abergel received support from France-BioImaging national research infrastructure (ANR-10-INBS-04).

Senior Editor

  1. Sara L Sawyer, University of Colorado Boulder, United States

Reviewing Editor

  1. Adam Frost, University of California, San Francisco (Adjunct), United States

Reviewers

  1. Adam Frost, University of California, San Francisco (Adjunct), United States
  2. Jonatas S Abrahão, Universidade Federal de Minas Gerais, Brazil

Version history

  1. Received: February 4, 2022
  2. Preprint posted: February 18, 2022 (view preprint)
  3. Accepted: July 27, 2022
  4. Accepted Manuscript published: July 28, 2022 (version 1)
  5. Version of Record published: September 22, 2022 (version 2)
  6. Version of Record updated: September 23, 2022 (version 3)

Copyright

© 2022, Villalta, Schmitt, Estrozi et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 2,842
    Page views
  • 894
    Downloads
  • 7
    Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Alejandro Villalta
  2. Alain Schmitt
  3. Leandro F Estrozi
  4. Emmanuelle RJ Quemin
  5. Jean-Marie Alempic
  6. Audrey Lartigue
  7. Vojtěch Pražák
  8. Lucid Belmudes
  9. Daven Vasishtan
  10. Agathe MG Colmant
  11. Flora A Honoré
  12. Yohann Couté
  13. Kay Grünewald
  14. Chantal Abergel
(2022)
The giant mimivirus 1.2 Mb genome is elegantly organized into a 30-nm diameter helical protein shield
eLife 11:e77607.
https://doi.org/10.7554/eLife.77607

Categories and tags

Research organism

Further reading

    1. Microbiology and Infectious Disease

    Lys417 acts as a molecular switch that regulates the conformation of SARS-CoV-2 spike protein

    Qibin Geng, Yushun Wan ... Fang Li
    Research Article Updated

    SARS-CoV-2 spike protein plays a key role in mediating viral entry and inducing host immune responses. It can adopt either an open or closed conformation based on the position of its receptor-binding domain (RBD). It is yet unclear what causes these conformational changes or how they influence the spike’s functions. Here, we show that Lys417 in the RBD plays dual roles in the spike’s structure: it stabilizes the closed conformation of the trimeric spike by mediating inter-spike–subunit interactions; it also directly interacts with ACE2 receptor. Hence, a K417V mutation has opposing effects on the spike’s function: it opens up the spike for better ACE2 binding while weakening the RBD’s direct binding to ACE2. The net outcomes of this mutation are to allow the spike to bind ACE2 with higher probability and mediate viral entry more efficiently, but become more exposed to neutralizing antibodies. Given that residue 417 has been a viral mutational hotspot, SARS-CoV-2 may have been evolving to strike a balance between infection potency and immune evasion, contributing to its pandemic spread.

    1. Evolutionary Biology
    2. Microbiology and Infectious Disease

    Diverse evolutionary pathways challenge the use of collateral sensitivity as a strategy to suppress resistance

    Rebecca EK Mandt, Madeline R Luth ... Amanda K Lukens
    Research Article Updated

    Drug resistance remains a major obstacle to malaria control and eradication efforts, necessitating the development of novel therapeutic strategies to treat this disease. Drug combinations based on collateral sensitivity, wherein resistance to one drug causes increased sensitivity to the partner drug, have been proposed as an evolutionary strategy to suppress the emergence of resistance in pathogen populations. In this study, we explore collateral sensitivity between compounds targeting the Plasmodium dihydroorotate dehydrogenase (DHODH). We profiled the cross-resistance and collateral sensitivity phenotypes of several DHODH mutant lines to a diverse panel of DHODH inhibitors. We focus on one compound, TCMDC-125334, which was active against all mutant lines tested, including the DHODH C276Y line, which arose in selections with the clinical candidate DSM265. In six selections with TCMDC-125334, the most common mechanism of resistance to this compound was copy number variation of the dhodh locus, although we did identify one mutation, DHODH I263S, which conferred resistance to TCMDC-125334 but not DSM265. We found that selection of the DHODH C276Y mutant with TCMDC-125334 yielded additional genetic changes in the dhodh locus. These double mutant parasites exhibited decreased sensitivity to TCMDC-125334 and were highly resistant to DSM265. Finally, we tested whether collateral sensitivity could be exploited to suppress the emergence of resistance in the context of combination treatment by exposing wildtype parasites to both DSM265 and TCMDC-125334 simultaneously. This selected for parasites with a DHODH V532A mutation which were cross-resistant to both compounds and were as fit as the wildtype parent in vitro. The emergence of these cross-resistant, evolutionarily fit parasites highlights the mutational flexibility of the DHODH enzyme.

    1. Biochemistry and Chemical Biology
    2. Microbiology and Infectious Disease

    Indole produced during dysbiosis mediates host–microorganism chemical communication

    Rui-Qiu Yang, Yong-Hong Chen ... Cheng-Gang Zou
    Research Article Updated

    An imbalance of the gut microbiota, termed dysbiosis, has a substantial impact on host physiology. However, the mechanism by which host deals with gut dysbiosis to maintain fitness remains largely unknown. In Caenorhabditis elegans, Escherichia coli, which is its bacterial diet, proliferates in its intestinal lumen during aging. Here, we demonstrate that progressive intestinal proliferation of E. coli activates the transcription factor DAF-16, which is required for maintenance of longevity and organismal fitness in worms with age. DAF-16 up-regulates two lysozymes lys-7 and lys-8, thus limiting the bacterial accumulation in the gut of worms during aging. During dysbiosis, the levels of indole produced by E. coli are increased in worms. Indole is involved in the activation of DAF-16 by TRPA-1 in neurons of worms. Our finding demonstrates that indole functions as a microbial signal of gut dysbiosis to promote fitness of the host.