Cell Cryo-ET reveals native HIV virion structure

Written | Edited by Wang Ye | Xi There are still nearly 2 million people worldwide infected with human immunodeficiency virus (HIV-1) every year, but there is currently no approved vaccine [1]
.

The invasion of the enveloped virus into the host is mediated by the glycoprotein on the surface of the envelope.
For HIV-1, it is its Env protein, which is responsible for recognizing and binding to the host cell surface receptor and completing the membrane fusion process.
It is the only target of neutralizing antibodies and is also a vaccine.
Designed primary target protein
.

The Env protein is matured by the cleavage of the precursor protein pg160, and contains two subunits, gp120 and gp41, which are linked by non-covalent bonds to form a heterodimer to form an Env homotrimer, in which gp41 can pass The transmembrane domain is anchored to the viral envelope, and the C-terminal cytoplasmic domain (CTD) is located within the viral cavity
.

During HIV-1 budding, the CTD of Env interacts with the matrix domain (MA) of the immature Gag protein to form a membrane-bound lattice that is located in the inner layer of the viral bilayer membrane
.

In order to obtain a uniform and stable Env trimer structure, the researchers previously introduced a pair of disulfide bonds (SOS) between gp120 and gp41, and mutated the isoleucine on the gp41 HR1 helix to a proline (IP).
After transformation, an Env trimer called SOSIP was constructed, and a series of Env structures were analyzed based on this construction [2]
.

However, because SOSIP contains many artificial modifications, it is still unknown whether it can truly reflect the structure of the natural Env trimer
.

In order to analyze the structural difference between Env on the surface of virus particles and its extracellular soluble part, on February 4, 2022, the research group of Kelly K.
Lee of the University of Washington and Michael B.
Zwick of the Scripps Institute collaborated on Cell Published the article Cryo-ET of Env on intact HIV virions reveals structural variation and positioning on the Gag lattice, with the aid of cryo-electron tomography (cryo-ET), combining sub-tomogram averaging and structural mass spectrometry The technique analyzes the structure of HIV-1 virus-like particles with Gag in both mature and immature states, revealing the variability of the structure of HIV-1 virus particles in the native state
.

On the surface of HIV-1 virions, there are both functional but unstable Env trimers and non-native Env proteins
.

To overcome the heterogeneity and instability of Env proteins, Michael B.
Zwick's group has previously obtained a series of Env protein variants through direct evolution [3].
The ADA.
CM Env used in this paper is derived from the type B isolate ADA.
A variant with 102 amino acids at the C-terminal end of its C-terminal domain (CTD) truncated
.

The study used virus-like particles called hVLPs as study samples, which have a double-layered envelope assembled outside the Gag protein layer, and the maturation of Gag requires enzymatic cleavage
.

hVLP contains the same components as real HIV-1 particles but does not have the ability to replicate, because the encapsulated viral genome lacks the Env gene, so the complete viral particles must be packaged in HEK 293 cells with ADA.
CM Env expressed on the surface [4] ]
.

Purified hVLPs are also treated with a mild oxidant, dithiodipyridine-2 (AT-2), to disrupt the nucleocapsid-RNA interaction and ensure that the particles are no longer infectious
.

Antibody binding experiments showed that AT-2 treatment did not affect the antigenicity of Env protein; although it was not infectious, hVLP treatment with AT-2 treatment could still induce syncytial and cytotoxic effects in infected cells, indicating that Env still had the ability to mediate virus ability to invade
.

Using this hVLP with a high surface Env protein density, the authors collected sufficient cryo-ET data and finally processed the HIV-1 virion structure with a resolution of 9.
1 Å
.

In the obtained structural data, most of the hVLPs were in the mature state, and about 3% of the particles were in the immature state, showing a clear and regular Gag grid under the capsule
.

Processing this fraction of sub-data yields a lower-resolution density map, but its Env structure is very close to the sub-nanometer-resolution Env structure obtained from the full data and the previously published Env ectodomain structure
.

Published structures of the Env ectodomain and hexameric Gag-CA could be matched to the corresponding densities, but the densities of the Env TMD-CTD did not appear in the averaged structures, indicating that the transmembrane domain (TMD) was not associated with the Env Extracellular segment or Gag-CA tight junction
.

Based on the presumed relative positions of the TMD-CTD to the Env extracellular segment, the former structure can be matched to the density map
.

The MA domain of the immature Gag N-terminal assembles into a hexamer in the form of a homotrimer along the inner layer of the viral envelope, and the MA is connected to the Gag-CA grid layer through a flexible connecting region
.

On the tomogram of immature hVLP, the regularly arranged dense points corresponding to Gag-MA can be clearly identified, which are located under the inner membrane layer; but the density of the Gag-MA layer cannot be distinguished in the averaged density map.
, the authors speculate that the arrangement of Gag-MA grids may be slightly different from the symmetry axes of Gag-CA or Env
.

Therefore, the structure of the MA grid can only be matched into the density map based on the presumed position of Gag-MA relative to Gag-CA (Fig.
1)
.

So the Env CTD and Gag-MA are discharged side by side at the bottom of the inner membrane layer, which just coincides with the rough tomogram
.

Moreover, in this composite model, the relative position of the Env CTD is just at the secondary axis of the edge of the Gag-CA grid, which rules out the assumption that it is located at the central hole in previous studies [5]
.

In this conformation, the Env CTD can directly interact with key residues on MA that are thought to be required for Env packaging
.

To further confirm the results of the structural analysis, the authors also collected cryo-electron tomography data of immature hVLPs with full-length trimers of ADA.
CM.
The analysis showed that the organization of Gag polyproteins was similar to that in the previous model, and the Env trimers were similar.
Also at the edge of the Gag-CA grille rather than the center hole
.

Figure 1 Composite model of the immature HIV-1 virion Env-Gag.
In the mature virion structure, the Gag polyprotein has been cleaved by enzymes, and the MA layer is fragmented instead of the continuous state in the aforementioned immature granules, indicating that MA structure needs to be remodeled during virion maturation
.

Moreover, the Env trimer is no longer co-located with the MA layer
.

The authors speculate that disruption of the membrane-bound Gag grid may help Env trimers acquire sufficient mobility to mediate the membrane fusion process
.

Next, we focus on analyzing the structure of the Env trimer
.

The authors found that hVLP-Env is essentially in a closed prefusion conformation with its extracellular segment attached to the viral membrane by a thin tripod-like stem-loop (Fig.
2)
.

Similar to the Env structure in immature hVLPs, the density of the Env transmembrane region in this subtomogram-averaged structure is also not visible, although its presence can be confirmed in gross tomographic images
.

Different numbers of Env were distributed on each mature hVLP, ranging from 8 to 82, and these trimers appeared to be randomly distributed, with no apparent aggregation or tight interaction with other trimers
.

hVLPs are basically spherical in shape, with an average diameter of about 100 nm.
Therefore, it is roughly estimated that a hVLP can accommodate up to 429 Env trimers.
Therefore, the number of Env trimers present on hVLP is much lower than its theoretical upper limit.
This allows sufficient surface area for each Env trimer to move
.

In addition, the authors also used hydrogen deuterium exchange mass spectrometry (HDX-MS) to analyze Env on native hVLPs, confirming that it is in a closed prefusion conformation, and its conformation is heterogeneous; meanwhile, by comparing the soluble BG505.
SOSIP construct, The overall conformation of the two was found to be similar, especially the organization of the gp120 subunit, but hVLP-Env was more regular in structure
.

Figure 2.
Averaged Env density map and matching structural model.
Side view.
What about the gp41 subunit? This is the biggest difference between hVLP and existing structures
.

In the previously published structure, the HR2 helix of Env is almost all in a long rod-like conformation, and the distal end is formed by Asp664.
In the hVLP-Env obtained in this paper, the length of the HR2 helix is ​​reduced by 10 amino acids, and the distal end is formed by Gln653
.

The remaining amino acids at the C-terminus of HR2 are bent to form a thin stem-loop connecting the extracellular segment with the viral membrane, and the tripod-like stem-loop enables the Env extracellular segment to be lifted about 10 Å away from the viral membrane
.

This also makes most of the membrane proximal outer region (MPER) embedded in the membrane, which is considered an ideal vaccine target for eliciting some of the most broadly neutralizing antibodies [6]
.

In addition, the authors observed that hVLP-Env tilted towards the membrane, resulting in a decrease in the density of the stem-loop region in the sub-fault average density map
.

The authors speculate that this tilt may help the virus resist neutralizing antibodies targeting MPER, and the analysis of neutralization experiments also verified this conjecture
.

Another major structural difference of the hVLP-Env gp41 subunit is that the complete density of the centrally located HR1 C-terminal helix (570-595) is not visible in the homogenized structure
.

The authors ruled out the possible effects of resolution and the introduction of symmetry, and HDX-MS analysis showed that some segments of this region had stronger main-chain hydrogen bonding interactions, which indicated that there was also structural variability in the central region compared to previously reported structures
.

Finally, the authors focus on glycosylation modifications
.

Sugar modifications at the glycosylation site in previously reported structures often only see the core N-acetylglucosamine, but in the hVLP-Env structure solved here, the extended glycosylation density is visible throughout the protein In particular, sugar modifications with multiple canonical sites resolved on the gp120 and gp41 subunits range in densities beyond existing structures, revealing a so-called "glycan shield"
.

The authors analyzed the sugar modifications near the receptor CD4 binding region, several important neutralizing antibody epitopes, and the region adjacent to the fusion peptide, showing that sugar side chains can protect the corresponding sites through steric hindrance
.

However, V3-loop and V1-/V2-loop antibodies bind to sugar modifications at supersites that produce no or only weak steric hindrance, and antibodies targeting these sites often need to bind adjacent sugar side chains to achieve To achieve its function, the authors speculate that this may also be one of the reasons why these sites can be effectively targeted by antibodies
.

In addition, the authors also found differences in the density of certain sugar modifications on each monomer of the Env trimer, showing heterogeneity of sugar modifications
.

Reviewing the full text, the authors analyzed the structure of HIV-1 virus particles in the native state by cryo-electron tomography: On immature particles, the orientation of Env on the bottom Gag grid is diverse, and the capture of this state provides a solution for Env packaging.
A few clues were drawn; sub-tomogram-averaged reconstructed virion structure combined with structural mass spectrometry revealed structural diversity in the center composed of gp41 subunits, heterogeneous glycosylation modifications on different Env monomers, and flexible stems The loops allow for tilting of the Env protein and different exposure patterns of neutralizing epitopes
.

The acquisition of the above results also provides new information on HIV-1 assembly and the structural diversity presented by the Env epitope
.

Original link: https://doi.
org/10.
1016/j.
cell.
2022.
01.
013 Publisher: Eleven References 1, Pandey, A.
, and Galvani, AP (2019).
The global burden of HIV and prospects for control.
Lancet HIV 6, e809–e811.
2, Ward, AB, and Wilson, IA (2015).
Insights into the trimeric HIV-1 envelope glycoprotein structure.
Trends Biochem.
Sci.
40, 101–107.
3, Leaman, DP, and Zwick, MB (2013).
Increased functional stability and homogeneity of viral envelope spikes through directed evolution.
PLoS Pathog.
9, e1003184.
4, Stano, A.
, Leaman, DP, Kim, AS, Zhang, L.
, Autin, L.
, Ingale, J.
, Gift, SK, Truong, J.
, Wyatt, RT, Olson, AJ, and Zwick, MB (2017).
Dense array of spikes on HIV-1 virion particles.
J.
Virol.
91, e00415-17.
5, Tedbury, PR, Novikova, M.
, Ablan, SD, and Freed, EO (2016).
Biochemical evidence of a role for matrix trimerization in HIV-1 envelope glycoprotein incorporation.
Proc.
Natl.
Acad.
Sci.
USA 113, E182–E190.
6, Rantalainen, K.
, Berndsen, ZT, Antanasijevic, A.
, Schiffner, T .
, Zhang, X.
, Lee, WH, Torres, JL, Zhang, L.
, Irimia, A.
, Copps, J.
, et al.
(2020).
HIV-1 Envelope and MPER Antibody Structures in Lipid Assemblies.
Cell Rep.
31, 107583.
Instructions for reprinting [Original article] BioArt original articles are welcome to forward and share, and reprinting is prohibited without permission.
The copyright of all works published is owned by BioArt
.

BioArt reserves all legal rights and violators will be held accountable
.