Structure of the two-component S-layer of the archaeon Sulfolobus acidocaldarius

  1. Lavinia Gambelli
  2. Mathew McLaren
  3. Rebecca Conners
  4. Kelly Sanders
  5. Matthew C Gaines
  6. Lewis Clark
  7. Vicki AM Gold
  8. Daniel Kattnig
  9. Mateusz Sikora
  10. Cyril Hanus
  11. Michail N Isupov
  12. Bertram Daum

  • Curated by eLife

    eLife logo

    eLife assessment

    This fundamental work substantially advances our structural understanding of S-layers in Archaea and how they are built to form formidable cell support structures able to stabilise the cytoplasmic membrane under harsh physicochemical conditions. The supporting evidence for the S-layer model is convincing, making excellent use of state-of-the-art 3D cryo-electron tomography reconstructions, although the proposed S-layer model would benefit from some additional validation.

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Surface layers (S-layers) are resilient two-dimensional protein lattices that encapsulate many bacteria and most archaea. In archaea, S-layers usually form the only structural component of the cell wall and thus act as the final frontier between the cell and its environment. Therefore, S-layers are crucial for supporting microbial life. Notwithstanding their importance, little is known about archaeal S-layers at the atomic level. Here, we combined single particle cryo electron microscopy (cryoEM), cryo electron tomography (cryoET) and Alphafold2 predictions to generate an atomic model of the two-component S-layer of Sulfolobus acidocaldarius . The outer component of this S-layer (SlaA) is a flexible, highly glycosylated, and stable protein. Together with the inner and membrane-bound component (SlaB), they assemble into a porous and interwoven lattice. We hypothesise that jackknife-like conformational changes, changes play important roles in S-layer assembly.

Article activity feed

  1. Version published to 10.7554/elife.84617 on eLife
  2. eLife

    Author Response

    Reviewer #1 (Public Review):

    Gambelli et al. provide a structural study of the SlaA/SlaB S-layer of the archaeon Sulfolobus acidocaldarius. S-layers form an essential component of most archaeal cell envelopes, where their self-assembling properties and activity as cell envelope support structures have raised substantial interest, both from researchers seeking to understand the fundamental biology of archaea, as well as researchers seeking to exploit the biomaterial properties of S-layers in biotechnological applications. Both interests are hampered by the paucity of structural information on archaeal S-layer assembly, structure, and function to date, in large part due to technical difficulties in their study.

    In this study, Gambelli and coworkers overcome these difficulties and report the high-resolution 3D cryoEM structures of the purified SlaA monomers at three different pH, as well as the medium resolution 3D cryoET structures of the SlaA/SlaB lattices determined from S-layer fragments isolated from the Sulfolobus cells.

    The structural work is generally well executed, although lacks in detail in places to allow a proper review, particularly in the cryoET. A further drawback of the current manuscript is that the structural work remains rather descriptive and speculative, with little validation of the proposed models.

    The authors run a plethora of representation, analyses, prediction, and simulation software on their structures resulting in an abundance of Figures that risk overloading the reader and in several cases bring little new insight beyond unsubstantiated speculation.

    We understand the reviewer’s concern about the number of figures presented in the manuscript. To avoid overloading the reader, we have further simplified the supplementary figures and provided additional context and explanations in the narrative of the manuscript to ensure that the reader can follow the data presented. We have also improved unclarities in legends, making sure that they provide clearer explanations of the data. Additionally, we have taken extra care to connect each figure to the main findings, emphasising how each piece of data contributes to the overall understanding of the structures.

    We find it difficult to agree with the assertion of unsubstantiated speculation. We carefully justified our interpretation of our data, referring to well-established principles and relevant literature. Nevertheless, we have attempted to provide further context and clarification in the revised manuscript. Where appropriate, we have acknowledged the limitations of our analyses and have made sure to note where further research is needed to confirm their findings.

    The structural description of the S. acidocaldarius S-layer will be of high general interest and the authors have made a substantial leap forward, but the current manuscript would benefit from a better validation and basic atomic description of the SlaA/SlaB S-layer.

    Specific points.

    • It is not possible to review the quality of the SlaA and SlaA/SlaB models in the cryoET reconstruction. No detailed fits of the map and model are shown, and no correlation statistics are given (the latter is also true for the higher resolution 3D reconstructions at pH4, 7, and 10). To be of use to the community, the S-layer model and cryoET maps should also be deposited in PDB and EMDB, and an autodep report and ideally the cryoET maps should be available.

    Maps and models for the SlaA single particle at pH4, 7 and 10 have now been released on the PDB database under the accession codes PDB-7ZCX, PDB-8AN3 and PDB-8AN2 and all validation statistics can be accessed there. We have also provided a standard cryoEM statistics table with the manuscript.

    We have also changed the main figures 4 and 5 to include more detail about the STA maps and models. We have deposited the sub-tomogram averaging map in the EMDB (EMD-18127) and models of the hexameric and trimeric pores in the Protein Databank under accession codes PDB-8QP0 and PDB-8QOX, respectively (with status release upon publication). We have also attached the map and models as supporting files to this rebuttal.

    • The authors spend a great deal on the MD simulation of the SlaA glycans and the description of the 'glycan shield' and its possible role in subunit electrostatics and intersubunit contacts. This does not result in testable hypotheses, however, and does not bring much more than vague speculation on the role of the glycans or the subunits contacts in S-layer assembly and stability.

    We propose that our glycan analysis does lead to a testable hypothesis, which could for example be tested by a future study involving the genetic or enzymatic ablation of glycosylation sites and the subsequent investigation of the structure and stability of the S-layer. We have included this statement in our manuscript to inspire future research in this direction.

    • For the primary description of the SlaA/B S-layer, more important would be a detailed atomic description and validation of the intermolecular contacts in the proposed lattice model. Given the low resolution of the cryoET, this would require MD simulation of the contacts. Lattice stability during MD simulation and/or the confirmation of lattice contacts by cross-linking mass spectrometry would go a great way in validating the proposed lattice model.

    We have improved our map and model by reprocessing our sub-tomogram averages (STA) using a different pipeline (Warp and M). We are now able to visualise more of SlaB, and the new map agrees with our Alphafold predictions of the SlaB trimer. The new map also clearly shows the interaction sites between SlaA and SlaB, as well as how SlaB integrates into the lipid bilayer. We have made new figures that now correlate the STA with the atomic model more clearly.

    Taking the reviewer’s suggestions on board, we have used Namdinator – a molecular dynamics-based flexible fitting software, to refine our model. Due to RAM limitations, we had to split our model into two pdb files. The first contains 6 SlaA monomers delineating a hexameric pore and the second, 3 SlaB monomers and 5 SlaA in the region of a trimeric pore. While the new models largely agree with the original, Namdinator did improve them. The IgG domains of SlaB now fill previously unoccupied areas of the map and any clashes have been removed. Notably, the way that SlaA is modelled is the only way in which the subunits can be reconciled with the map. This is especially true for the surface glycans, which in our model are excluded from any of the intermolecular interfaces and thus remain free to move around in the solvent. In any other SlaA configuration, there would be severe clashes between neighbouring polypeptide backbones or proteins and surface glycans and thus be sterically or entropically unfavourable.

    Unfortunately, full MD simulations of the entire S-layer array would necessitate the simulation of at least 36 SlaA monomers, including glycans, in addition to 9 SlaB monomers integrated into a membrane and solvent environment, implying >8 Million atoms. Such largescale models would only enable the simulation of very short simulation times (on the order of no more than 100 nanoseconds). Such time scales would preclude the observation of major changes, even if the model was sub-optimally configured.

    • The discussion of the subunit electrostatics and the role they could play in subunit assembly/disassembly remains superficial and speculative. No real model or hypothesis is put forward, let alone validated.

    We have rephrased the discussion to clearly state our hypothesis regarding S-layer disassembly. Hopefully, it should now be clearer that from our data, we deduce that S-layer disassembly at high pH is likely not driven by protein unfolding or pH-induced conformational change. We hypothesise that instead the pH-induced disassembly is likely caused by a weakening or abolishment of hydrogen bonds, as the proton concentration is reduced.

    • The authors solve the cryoEM structure of SlaA released and purified form S. acidocaldarius S-layers by an alkaline pH shift. When shifted back to acidic pH, does this native material self-assemble in vitro? If not, do the authors have an explanation for this? Are components missing or could the solved structures represent SlaA conformations that are no longer assembly competent?

    We have previously shown that S. acidocaldarius S-layers disassembled by a pH shift from acidic to alkaline reassemble when the pH is shifted back to acidic. We also demonstrated that this disassembly / reassembly works with both SlaB present and absent, showing that SlaA alone can assemble into an S-layer (Gambelli et al, PNAS, 2019). This means that the SlaA protein that we imaged in this manuscript is indeed reassembly competent. We have included a sentence clarifying this in the first paragraph of the Results section and have discussed our hypothesis for the mechanism underlying assembly and disassembly in detail.

    Reviewer #2 (Public Review):

    Gambelli et al. investigated the surface layer (S-layer) of Sulfolobus acidocaldarius by using combined single particle cryo-electron microscopy (cryoEM), cryo-electron tomography (cryoET), and Alphafold2 predictions to generate an atomic model of this outermost cell envelope structure. As known from previous studies, the two-dimensional lattice comprises two distinct S-layer glycoproteins (SLPs) termed SlaA, the outer component interacting with the harsh living environment of this archaeon, and SlaB, comprising a dominant hydrophobic domain, which anchors this SLP in the cytoplasmic membrane, respectively. The interwoven S-layer lattice of S. acidocaldarius shows a hexagonal lattice symmetry with a p3 topography. It is built very complex as the unit cell constitutes of one SlaB trimer and three SlaA dimers (SlaB3/3SlaA2). Despite the complexity of this distinct proteinaceous S-layer lattice, the authors not only investigated the SLP structures but also considered the glycans in their structure predictions.

    The strengths of this study are that it was possible, and the first approach taken, to divide the Y-shaped SlaA SLP, starting from the N-terminus into six domains, D1 to D6. As previous studies revealed that SlaA assembly and disassembly are pH-sensitive processes, the structure of SlaA was investigated at different pH conditions. This approach led to the striking result that the cryoEM maps of SlaA D1 to D4 are virtually identical at the three pH conditions, demonstrating remarkable pH stability of these protein domains. For SlaA at low pH, however, the domains D5 and D6 were too flexible to be resolved in the cryoEM maps. Nevertheless, the authors were able to hypothesize that jackknife-like conformational changes of a link between domains D4 and D5, as well as pH-induced alterations in the surface charge of SlaA play important roles in S-layer assembly. This study showed in addition, that the surface charges of SlaA shift significantly from positive at acidic pH to negative at basic pH. A comparison of the surface charge between glycosylated and non-glycosylated SlaA showed that the glycans contribute considerably to the negative charge of the protein at higher pH values. This change in electrostatic surface potential may therefore be a key factor in disrupting protein-protein interactions within the S-layer, causing its disassembly as it is highly desired for new practical applications in biomolecular nanotechnology and synthetic biology. An excellent approach was to use exosomes to determine the structure of the entire S-layer structure comprising of SlaA and SlaB. By this approach, effectively two zones in the SlaA assembly could be distinguished: an outer zone constituted by D1 to D4, and one inner zone formed by D5 and D6. Moreover, for the first time, deeper insights into how SlaA forms the hexagonal and triangular pores within the S-layer lattice of S. acidocaldarius are provided. Very interesting are the found SlaA dimers, which are suggested to be formed by two SlaA monomers through the D6 domains, with each SlaA dimer spanning two adjacent hexagonal pores.

    The weaknesses in this work are in the introduction, where the citation is incomplete. In the comparisons drawn between archaeal and bacterial S-layers, basic citations are missing for the latter. One gets the impression that there is a deliberate avoidance of citing individual prominent S-layer research groups here. The same is true for citations of glycosylation of archaeal S-layer proteins and Sulfolobus mutants lacking SlaB.

    We thank the reviewer for suggesting the inclusion of additional references. We would like to reassure the reviewer that we did not intend any deliberate omissions. Instead, we aimed to focus on archaeal S-layers and thus did not provide a detailed overview of bacterial S-layers. We have now incorporated more references on bacterial S-layers, hoping that this will be provide a more balanced overview.

    The authors show many pictures and schematic drawings of high quality. In the main text, these illustrations should be briefly commented on if there is any ambiguity. For example, it is somewhat difficult to understand that in one schematic drawing the angle between the SlaA longitudinal axis and the membrane plane is 28 degrees and at the same time in another schema, the angle of the longitudinal axes in SlaA dimers is given as 160 degrees.

    We thank the reviewer for their appreciation for our figures. To clarify, the angles mentioned are two different ones. The 28 degrees angle is located between the cytoplasmic membrane and the longitudinal axis of an SlaA monomer in the assembled S-layer. The 160 degrees angle is located between two SlaA monomers forming a dimer.

    The authors argue that by a pH shift to 10, SlaA disassembles and exists exclusively as a single molecule. The presence of exclusively single SlaA proteins and the purity of the fractions were assessed by SDS/PAGE analysis and cryoEM micrographs. However, one can doubt that, due to the strong denaturing effect of SDS and the subsequent dissociation of protein complexes, SlaA dimers or oligomers could have been determined with SDS/PAGE.

    To clarify, we did not assess the assembly state of the S-layer by SDS PAGE, as we are aware that assembled S-layers would not travel into the gel. Instead, we assessed the assembly state by negative stain electron microscopy. Class averages of purified SlaA did not reveal any dimers or higher oligomers.

    Moreover, the shown representative micrographs (supplementary figure 2, a-c) show a heterogeneous structure and thus, do not support the exclusive presence of disassembled SlaA monomers.

    We are not sure what exactly the reviewer is referring to, there are only single SlaA particles visible in supplementary figure 2, a-c. (new ) Larger, amorphous “blobs” in the panels are likely ethane contaminations on the cryoEM grid.

    An interesting finding is SlaA dimerization. SlaA dimers can obviously be found in co-existence with SlaA-only S-layer as shown in supplementary figure 15. A short discussion on whether dimers are an intermediate structure in the process of S-layer lattice formation from monomeric SlaA or if this structure was just a coincident observation could help the reader to better understand the meaning of these dimeric structures and at which stage they are formed.

    We thank the reviewer for their suggestion and added a brief statement to the discussion to clarify this point: “Their co-existence with assembled S-layer may indicate that SlaA dimers are an intermediate of S-layer assembly or disassembly.” The figure numbering was updated, so supplementary figure 15 has now become Figure 4-figure supplement 4.

  3. eLife

    eLife assessment

    This fundamental work substantially advances our structural understanding of S-layers in Archaea and how they are built to form formidable cell support structures able to stabilise the cytoplasmic membrane under harsh physicochemical conditions. The supporting evidence for the S-layer model is convincing, making excellent use of state-of-the-art 3D cryo-electron tomography reconstructions, although the proposed S-layer model would benefit from some additional validation.

  4. eLife

    Reviewer #1 (Public Review):

    Gambelli et al. provide a structural study of the SlaA/SlaB S-layer of the archaeon Sulfolobus acidocaldarius. S-layers form an essential component of most archaeal cell envelopes, where their self-assembling properties and activity as cell envelope support structures have raised substantial interest, both from researchers seeking to understand the fundamental biology of archaea, as well as researchers seeking to exploit the biomaterial properties of S-layers in biotechnological applications. Both interests are hampered by the paucity of structural information on archaeal S-layer assembly, structure, and function to date, in large part due to technical difficulties in their study.

    In this study, Gambelli and coworkers overcome these difficulties and report the high-resolution 3D cryoEM structures of the purified SlaA monomers at three different pH, as well as the medium resolution 3D cryoET structures of the SlaA/SlyB lattices determined from S-layer fragments isolated from the Sulfolobus cells.

    The structural work is generally well executed, although lacks in detail in places to allow a proper review, particularly in the cryoET. A further drawback of the current manuscript is that the structural work remains rather descriptive and speculative, with little validation of the proposed models.

    The authors run a plethora of representation, analyses, prediction, and simulation software on their structures resulting in an abundance of Figures that risk overloading the reader and in several cases bring little new insight beyond unsubstantiated speculation.

    The structural description of the S. acidocaldarius S-layer will be of high general interest and the authors have made a substantial leap forward, but the current manuscript would benefit from a better validation and basic atomic description of the SlaA/SlaB S-layer.

    Specific points.

    - It is not possible to review the quality of the SlaA and SlaA/SlaB models in the cryoET reconstruction. No detailed fits of the map and model are shown, and no correlation statistics are given (the latter is also true for the higher resolution 3D reconstructions at pH4, 7, and 10). To be of use to the community, the S-layer model and cryoET maps should also be deposited in PDB and EMDB, and an autodep report and ideally the cryoET maps should be available.

    - The authors spend a great deal on the MD simulation of the SlaA glycans and the description of the 'glycan shield' and its possible role in subunit electrostatics and intersubunit contacts. This does not result in testable hypotheses, however, and does not bring much more than vague speculation on the role of the glycans or the subunits contacts in S-layer assembly and stability. For the primary description of the SlaA/B S-layer, more important would be a detailed atomic description and validation of the intermolecular contacts in the proposed lattice model. Given the low resolution of the cryoET, this would require MD simulation of the contacts. Lattice stability during MD simulation and/or the confirmation of lattice contacts by cross-linking mass spectrometry would go a great way in validating the proposed lattice model.

    - The discussion of the subunit electrostatics and the role they could play in subunit assembly/disassembly remains superficial and speculative. No real model or hypothesis is put forward, let alone validated.

    - The authors solve the cryoEM structure of SlaA released and purified form S. acidocaldarius S-layers by an alkaline pH shift. When shifted back to acidic pH, does this native material self-assemble in vitro? If not, do the authors have an explanation for this? Are components missing or could the solved structures represent SlaA conformations that are no longer assembly competent?

  5. eLife

    Reviewer #2 (Public Review):

    Gambelli et al. investigated the surface layer (S-layer) of Sulfolobus acidocaldarius by using combined single particle cryo-electron microscopy (cryoEM), cryo-electron tomography (cryoET), and Alphafold2 predictions to generate an atomic model of this outermost cell envelope structure. As known from previous studies, the two-dimensional lattice comprises two distinct S-layer glycoproteins (SLPs) termed SlaA, the outer component interacting with the harsh living environment of this archaeon, and SlaB, comprising a dominant hydrophobic domain, which anchors this SLP in the cytoplasmic membrane, respectively. The interwoven S-layer lattice of S. acidocaldarius shows a hexagonal lattice symmetry with a p3 topography. It is built very complex as the unit cell constitutes of one SlaB trimer and three SlaA dimers (SlaB3/3SlaA2). Despite the complexity of this distinct proteinaceous S-layer lattice, the authors not only investigated the SLP structures but also considered the glycans in their structure predictions.

    The strengths of this study are that it was possible, and the first approach taken, to divide the Y-shaped SlaA SLP, starting from the N-terminus into six domains, D1 to D6. As previous studies revealed that SlaA assembly and disassembly are pH-sensitive processes, the structure of SlaA was investigated at different pH conditions. This approach led to the striking result that the cryoEM maps of SlaA D1 to D4 are virtually identical at the three pH conditions, demonstrating remarkable pH stability of these protein domains. For SlaA at low pH, however, the domains D5 and D6 were too flexible to be resolved in the cryoEM maps. Nevertheless, the authors were able to hypothesize that jackknife-like conformational changes of a link between domains D4 and D5, as well as pH-induced alterations in the surface charge of SlaA play important roles in S-layer assembly.
    This study showed in addition, that the surface charges of SlaA shift significantly from positive at acidic pH to negative at basic pH. A comparison of the surface charge between glycosylated and non-glycosylated SlaA showed that the glycans contribute considerably to the negative charge of the protein at higher pH values. This change in electrostatic surface potential may therefore be a key factor in disrupting protein-protein interactions within the S-layer, causing its disassembly as it is highly desired for new practical applications in biomolecular nanotechnology and synthetic biology.
    An excellent approach was to use exosomes to determine the structure of the entire S-layer structure comprising of SlaA and SlaB. By this approach, effectively two zones in the SlaA assembly could be distinguished: an outer zone constituted by D1 to D4, and one inner zone formed by D5 and D6. Moreover, for the first time, deeper insights into how SlaA forms the hexagonal and triangular pores within the S-layer lattice of S. acidocaldarius are provided. Very interesting are the found SlaA dimers, which are suggested to be formed by two SlaA monomers through the D6 domains, with each SlaA dimer spanning two adjacent hexagonal pores.

    The weaknesses in this work are in the introduction, where the citation is incomplete. In the comparisons drawn between archaeal and bacterial S-layers, basic citations are missing for the latter. One gets the impression that there is a deliberate avoidance of citing individual prominent S-layer research groups here. The same is true for citations of glycosylation of archaeal S-layer proteins and Sulfolobus mutants lacking SlaB.
    The authors show many pictures and schematic drawings of high quality. In the main text, these illustrations should be briefly commented on if there is any ambiguity. For example, it is somewhat difficult to understand that in one schematic drawing the angle between the SlaA longitudinal axis and the membrane plane is 28 degrees and at the same time in another schema, the angle of the longitudinal axes in SlaA dimers is given as 160 degrees.
    The authors argue that by a pH shift to 10, SlaA disassembles and exists exclusively as a single molecule. The presence of exclusively single SlaA proteins and the purity of the fractions were assessed by SDS/PAGE analysis and cryoEM micrographs. However, one can doubt that, due to the strong denaturing effect of SDS and the subsequent dissociation of protein complexes, SlaA dimers or oligomers could have been determined with SDS/PAGE. Moreover, the shown representative micrographs (supplementary figure 2, a-c) show a heterogeneous structure and thus, do not support the exclusive presence of disassembled SlaA monomers.
    An interesting finding is SlaA dimerization. SlaA dimers can obviously be found in co-existence with SlaA-only S-layer as shown in supplementary figure 15. A short discussion on whether dimers are an intermediate structure in the process of S-layer lattice formation from monomeric SlaA or if this structure was just a coincident observation could help the reader to better understand the meaning of these dimeric structures and at which stage they are formed.

  6. Version published to 10.1101/2022.10.07.511299v1 on bioRxiv