Abstract
Human T cell leukemia virus type 1 (HTLV-1) immature particles differ in morphology from other retroviruses, suggesting a distinct way of assembly. Here we report the results of cryo-electron tomography studies of HTLV-1 virus-like particles assembled in vitro, as well as derived from cells. This work shows that HTLV-1 uses a distinct mechanism of Gag–Gag interactions to form the immature viral lattice. Analysis of high-resolution structural information from immature capsid (CA) tubular arrays reveals that the primary stabilizing component in HTLV-1 is the N-terminal domain of CA. Mutagenesis analysis supports this observation. This distinguishes HTLV-1 from other retroviruses, in which the stabilization is provided primarily by the C-terminal domain of CA. These results provide structural details of the quaternary arrangement of Gag for an immature deltaretrovirus and this helps explain why HTLV-1 particles are morphologically distinct.
Similar content being viewed by others
Main
The Retroviridae family includes two important human pathogens infecting T cells, human immunodeficiency virus type 1 (HIV-1) and human T cell leukemia virus type 1 (HTLV-1). The global prevalence suggests that the number of people living with HTLV-1 (a member of the Deltaretrovirus genus) ranges from 5 to 10 million, which is likely an underestimate1. While most HTLV-1 infections remain asymptomatic, approximately 5% lead to aggressive diseases such as HTLV-1-associated myelopathy and tropical spastic paraparesis (HAM-TSP) or adult T cell leukemia and lymphoma (ATLL). ATLL is an aggressive form of cancer with a median survival rate of less than 1 year2,3.
The retroviral structural protein Gag forms the immature protein shell of nascent virus particles4. All retroviral Gag proteins contain three canonical domains (Extended Data Fig. 1a): matrix (MA), capsid (CA), consisting of independently folded N-terminal (CA-NTD) and C-terminal (CA-CTD) domains, and nucleocapsid (NC). During immature virus particle formation, these domains function in membrane binding, viral lattice self-assembly and viral genomic RNA (vgRNA) packaging, respectively. Gag oligomerization is primarily driven by interactions between CA domains and these interactions determine virus particle morphology and size5. The immature Gag lattice has local six-fold symmetry and is incomplete; for example, in HIV-1, it covers ~60% of the available membrane surface area inside a virion. Upon maturation, the viral-encoded protease cleaves Gag at defined positions, leading to a cascade of structural changes that rearrange the virion interior6. In a mature virion, MA remains associated with the viral membrane, while the CA protein forms a core consisting of CA hexamers and pentamers. The CA core contains the condensed NC–vgRNA complex, reverse transcriptase and integrase. Expression of Gag in mammalian cells, as well as the bacterial expression and purification of certain truncated variants thereof, is sufficient for assembly of virus-like particles (VLPs) in vitro7 that have authentic immature Gag or mature CA architectures8,9,10,11.
Despite substantial sequence variation among retrovirus species, CA shows a strongly conserved tertiary fold, with six to seven α-helices in the CA-NTD and four α-helices in the CA-CTD12,13,14,15,16,17,18. The latter harbors the highly conserved major homology region (MHR), which has been implicated in preserving the CA-CTD protein fold and in establishing relevant interactions in the assembly of retroviral lattices9,19. HIV-1 (Lentivirus) and Rous sarcoma virus (RSV; Alpharetrovirus) have a Gag cleavage product between CA and NC named the spacer peptide (SP) and Mason-Pfizer monkey virus (M-PMV; Betaretrovirus) has a similar segment in CA named the spacer peptide-like region20,21. The SP region of Gag is important for immature assembly and in the regulation of maturation4. Similarly, Gag and CA lattice formation in immature and mature virus particles, respectively, requires CA-CTD dimerization, which is established by hydrophobic residues in helices 8 or 9 of the CA-CTD22.
The CA-CTD has been proposed to have a dominant role for immature assembly of retroviruses. Variants of HIV-1 Gag containing only the C-terminal half of CA and downstream regions can form VLPs when expressed in cells23,24. Cryo-electron tomography (cryo-ET) and subtomogram averaging of purified immature virus particles or in vitro assembled immature-like VLPs of lentiviruses, alpharetroviruses, betaretroviruses and gammaretroviruses demonstrate that CA-CTD dimerization is a key factor in the conserved structural arrangement of the CA-CTD hexamer of immature Gag lattices8,18,25,26. These studies also show that the immature CA-NTD arrangement differs substantially among retrovirus genera. In summary, these structural studies are consistent with the hypothesis that the CA-CTD is responsible for forming the essential stabilizing immature assembly interfaces in retroviruses, while the CA-NTD may have a primary role in determining immature lattice curvature and particle size and represents a major binding site for host cell factors27.
It is not certain whether the function of the CA-NTD in assembly is conserved for retroviruses and whether there are other CA-NTD arrangements that have not been reported. Interestingly, previous mutagenesis experiments targeting residues in the CA-CTD of HTLV-1 Gag did not affect VLP budding28, suggesting that, unlike other retroviruses, the CA-CTD is not the key determinant of immature assembly. Morphological studies of HTLV-1 particles using cryo-electron microscopy (cryo-EM) also showed peculiar differences in Gag lattice shape, such as areas of flat lattices and varying membrane–CA distance29,30,31. Taken together these studies suggest that HTLV-1 may use different Gag assembly mechanisms than other retroviruses.
However, understanding Gag assembly mechanisms in HTLV-1 is limited by the current lack of structural information for this virus. Given that HTLV-1 is a human pathogen of importance and the ongoing development of inhibitors targeting HIV-1 CA (CA assembly inhibitors)32, it is critical to generate this structural information.
Here, we report cryo-ET and subtomogram averaging results of full-length and truncated HTLV-1 Gag assemblies. Together with supporting mutagenesis experiments, this allows us to describe the key immature HTLV-1 assembly determinants, delineating the contributions of the CA-NTD and CA-CTD to assembly. Our results show that HTLV-1 uses a distinct CA-NTD arrangement to form the immature Gag lattice and that, in contrast to other retroviral structures, the CA-NTD is the key determinant of immature HTLV-1 assembly.
Results
HTLV-1 Gag VLPs display unusual variability
To gain insight into the immature architecture of HTLV-1, we used cryo-ET to visualize mammalian cell-derived Gag-based VLPs (Table 1, Fig. 1a and Extended Data Fig. 1c). As previously reported30,31, the shape and size of the VLPs varied substantially. The majority of VLPs were spherical; however, the local curvature ranged from flat patches to sharp kinks (Fig. 1a and Supplementary Video 1). As with other retroviruses, the individual domains of Gag formed distinct layers of density starting at the inner leaflet of the viral membrane (corresponding to MA) and extending toward the center of the VLP by ~25 nm (Fig. 1b,c). The CA layer exhibited a regular organization; hence, we used subtomogram averaging (Extended Data Fig. 2a) to obtain higher-resolution information for this layer. Alignments focused on the CA layer resulted in a blurred density for the membrane, which is consistent with the measured variation in the CA-to-membrane distance (mean distance of 17.7 nm, s.d. of ±0.8 nm) (Extended Data Fig. 2b,c). The distance of flat Gag lattice regions to the membrane was previously reported to be greater than the distance of curved regions30.
a, Computational slice (thickness, 8.8 nm) through a cryo-electron tomogram containing HTLV-1 Gag-based VLPs. Protein density is black. Scale bar, 50 nm. The shown tomogram is representative of the 85 tomograms acquired. b, Enlarged view of the Gag lattice within VLPs, as annotated by a white rectangle in a. The arrowheads designate the different layers of the radially aligned Gag lattice underneath the viral membrane (VM): NC-RNP (1), CA-CTD (2), CA-NTD (3), MA (4), inner leaflet (5) and outer leaflet (6). c, The 1D radial density plot of the Gag lattice in immature HTLV-1 Gag-based VLPs. Two zero-value reference points are reported for the distance measurement. The primary x axis reference zero value is set at the VM (top x axis). The secondary x axis reference zero value is set at the local density minimum between the CA-NTD and CA-CTD (bottom x axis) and is provided to allow a straightforward comparison to Fig. 2c. The distance of the different layers from these reference points is given in nanometers. The annotation with arrowheads is as in b. AU, arbitrary units. d–f, Isosurface representations of the subtomogram average of the CA hexamer from HTLV-1 Gag-based VLPs. The CA-NTDs of the central hexamer are colored cyan, with one monomer highlighted in blue. Two additional CA-NTD monomers from adjacent hexamers are also colored in cyan, to highlight the trimeric interhexamer interface. The CA-CTDs of the central hexamer are colored orange and red. The hexameric arrangement is indicated by a small hexagon. d, CA lattice as seen from the outside of the VLP and rotated by 90° to show a side view. e, Top view of the CA-NTD, with the trimeric interhexamer interface and the intrahexameric interface indicated with a dashed red triangle and dashed yellow hexagon, respectively. CA monomers in the trimeric interhexamer interface are annotated with I, II and III. f, Top view of the CA-CTD hexamer. A dashed orange ellipsoid highlights one CA-CTD dimer linking adjacent hexamers. g,h, Molecular models of the CA-NTD and CA-CTD rigid-body fitted into the EM density of the immature CA lattice. Coloring as in d–f. g, Trimeric CA-NTD interactions linking hexamers, involving residues spanning helices 4 and 5. h, Interactions around the hexamer, involving helices 1, 2 and 3 from adjacent CA-NTDs. i, Model of the CA-CTD dimer.
Retroviral Gag lattices are spherical and the CA layer follows local C6 symmetry8,18,33. However, to account for the observed heterogeneity of HTLV-1 Gag VLP shapes, we applied local C2 symmetry during subtomogram averaging. The resulting maps had a CA-CTD layer that was resolved to a lower resolution than the CA-NTD layer, which we predict is because of heterogeneity caused by flexibility of the CA-CTD. Three-dimensional (3D) classification focused on the CA-CTD yielded a class in which all helices were resolved. In contrast, the CA-NTD did not show increased heterogeneity. Therefore, we used both domains as independent species in a multiparticle refinement using the software M34. This approach yielded 5.9-Å and 6.2-Å resolution maps of the CA-NTD and CA-CTD layers, respectively (Fig. 1d–f and Extended Data Fig. 2d,e). This allowed for the generation of a model of the immature HTLV-1 CA lattice by rigid-body fitting models of HTLV-1 CA-NTD and CA-CTD into the EM density map (Extended Data Fig. 2f). Densities for MA and NC above and below CA, respectively, are present in the one-dimensional (1D) density plot (Fig. 1c) but MA and NC are not resolved in the subtomogram averages, suggesting that neither domain follows the same organization as the CA lattice.
Two CA-NTD interfaces stabilize the immature lattice
The CA-NTD establishes lateral interactions that form two basic building blocks shaping the lattice: a CA-NTD trimer and a CA-NTD hexamer (Fig. 1e,g,h and Supplementary Video 2). For both the trimer and the hexamer, each CA-NTD monomer contacts two adjacent monomers around the local pseudo-symmetry axes. The trimer is formed by interactions spanning residues in helices 4 and 5 (Fig. 1e,g). Similarly, the hexamerization interface involves helices 1, 2 and 3 of two neighboring monomers (annotated I and I′ in Fig. 1e,h). Unlike in some other retroviruses8,18,25, there is no dimerization interface in the HTLV-1 CA-NTD.
CA-CTD forms isolated dimers but not a continuous lattice
Within the CA-CTD layer, the only intermolecular contact is a dimerization interface formed by helix 8 (Fig. 1f,i and Supplementary Video 2). This is similar to other immature retroviral Gag assemblies, where contacts between hexamers are established by a dimeric interface involving two neighboring CA-CTD monomers (Extended Data Fig. 3a). Unlike other retroviruses, where the CA-CTD forms interhexameric and intrahexameric contacts, the CA-CTD of HTLV-1 only forms interhexameric contacts. The positioning of the individual HTLV-1 CA-CTD dimers places them out of reach of the adjacent dimers around the hexameric CA ring (Extended Data Fig. 3b), not enabling them to contribute to intrahexamerization interfaces forming a lattice.
HTLV-1 lacks a six-helix bundle (6HB), which is reported to be critical for immature CA hexamer formation for other retroviruses. For HIV-1, the 6HB is formed by residues in CA and SP1 (refs. 9,35). Similarly, in RSV, M-PMV and murine leukemia virus (MLV) the immature lattices are stabilized by interactions downstream of the CA layer8,18,25 (Extended Data Fig. 3b). HTLV-1 Gag contains only the canonical Gag domains and is not reported to have a spacer peptide between CA and NC. This is consistent with our observation that there are no ordered densities C-terminal to CA, which would be indicative of the presence of a 6HB or another stabilization-providing region. Hence, the absence of an assembly element such as a 6HB may explain why HTLV-1 has a less well-organized CA-CTD layer.
The HTLV-1 CA-CTD effectively acts to help stabilize the CA-NTD interhexameric interactions. The positions of individual dimers are constrained only by the flexible linker between CA-NTD and CA-CTD. The role of CA-CTD seems to be in further crosslinking the NTD layer, while not constraining the shape of the lattice to a single curvature. Instead, the variable distance between the CA-CTD dimers allows for heterogeneity in particle size and shape.
High-resolution model of the immature HTLV-1 CA-NTD lattice
To analyze the structure of immature HTLV-1 lattice interfaces in more detail, we engineered an HTLV-1 Gag truncation construct for bacterial expression, purification, in vitro assembly and analysis by cryo-ET. Our construct lacked the N-terminal 125 residues of Gag (MA126CANC) (Extended Data Fig. 1a and Extended Data Fig. 4a), similar to other Gag truncation variants that have been previously used to study immature HIV-1, equine infectious anemia virus (EIAV) and M-PMV assemblies9,20,26.
The purified HTLV-1 MA126CANC assembled into hollow tubes (Fig. 2a and Supplementary Video 3), similar to immature-like tubes observed for EIAV, M-PMV and HIV-1 (refs. 20,26,36). Cryo-ET of the HTLV-1 tubular assemblies revealed a density profile of the CANC layer matching the profile observed for HTLV-1 Gag-based VLPs (compare Figs. 1c and 2c). The MA126CANC tubes exhibited varying helical parameters. Tubes were sorted into nine groups and each group was processed separately by subtomogram averaging (Extended Data Fig. 5). The two groups with the largest number of tubes had similar geometry and, thus, they were pooled and processed together. The resulting final reconstruction of the CA-NTD layer had an overall estimated resolution of 3.4 Å (Extended Data Fig. 6a,b). The helical pitch and densities for large side chains were clearly visible at this resolution, which enabled us to build a refined model of the immature CA-NTD lattice interactions (Extended Data Fig. 6c and Table 1). The in vitro assembled tubes (Fig. 2d) had the same arrangement of the CA-NTD layer as the full-length HTLV-1 Gag VLPs (Cα root-mean-square deviation (r.m.s.d.) = 1.2 Å), underscoring the biological relevance of our in vitro system. Hence, our refined model allowed identification of the side chains likely to participate in the CA-NTD interfaces that stabilize the immature lattice.
a, Computational slice (thickness, 8.8 nm) through a cryo-electron tomogram containing HTLV-1 MA126CANC tubes. Protein density is black. Scale bar, 50 nm. The tomogram is representative of the 69 tomograms acquired. b, Enlarged view of the CANC lattice within tubes, as annotated by a white rectangle in a. The arrowheads designate the different layers of the tubular lattice: NC-RNP (1), CA-CTD (2) and CA-NTD (3). c, The 1D radial density plot of the CANC lattice in MA126CANC tubes, measuring the distance of the individual CA domains and NC-RNP from the linker between the CA-NTD and CA-CTD. d, EM density map of the immature HTLV-1 CA-NTD hexamer at 3.4-Å resolution, seen from the outside of the tube (left) and a side view (right). The three symmetry-independent trimer positions are colored in cyan, blue and pink. The remaining three CA-NTD domains of the central hexamer are colored in dark gray. Note the missing density for the CA-CTD in the side view on the right. e, Zoomed-in view into the CA-NTD intrahexamer interface (annotated with a yellow dashed hexagon in d). Labeling is as in Fig. 1. Assembly-relevant residues and the corresponding helices are annotated. f, Left: zoomed-in view into the trimeric CA-NTD interface (annotated with a red triangle in d). Residues within the trimeric interface that were analyzed using mutagenesis experiments are shown and colored according to the indicated coloring scheme. Right: for simplicity, one CA-NTD domain with the highlighted and colored residues is shown in side view to allow an easier appreciation of the residue location within the CA-NTD. The C2 hexamer center of the isosurface view in d is annotated by a schematic hexamer and the trimeric interfaces in d,f are annotated by black triangles.
Residues involved in stabilizing the immature CA-NTD lattice
Previous cell-culture-based mutagenesis studies of CA revealed the critical role of the CA-NTD for particle morphogenesis, while CA-CTD amino acid substitutions did not affect budding. Specifically, CA-NTD residues M17 and Y61 have been identified as essential for immature HTLV-1 Gag lattice assembly and particle formation37. In our model, these residues are located in helices 1 and 3, respectively, which form the hexamerization interface together with helix 2 (Fig. 2e). To further characterize this interface, we substituted residue L43, which interacts with M17 and Y61 in our model. The L43A substitution reduced particle release by more than 50% and further morphological characterization of released L43A particles showed aberrant and scarce patches of Gag lattice underneath the viral membrane (Extended Data Fig. 7a,b and Supplementary Table 1).
Double-alanine swap experiments on a set of adjacent residues were performed to characterize the trimeric CA-NTD interface. Residue pairs H72-Q73, Y90-N91, P92-L93, G95-P96, L97-R98 and V99-Q100 were substituted (Fig. 2f) and screened for altered particle production efficiency. Previously, the G95-P96 pair was already shown in an insertion–substitution approach to reduce particle formation28, supporting the importance of this region in VLP assembly. Our mutagenesis experiments revealed a substantial role for the H72-Q73, P92-L93 and V99-Q100 residue pairs in immature Gag assembly and particle release (Extended Data Fig. 7c and Supplementary Table 1). The Y90-N91, G95-P96 and L97-R98 residue pair substitutions not only reduced the particle production but also greatly impacted cellular Gag levels, suggesting these substitutions to also have an effect on expression or stability of the Gag protein. Overall, we conclude from these observations that residues within the trimer interface have important roles for both the behavior of the individual Gag molecules and Gag lattice assembly and stability.
The CA-CTD layer can adapt to variable curvatures
The CA-CTD layer in the tubular arrays was arranged in a similar way as in the VLPs produced from cells (Fig. 1f and Extended Data Fig. 6d), also forming a layer of disconnected dimers. However, in this case, the qualitative difference between CA-NTD and CA-CTD was even more dramatic. The CA-CTD density in different individual tubes did not permit structural analysis other than coarse positioning of the CA-CTD dimer, as the CA-CTD layer was only visible at lower resolution (compare Fig. 2d and Extended Data Fig. 6d). Nevertheless, this allowed us to assess large-scale changes in the CA layer between the different tube geometries. Unlike the CA-NTD layer, the CA-CTD layer was observed to have larger movement of its domains. While the distance between the adjacent symmetry-independent CA-NTD trimers was nearly constant (42.1–42.6 Å), the distance between symmetry-independent CA-CTD dimers was skewed such that it was decreased in the direction of tube curvature (31.8 Å) and increased along the tube axis (35.6 Å) (Extended Data Fig. 6e). Whether there are different conformations of CA-CTD dimers and to what extent these might contribute to the flexibility or fluidity of the layer could not be discerned from our data.
Biophysical characterization of HTLV-1 CA and its domains
Given the peculiar behavior of the CA-CTD within the immature lattice, we sought to further characterize full-length CA and its two domains separately using biophysical approaches. To this end, HTLV-1 CA (residues 131–344), CA-NTD (residues 131–258) and CA-CTD (residues 259–344) (Extended Data Fig. 4a) were expressed in Escherichia coli and purified. The CA, CA-NTD and CA-CTD proteins were subjected to nano differential scanning fluorimetry (nanoDSF), which monitors changes in intrinsic tryptophan fluorescence (ITF) upon protein unfolding as a function of temperature38 (Extended Data Fig. 8; see also highlighted tryptophan residues within HTLV-1 CA in Extended Data Fig. 1b). We also conducted backreflection and dynamic light scattering (DLS) experiments to measure turbidity and determine cumulant radius as a means of informing about aggregation and assembly properties. We aimed to characterize the properties of CA and its domains in both a soluble form and a potentially assembled state. To this end, we carried out the measurements at two different time points and protein concentrations: (1) a preassembly condition, directly after transferring the protein to assembly buffer at a concentration normalized for tryptophan content, and (2) a postassembly condition, after incubation under assembly conditions. We note that we were able to obtain reproducible assemblies only for HTLV-1 CA (Extended Data Fig. 4b) but not its domains.
These data revealed the overall similar behavior of full-length CA and CA-NTD, while CA-CTD behaved differently. CA-CTD reproducibly showed the highest melting temperature (Tm) in both conditions, suggesting that CA-CTD is the most stable of the three protein variants we tested (Extended Data Fig. 8a and Supplementary Table 2). Both full-length CA and CA-NTD showed a higher propensity to either aggregate or potentially form regular structures (Extended Data Fig. 8b,c), while CA-CTD did not show a substantial increase in turbidity or cumulant radius in any of the conditions tested.
For comparison of our HTLV-1 measurements to another retrovirus species, we repeated the same set of experiments with purified HIV-1 CA (residues 133–363), CA-NTD (133–278) and CA-CTD (279–363) (Extended Data Fig. 4c, Extended Data Fig. 8d–f and Supplementary Table 2). Interestingly, HIV-1 CA and its domains behaved almost identically to their HTLV-1 counterparts in the different tested conditions, showing similar trends for melting curves, turbidity and cumulant radius. The minor differences observed were in slight changes in the respective inflection point temperatures and the signal strength for fluorescence ratio, most likely caused by the number of tryptophan residues and their location within the individual CA constructs.
For the tested conditions, we were only able to reproducibly obtain assemblies that could be visualized by negative staining TEM for HIV-1 CA but not the CA-NTD or CA-CTD alone (Extended Data Fig. 4d).
Discussion
Noncanonical stabilization of immature HTLV-1 by the CA-NTD
The work presented here expands our knowledge of the immature virion Gag architecture to five of the six genera within the subfamily Orthoretrovirinae (Fig. 3). Importantly, we identified an unconventional mechanism of immature Gag lattice stabilization that is, to date, unique to HTLV-1. Our structural experiments are consistent with a mode in which the CA-NTD is the only lattice-forming CA domain in immature HTLV-1 particles. This is supported by a previous study that did not find substitutions in the CA-CTD affecting particle budding28 and our study showing that, in our in vitro MA126CANC assemblies, the CA-CTD is less organized than the CA-NTD (or the CA-CTD of other retroviruses).
Schematic representation of the CA-NTD and CA-CTD arrangements within immature retroviruses from different genera. HTLV-1, Deltaretrovirus (as determined in this study); HIV-1, Lentivirus; RSV, Alpharetrovirus; M-PMV, Betaretrovirus; MLV, Gammaretrovirus.
The structural interactions that differentiate retroviral genera can be defined as contributions of the individual interfaces and the orientation of the CA-NTD with respect to the local symmetry axes. The CA-NTD arrangement in HTLV-1 (Deltaretrovirus) is similar to that of MLV (Gammaretrovirus) (Fig. 3), with both having a tight packing of the trimerization interface and helix 1 pointing toward the center of the hexamer. This results in positioning of CA-NTD in HTLV-1 and MLV closer to that observed in mature CA lattice conformations18. When using gag and pol sequences for establishing evolutionary relationships, Lentivirus and Betaretrovirus genera are closer to Deltaretrovirus than Gammaretrovirus39. We performed a multiple-sequence alignment (MSA) for different retroviral CA sequences and generated a phylogenetic tree to evaluate how these evolutionary relationships change when we take into account only CA40,41 (Extended Data Fig. 9). Indeed, we found the CA relationship for gammaretroviruses and deltaretroviruses to be slightly closer to each other than to lentiviruses, which is the other way compared to full-length Gag comparisons. We interpret this CA phylogeny analysis to be in line with our observation that the quaternary immature CA arrangement in HTLV-1 and MLV is more alike.
CA domain properties do not solely define assembly behavior
Our biophysical experiments of HTLV-1 and HIV-1 CA and their respective domains indicate that the different assembly properties of HTLV-1 and HIV-1 are not derived solely from features within the individual CA-NTD domains but must also be influenced by a defined interplay between the CA domains. In addition, it is likely that their context within Gag and the interaction with neighboring domains, such as NC and MA, have an important role. NanoDSF does not necessarily consider immature or mature CA behavior but provides insight into protein stability and its propensity to interact or aggregate. Hence, we conclude that the biophysics experiments conducted here provide interesting comparative insights into HTLV-1 and HIV-1 CA but do not necessarily allow further conclusions on how the proteins themselves define immature or mature assembly characteristics.
Future experiments will be required to understand the differential contribution of CA domains to immature assembly in retroviruses. One potential avenue could be to determine high-resolution structures of HTLV-1 and HIV-1 CA chimeras42 and to solve the structure of the MA lattice in HTLV-1, as recently achieved for HIV-1 (ref. 43).
Roles of HTLV-1 CA domains in immature lattice stabilization
A unique feature reported for HTLV-1 is the varying distance of the CA layer to the viral membrane in immature Gag-based VLPs. These differences in spacing correspond to the differences in curvature and flat patches of the HTLV-1 Gag lattice as reported here (Fig. 1a) and previously30. Unlike RSV and M-PMV, which also display a larger but uniform distance of their CA layer to the viral membrane8,44, HTLV-1 does not contain any noncanonical Gag domains between MA and CA. Therefore, the distance between membrane and CA layer is likely determined by specific properties of MA and CA and not the presence of additional domains.
The variable distance of CA to the membrane could be explained by either MA detachment or a substantial structural change of MA. Our data do not provide any evidence for MA detachment from the membrane in areas of increased flexibility, as we regularly can see protein density attached to the inner leaflet of Gag-associated viral membrane (Supplementary Video 1). In addition, previous studies reported HTLV-1 MA to bind to membranes with higher affinity and efficiency but with lower specificity45,46,47, arguing against MA detachment from the viral membrane.
A study of HIV-1 and HTLV-1 Gag chimeras showed that the HTLV-1 CA-NTD is the curvature-defining element, while MA regulates the CA–membrane distance42. This was convincingly demonstrated, as any construct with the CA-NTD from HTLV-1 contained a flat lattice. A curved membrane over a flat lattice was observed only in those constructs where both CA-NTD and MA from HTLV-1 were present. The clear preference of HTLV-1 CA-NTD to shape lattices into flat regions appears to be compensated for by more flexibility in HTLV-1 Gag layers both above and below it. On the proximal side, MA seems to tolerate a larger membrane–CA distance range (Extended Data Fig. 2b), allowing to speculate whether MA can undergo a conformational change. On the distal side, CA-CTD forms sparser lateral interactions than other retroviruses, which makes the location of CA-CTD dimers less constrained with respect to the CA-NTD network. This is apparent from the blurring of the CA-CTD layer in our subtomogram averages. Accordingly, residues in the MHR, which have been implicated in immature assembly for most retroviruses, are not positioned to form interactions stabilizing the immature HTLV-1 CA-CTD hexamer. To date most structurally studied retroviruses have a stabilization element, which forms at the six-fold symmetry axis below the CA-CTD9,18,25,35,48,49. Our results here find no such domain in HTLV-1, further highlighting that HTLV-1 is structurally unique.
The fundamental difference in the organization of the CA-CTD in HTLV-1 compared to other retroviruses is quite striking. While the CA-CTD is a critical lattice-forming element for other retroviruses6,23,36,50, some of the canonical CA-CTD functions are taken up by the CA-NTD in HTLV-1 (Fig. 4).
Schematic representation of the differential role of the CA-NTD and CA-CTD in the assembly of immature HTLV-1 particles, compared to other retroviruses with known immature structures (using HIV-1 as an example). In the case of HTLV-1, immature lattice interactions are driven by the CA-NTD, while the CA-CTD drives particle formation in HIV-1.
The role of the HTLV-1 CA-CTD in immature HTLV-1 assembly
This raises a question regarding the function of the CA-CTD in immature HTLV-1 assembly. One possibility is that, in HTLV-1, a direct interaction between the NC–RNA complex and the CA-CTD may contribute to immature lattice formation. In support of this hypothesis, the HTLV-1 CA-CTD possesses a stretch of positively charged residues at its base (toward the VLP center) in helices 9 and 10 that could interact with NC and/or RNA (see underscored residues in Extended Data Fig. 1b). This would be similar to what was previously shown for M-PMV, where a basic RKK motif at the CA-CTD base is involved in promoting virus assembly and RNA packaging51. EIAV also has a basic motif, RHR, and density is observed to interact with this motif in slightly acidic in vitro assembly conditions26, which is consistent with the proposed model.
Moreover, the C-terminal region of the NC domain of HTLV-1 is negatively charged, a feature absent in other retroviruses. Accordingly, it has been shown that HTLV-1 NC is a weak RNA chaperone and nucleic acid binder52. The authors posed the hypothesis that the NC C terminus interacts intramolecularly with its zinc fingers effectively blocking nonspecific RNA binding. It was further suggested that, after Gag oligomerization into a lattice, lateral intermolecular NC interactions form between the C terminus and zinc fingers. Additional studies focusing on the HTLV-1 CA-CTD, especially considering the lack of exhaustive CA-CTD mutagenesis experiments, are warranted to more accurately determine the interactions in this region of the immature lattice.
Conclusions
HTLV-1 infectious spread occurs predominantly by cell-to-cell contact and not by cell-free virus infection of permissive T cells. The implication of the unique features of HTLV-1 Gag lattice stabilization shown here on HTLV-1 cell–cell transmission is unclear. Future work using cellular tomography on assembling and budding virions could provide a clearer understanding of the structural intermediates in the HTLV-1 lifecycle that are important in addressing this outstanding question in the field.
In addition to HTLV-1, the Deltaretrovirus genus contains multiple members, such as bovine leukemia virus. Structural comparison of the immature Gag lattice in other deltaretroviruses could help us better understand the general assembly principles and could clarify whether the observed dominance of the HTLV-1 CA-NTD in the immature Gag lattice is conserved within the genus. Our findings provide a structural basis for guiding future studies on HTLV-1 assembly and maturation, as well as for guiding the discovery of targets for therapeutic intervention. The distinct and unique nature of the HTLV-1 lattice, that is, the absence of hexameric CTD interactions and of a clear inositol hexakisphosphate (IP6)-binding site as observed in HIV-1 (ref. 53), will necessitate alternative drug targeting approaches than, for example, the structure-based maturation inhibitors developed against HIV-1 (refs. 54,55,56).
Methods
Mammalian cell culture and Gag-based VLP production and purification
The codon-optimized HTLV-1 Gag sequence, subcloned into the pCMV expression vector was ordered from Thermo Fisher Scientific. HEK293T cells (American Type Culture Collection (ATCC), CRL-3216), cultured in 10 × 10-cm dishes, were transfected with the expression plasmid using Lipofectamine LTX Reagent with PLUS Reagent (Thermo Fisher Scientific, 15338100) according to the manufacturer’s instructions. Cultivation medium containing released HTLV-1 Gag-based VLPs was obtained 40 h after transfection by ultracentrifugation. This and all further centrifugation steps were performed at 4 °C. Cultivation medium was first clarified by centrifugation for 5 min at 1,500g and subsequent filtering through a syringe-mounted PVDF filter with 0.45-µm pore size (Merck Millipore). VLPs were pelleted through a 20% sucrose cushion at 125,000g for 120 min. The pellet was briefly air-dried and resuspended in 150 μl of PBS. Pooled and resuspended pellets were applied on a 6–18% Optiprep gradient and centrifuged at 235,000g for 90 min. Fractions containing Gag-based VLPs were pooled, diluted 1:4 in PBS and pelleted without a cushion at 260,000g for 45 min. The final pellet was then resuspended in 12 μl of PBS.
Cryo-ET for HTLV-1 Gag-based VLPs
Grid preparation for HTLV-1 Gag-based VLPs
The resuspended Gag-based VLP pellet was kept on ice and mixed with 10-nm colloidal gold suspension before vitrification. Then, 2.5 µl of solution was applied on glow-discharged C-flat 2/2 3C grids and vitrified in liquid ethane using a Leica EM GP2 plunge-freezer. Grids were blotted in back-side mode for 3–4 s. The humidity chamber was conditioned to 8 °C and 90% relative humidity. Grids were stored in liquid nitrogen until imaging.
Imaging of HTLV-1 Gag-based VLPs
Data were collected on a Thermo Fisher Scientific Krios G3i equipped with Gatan Quantum K3 using SerialEM for data acquisition57. Areas of interest for high-resolution data collection were identified in low-magnification montages. Before tomogram acquisition, gain references were acquired and the filter was fully tuned using DigitalMicrograph. Microscope tuning was performed using the FEI AutoCTF software or SerialEM tuning functions. The slit width of the filter was set to 20 eV. The magnification was set to ×80,000, resulting in a pixel spacing of 1.381 Å. Tilt images were acquired as movies in super-resolution mode consisting of eight dose fractions. The dose rate was set at ~18 electrons per pixel per s. Tilt series were acquired using a dose-symmetric tilt scheme58, with a tilt range from 0° to −60° and +60° in 3° steps, and at nominal defocus between −1.25 and −3.5 µm. The total dose per tilt image was 3.5 e− per Å2.
Cryo-ET image processing of HTLV-1 Gag-based VLPs
A schematic overview of the individual processing steps for the HTLV-1 Gag-based VLPs is shown in Extended Data Fig. 2a. Tilt images with blocked field of view were removed before further steps. Tilt series were aligned using the Etomo package (as part of the IMOD package version 4.9.12) and exposure-filtered, as published previously, using a custom Matlab script9,59,60. Contrast transfer function (CTF) estimation was performed using ctffind (4.1.10).
Template matching was performed essentially as previously published61. A template was generated using subtomogram alignment of the CA layer from round VLPs from 12 tomograms. A cylindrical mask centered at a CA hexamer and encompassing two adjacent CA rings was applied to the template. In-plane and cone angles were scanned in steps over a range of 180° and 360°, in 6° and 8° steps, respectively. Lattice connectivity analysis (as described previously61) was performed to remove peaks outside of lattice and annotate VLPs. The following criteria were used to assess pairs of local cross-correlation maxima: the minimum and maximum spacing between hexamers was 45 Å and 110 Å; the minimum and maximum curvature was −15° and 25°. Only networks containing >20 cross-correlation peaks were considered as patches of lattice and used for subsequent steps.
For initial subtomogram alignment, tomograms were reconstructed in novaCTF62 using two-dimensional (2D) CTF correction of 2× binned stacks with the multiplication algorithm. Bin8 and bin4 alignments were performed in Dynamo (version 1.1.133)63 and subTOM26. A conservative low-pass filter of ≥25 Å was used during these alignments. Given the irregular and variable shape of Gag-based VLPs, we refrained from using C6 rotational symmetry, which is conventionally used for processing of immature retroviral lattices. Instead, we used C2, which allows for capturing CA hexamer deformation. After bin4 alignments, the VLP list was manually curated to remove particle formations outside enveloped VLPs.
Preprocessing for the Warp-RELION-M pipeline was performed in Warp (version 1.0.9)64. Tilt series alignments from Etomo (IMOD version 4.9.12) and subtomogram alignments from Dynamo (version 1.1.133) and subTOM were imported and used for subtomogram reconstruction. The 3D refinement was performed in RELION 3.0.8 at bin4 and bin2, using subtomograms reconstructed in Warp at the respective binning. The 3D classification was performed in RELION 3.0.8 (ref. 65) at bin2 to select a more homogeneous CA-CTD population. A cylindrical mask encompassing six CA-CTD dimers was used during the classification. For both CA-NTD and CA-CTD, a separate species was set up masking the respective layer. The CA-NTD species contained all particles from the 3D refinement at bin2, whereas the CA-CTD species contained just particles from a selected class (Extended Data Fig. 2a). Five rounds of multiparticle refinement were performed, refining particle positions, 2D and 3D warping and CTF parameters. Frame alignment refinement was performed in the last two rounds.
Model building for the immature HTLV-1 CA lattice from the Gag-based VLP-derived structure
The model of the immature HTLV-CA assembly was generated by rigid-body fitting the HTLV-1 CA-NTD derived from Protein Data Bank (PDB) 1QRJ (ref. 12) into our EM density using the fit in map option in UCSF Chimera. As we noted that the experimentally derived HTLV-1 CA-CTD from PDB 1QRJ resulted in severe clashes on the CA-CTD dimer interface upon rigid-body fitting, we used a computationally predicted model of the CA-CTD (CA residues 129–207) using ColabFold66. Interestingly, this predicted model fitted into our EM density better and resulted in fewer clashes across the dimeric interface. Hence, we used this rigid-body-fitted prediction for modeling the CA-CTD.
Protein purification and in vitro assembly
Cloning of truncated Gag constructs for in vitro assembly
The sequences encoding HTLV-1 MA126CANC (Gag residues 126–344), HTLV-1 CA (131–344), HTLV-1 CA-NTD (131–258) and HTLV-1 CA-CTD (259–344) were cloned into pET28 expression vector in frame with 6xHis-SUMO tags using standard molecular cloning methods. The same procedure was done for cloning HIV-1 CA (133–363), HIV-1 CA-NTD (133–278) and HIV-1 CA-CTD (279–363).
Protein expression and purification of HTLV-1 MA126CANC
An overnight culture of BL21 carrying the pET28/His-SUMO-MA126CANC construct was prepared to inoculate a total of 2 L of lysogeny broth (LB) medium supplemented with kanamycin. Following incubation at 37 °C with shaking at 210 r.p.m. until the bacterial culture reached an optical density at 600 nm (OD600) of 0.5–0.7, protein expression was induced by the addition of IPTG at a final concentration of 1 mM. Induction was carried out for 6 h at 37 °C with shaking at 210 r.p.m.
Bacterial cells were harvested by centrifugation at 6,000g for 15 min. The resulting cell pellet was dissolved in resuspension buffer (20 mM Tris, 500 mM NaCl, 2 µM ZnCl2, 5% glycerol, 1 mM PMSF and 1 mM TCEP, pH 8.0).
Protein was extracted from cells by cell lysis through three freeze–thaw cycles involving freezing at −80 °C followed by thawing at 42 °C. After 45–60 min of centrifugation of the lysate at 21,000g and 4 °C, the supernatant was collected and subjected to nucleic acid precipitation by treatment with 10% PEI (polyethylenimine) at a final concentration of 0.3% under stirring at 4 °C for 30 min. Afterward, the mixture was centrifuged for 10 min at 10,000g and 4 °C. The remaining supernatant was treated with ammonium sulfate at a final concentration of 40% under stirring at 4 °C overnight for precipitation of His-SUMO-MA126CANC. Precipitated protein was collected by centrifugation at 10,000g and 4 °C for 20 min.
Purification of His-SUMO-MA126CANC was carried out using anion-exchange chromatography and affinity chromatography. For anion-exchange chromatography the protein was dissolved in buffer composed of 20 mM Tris and 2 mM TCEP at pH 8.0 and was then applied to a 5-ml HiTrap SP HP column (Cytiva, 17115201) before being eluted with high-salt buffer composed of 20 mM Tris, 500 mM NaCl and 2 mM TCEP at pH 8.0.
The protein sample was then transferred to a 1-ml HisTrap FF column (Cytiva, 17531901) for protein purification by affinity chromatography. Bound protein was treated with wash buffer (20 mM Tris, 500 mM NaCl, 20 mM imidazole and 2 mM TCEP, pH 8.0) before it was eluted with high-concentration imidazole buffer (20 mM Tris, 500 mM NaCl, 250 mM imidazole and 2 mM TCEP, pH 8.0).
To remove imidazole, the protein solution was dialyzed overnight at 4 °C in tubes of the Pur-A-Lyzer Maxi 6000 dialysis kit (Sigma, PURX60100-1KT) against dialysis buffer (20 mM Tris, 500 mM NaCl and 0.5 mM TCEP, pH 8.0). During dialysis, the sample was treated with N-terminally His-tagged Ulp1 protease for removal of the His-SUMO tag from MA126CANC. The protease was later removed by reapplication of the sample to a 1-ml HisTrap FF column resulting in high-affinity binding of His-Ulp1 to the nickel Sepharose resin. MA126CANC protein exhibits low affinity for the packing material of this column. Hence, it was eluted with 20 mM imidazole wash buffer. After another dialysis step, 150–200-µl aliquots of purified protein were flash-frozen in liquid nitrogen and stored at −80 °C.
Assembly of MA126CANC tubes
Protein stored in dialysis buffer of 500 mM NaCl was first mixed with an equal volume of salt-free buffer (20 mM Tris and 0.5 mM TCEP, pH 8.0). The protein sample was then transferred into a Pierce concentrator tube with a 10-kDa molecular weight cutoff (MWCO) (Thermo Fisher Scientific, 88513) and protein concentration was increased to 12–20 mg ml−1 by centrifugation at 15,000g and 4 °C for 12 min. Then, 10 µl of protein (120–200 µg) was mixed with 2 µl of 5 mg ml−1 GT50 nucleotides, 1 µl of 50 mM EDTA and 37 µl of salt-free buffer.
The reaction mixture with a final NaCl concentration of 50 mM was incubated overnight at 4 °C. After incubation, an aliquot of the sample was subjected to negative staining for confirmation of VLP presence by transmission EM on a Tecnai T10.
Protein expression and purification of HTLV-1 CA, CA-NTD and CA-CTD
Overnight cultures of BL21 carrying pET28a vectors containing a His-SUMO tag, with HTLV-1 CA, CA-NTD and CA-CTD inserts were grown at 37 °C and 220 r.p.m. until an OD600 of 2.5 was reached. These were used to inoculate 1 L of LB medium for each construct and grown at 37 °C and 220 r.p.m. until an OD600 of 0.6 was reached, after which protein expression was induced with 1 mM IPTG for 6 h. The cultures were centrifuged at 6,000g and 4 °C for 15 min to collect the cell pellets, which were subsequently dissolved in lysis buffer (20 mM Tris, 500 mM NaCl, 2 µM ZnCl2, 5% glycerol, 1 mM PMSF and 1 mM TCEP, pH 8.0) and subjected to 3–5 freeze–thaw cycles to disrupt the cells.
The resulting lysed cells were centrifuged at 21,000g for 45 min to separate protein from insoluble cell debris, followed by protein purification from nucleic acid by precipitating the nucleic acid with 10% PEI at a final concentration of 0.3% for 10 min under stirring at 4 °C.
The solutions were centrifuged again, at 10,000g and 4 °C for 10 min to collect the precipitated nucleic acid and ammonium sulfate was subsequently added to the supernatant to a saturation of 40% under stirring at 4 °C overnight to precipitate the proteins. The solutions were centrifuged again at 13,000g to collect the proteins in a pellet before resuspending the protein pellets in wash buffer (20 mM Tris, 20 mM imidazole, 500 mM NaCl and 2 mM TCEP, pH 8).
The proteins were then purified using affinity chromatography with a 1-ml HisTrap FF column according to the manufacturers’ instructions (Cytiva) and elution buffer containing 20 mM Tris, 250 mM imidazole, 500 mM NaCl and 2 mM TCEP, pH 8.
Following purification, the His-SUMO tag was cleaved from the proteins using Ulp1 protease and the proteins placed for dialysis at 4 °C overnight in dialysis buffer (20 mM Tris, 500 mM NaCl and 5 mM TCEP, pH 8). A second affinity chromatography purification step with Ni-NTA columns was performed to purify the proteins from the cleaved His-SUMO tag, followed again by dialysis in dialysis buffer at 4 °C overnight.
Protein expression and purification of HIV-1 CA, CA-NTD and CA-CTD
E. coli BL21 cells were transformed with the respective pET28 vectors and grown on agar plates supplemented with 50 µg ml−1 kanamycin. Overnight cultures of the BL21 cells carrying the respective vectors were used to inoculate 1 L of LB medium supplemented with kanamycin and 10 mM HEPES. The cultures were grown at 37 °C and 220 r.p.m. until an OD600 of 0.9–1.1 was reached, whereby protein expression was induced with 0.3 mM IPTG and protein expressed for 5 h at 30 °C and 200 r.p.m.
The cultures were centrifuged at 6,000g and 4 °C for 15 min to collect the cell pellets that were subsequently dissolved in lysis buffer (20 mM Tris, 500 mM NaCl, 2 μM ZnCl2, 5% glycerol, 1 mM PMSF and 5 mM β-mercaptoethanol, pH 8.0) and subjected to 3–5 freeze–thaw cycles to disrupt the cells.
The resulting lysed cells were centrifuged at 21,000g for 45 min to separate protein from insoluble cell debris, followed by protein purification using affinity chromatography with a 1-ml HisTrapTM FF column according to the manufacturer’s instructions (Cytiva) and elution buffer containing 20 mM Tris, 250 mM imidazole, 500 mM NaCl and 2 mM TCEP at pH 8. Following purification, the His-SUMO tag was cleaved from the proteins using Ulp1 protease and the proteins placed for dialysis at 4 °C overnight in dialysis buffer (20 mM Tris, 500 mM NaCl and 5 mM TCEP, pH 8). A second affinity chromatography purification step with Ni-NTA columns was performed to purify the proteins from the cleaved His-SUMO tag. Proteins were then snap-frozen and stored at −80 °C.
Assembly reactions and preparation for nanoDSF
HTLV-1 and HIV-1 CA, CA-NTD, and CA-CTD were separately concentrated in Pierce concentrator tubes (10-kDa MWCO for CA and CA-NTD and 3-kDa MWCO for CA-CTD). Concentrations of monomeric HTLV-1 CA, CA-NTD and CA-CTD were normalized to the tryptophan content of the respective constructs, with 0.96 mg ml−1 for CA, 1.15 mg ml−1 for CA-NTD and 0.70 mg ml−1 for CA-CTD under preassembly conditions (50 mM MES and 200 mM NaCl, pH 6) in a volume of 10 μl and placed in Prometheus Series high-sensitivity capillaries (NanoTemper, PR-C006). The same was performed for HIV-1 CA (0.82 mg ml−1), CA-NTD (0.66 mg ml−1) and CA-CTD (2 mg ml−1) but with a different preassembly buffer (50 mM MES, 150 mM NaCl and 5 mM β-mercaptoethanol, pH 6).
For measurement of HTLV-1 postassemblies, the proteins were first placed in assembly buffer (50 mM MES and 200 mM NaCl, pH6) with a final concentration of ~680 µM and were incubated at 4 °C for 4 h and 26 °C overnight for assembly to occur. A volume of 10 μl was then placed in Prometheus Series high-sensitivity capillaries (NanoTemper, PR-C006). For visualizing assemblies of HTLV-1 CA on a Tecnai T10 transmission EM instrument, an aliquot was used for negative staining with 2% uranyl acetate (UA).
Similarly, for measurement of HIV-1 postassemblies, protein was concentrated to 4 mg ml−1 (150 µM) and dialyzed into 50 mM MES, 150 mM NaCl and 5 mM β-mercaptoethanol at pH 6 overnight. IP6 was then added to a final concentration of 4 mM and the protein was incubated at 30 °C for 2 h. A volume of 10 µl was then placed in Prometheus Series high-sensitivity capillaries (NanoTemper, PR-C006). For visualizing assemblies of HIV-1 CA on a Tecnai T10 transmission EM instrument, an aliquot was used for negative staining with 2% UA.
NanoDSF, backreflection and static light scattering
The capillaries were loaded into a Prometheus Panta instrument (NanoTemper Technologies). NanoDSF, backreflection and DLS were measured continuously over a temperature ramp of 5 °C min−1 from 15 °C to 95 °C at 40% light-emitting diode excitation power. All scans were single reads of three replicates performed at least three times. Data were analyzed using custom Python scripts (Python version 3.11.4).
Cryo-ET for HTLV-1 MA126CANC tubes
Grid preparation for MA126CANC tubes
After in vitro assembly, MA126CANC tubes were kept at 4 °C until plunge-freezing. Then, 2/2 3C C-Flat grids coated with a 2-nm support carbon layer were glow-discharged in the presence of amylamine. Grids were then incubated on a 5-μl drop of sample for 10 min. The rest of the sample was mixed with 10-nm colloidal gold and 2.5 μl of this solution was added to the incubated grids before vitrification. The samples were vitrified in liquid ethane using a Leica GP2 plunger with front-side blotting (blot time, 3–4 s; humidity, 90–95%; temperature, 10 °C) and stored in liquid nitrogen until imaging.
Imaging of MA126CANC tubes
For in vitro assembled MA126CANC tubes, we acquired data on two systems. One dataset was collected on a Thermo Fisher Scientific Krios G3i equipped with Gatan Quantum K3 (system 1). The second dataset was collected on a FEI Titan Krios, operated at 300 keV, equipped with a Gatan Quantum 967 LS energy filter and a Gatan K2xp direct electron detector (system 2). The slit width of the filter was set to 20 eV on both systems. SerialEM was used for data acquisition in both cases. Areas of interest for high-resolution data collection were identified in low-magnification montages. Before tomogram acquisition, gain references were acquired and the filter was fully tuned. Filter tuning was performed in DigitalMicrograph. Microscope tuning was performed using the FEI AutoCTF software or SerialEM tuning functions.
On system 1, the magnification was set to ×80,000, resulting in a pixel spacing of 1.381 Å. Tilt images were acquired as movies in super-resolution consisting of eight dose fractions. The dose rate was set at ~21 eps. On system 2, the magnification was set to ×105,000, resulting in a pixel spacing of 1.327 Å. Tilt images were acquired as movies in super-resolution consisting of ten dose fractions. The dose rate was set at ~5.4 eps.
Tilt series were acquired using a dose-symmetric tilt scheme58, with a tilt range from 0° to −60° and +60° in 3° steps, at a nominal defocus between −1.5 and −3.5 µm. The total dose per tilt image was 3.5 e− per Å2.
Cryo-ET image processing of MA126CANC tubes
A schematic overview over the individual processing steps for the MA126CANC tubes is shown in Extended Data Fig. 5a. Before tilt series alignment, movies were aligned using the IMOD ‘alignframes’ function. Tilt images with a blocked field of view were removed before further steps. Tilt series were aligned using the Etomo (part of the IMOD package, version 4.9.12) and exposure-filtered using a custom Matlab (version R2018b) script9,59. CTF estimation was performed using ctffind (4.1.10).
For initial subtomogram alignment described below, CTF-corrected tomograms were reconstructed in novaCTF using the multiplication algorithm with a slab size of 15 nm. Bin8 and bin4 alignments were performed in Dynamo (version 1.1.133) and subTOM. A conservative low-pass filter of ≥25 Å was used during these alignments.
Preprocessing for the Warp-RELION-M pipeline was performed in Warp. Tilt series alignments from Etomo and subtomogram alignments from Dynamo and subTOM were imported and used for subtomogram reconstruction. The 3D refinement was performed in RELION 3.0.8 at bin4 and bin2, using subtomograms reconstructed in Warp at the respective binning.
Classification of tube geometries for structure determination
Because the tube geometry was variable for the CANC tubes, tubes were sorted into groups containing similar tube geometries. Specifically, the angle between the tube axis and the hexamer–hexamer connection was used to distinguish between different tube geometries (Extended Data Fig. 5b).
To this end, first, a de novo reference was generated for each tube separately by subtomogram alignment of regularly spaced volumes along a given tube using Dynamo and subTOM. The initial alignments were performed at bin8. The first step was to remove particles that converged to the same spatial position using distance cleaning. A cross-correlation threshold was set manually for each tomogram upon visual inspection to remove misaligned and bad subvolumes. The positions of subtomograms derived from the initial alignments were used for geometry analysis. Tubes were sorted into groups such that the maximum angle difference within one group was 5°. Subtomogram alignment and tube grouping were then refined at bin4.
Merging datasets for multiparticle refinement
Because the datasets acquired on the two microscope systems differed in pixel size, they were processed separately until the multiparticle refinement step. Then, because the pixel size of system 1 was not calibrated, we used the fitmap function of UCSF Chimera67 to maximize the overlap between maps generated using systems 1 and 2, while varying the pixel size of system 1. For further processing, we used this estimated pixel size rather than the nominal value. On this basis, we estimated the actual pixel size on system 1 to be 1.339 Å. To unify the pixel size used for multiparticle refinement, we used a box size that yielded an integer upon dividing with both pixel sizes. A box size of 321.3 Å fulfilled the criteria.
Multispecies refinement of MA126CANC tubes
The 3D refinement in RELION 3.0.8 was performed for each group of tubes separately at bin4 and bin2. Subtomograms were reconstructed in Warp. The respective halfmaps and particle coordinate star files were used as a starting point for multiparticle refinement in M, gradually adding refinement options. Five rounds of refinement were performed as described previously68. Then, 2D and 3D warping was performed in the first round, particle pose refinement was added in the second round, CTF refinement was added in the third round and stage angle refinement and movie refinement were added in the fourth and final iterations. This yielded nine CA-NTD hexamer structures at resolutions from 3.7 to 7.0 Å. To increase the resolution for model building, the two most populated groups sharing similar geometry were pooled, subtomograms were reconstructed using M at a pixel size of 1.77 Å and the pooled population was subjected to Autorefine 3D in RELION. Afterward, two final rounds of M refinement were performed. The final 3D refinement was performed in RELION, with subtomograms reconstructed using M at bin1.
Model refinement
The initial model for HTLV-1 CA-NTD was obtained by adjusting PDB 1QRJ, by flexibly fitting it into the EM density using the Isolde plugin of ChimeraX69,70. The final model was then built by iterating between real-space refinement in PHENIX71 and manual adjustments in Coot72.
Data visualization and figure preparation
The 3D volumes were visualized using UCSF Chimera (version 1.17.3)67, ChimeraX (versions 1.7.1 and 1.8)70 and IMOD (version 4.9.12)60. Images of 3D volumes and models were made using Chimera or ChimeraX or by ray tracing in PyMOL (version 1.3r1; Schrödinger). Figures were prepared using Adobe Illustrator. Videos were generated in ChimeraX and Adobe Premiere Pro 2023.
The 1D density plots were calculated from the bin1 (CANC tubes) or bin2 (Gag VLPs) unsharpened unmasked average, assuming a tube diameter of 60 nm and VLP diameter of 130 nm.
Mutagenesis of residues involved in stabilizing the immature CA-NTD
To characterize the residues involved in stabilizing the immature CA-NTD, we conducted mutagenesis on key residues. First, we substituted residue L43 to alanine, as L43 may contact M17 and Y61. Second, we conducted mutagenesis of the trimeric CA-NTD interface. Specifically, residue pairs H72-Q73, Y90-N91, P92-L93, G95-P96, L97-R98 and V99-Q100 in the CA-NTD were created using the Gibson assembly method73. The correct creation of the mutants was verified by Sanger DNA-sequencing analysis. The efficiency of immature particle production was analyzed by quantifying Gag proteins in released particles through harvesting cell culture supernatants and using immunoblot analysis with a mouse monoclonal anti-HTLV p24 (CA) antibody (Santa Cruz Biotechnology, sc-53891). To determine the particle production (that is, particle release) efficiency of mutants relative to that of the wild type (WT), the Gag expression levels detected in cell lysates of the WT and mutants were normalized relative to the respective glyceraldehyde 3-phosphate dehydrogenase (GAPDH) level. The mutant particle release relative to that of the WT was then determined by the ratio of Gag levels detected from cell culture supernatants to that from the normalized Gag levels from the cell lysate, with WT particle release being set to 100. Results were plotted using GraphPad Prism 6.0. The relative significance between the WT and mutant was determined using an unpaired t-test. Cryo-EM analysis of immature particle morphologies of the L43A mutant was performed by producing VLPs from cells and concentrating by ultracentrifugation.
Cryo-EM analysis of L43A particle morphology
The pN3-HTLV-1-Gag L43A and the HTLV-1 Env expression plasmids were cotransfected into HEK293T/17 cells (ATCC, CRL-11268) using GenJet (version II) at a 10:1 molar ratio as previously described31,74. At 48 h after ransfection, cell culture supernatants were harvested and centrifuged at 1,800g for 10 min, followed by passing through a 0.2-µm filter. Particles were then concentrated by ultracentrifugation in a 50.2 Ti rotor at 35,000 r.p.m. for 90 min through an 8% Optiprep cushion. Particle pellets were resuspended in about 200 µl of STE buffer before ultracentrifugation through a 10–30% Optiprep step gradient at 45,000 r.p.m. for 3 h. The particle band was removed from the gradient by puncturing the side of the thin wall ultraclear centrifuge tube (Beckman Coulter) with a hypodermic syringe needle and pelleted in STE buffer at 40,000 r.p.m. for 1 h using an SW55 Ti rotor. The resulting particle pellet was then resuspended in approximately 10 µl of STE buffer and frozen at −80 °C until analyzed by cyro-EM.
Cryo-EM analysis of L43A particles was performed as previously described31,37,42,74. Particle samples were thawed on ice and approximately 3.0 µl of purified virus particles were applied to glow-discharged Quantifoil R1.2/1.3 400-mesh holey carbon-coated copper grids. The grids were then manually blotted by a piece of filter paper and plunge-frozen in liquid ethane. The frozen grids were stored in liquid nitrogen until imaging analysis.
Cryo-EM analysis was performed using a Tecnai FEI G2 F30 FEG transmission EM instrument (FEI) at liquid nitrogen temperature operating at 300 kV. Images were recorded at a nominal magnification of ×20,000 at 5–10 μm underfocus using a Gatan 4k-by-4k charge-coupled device camera. At least 250 individual immature particles of L43A mutant were collected.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The EM density maps and models for the immature HTLV-1 CA-NTD lattice and representative tomograms were deposited to the EM Data Bank under accession codes EMD-17929, EMD-17930, EMD-17931, EMD-17932, EMD-17933, EMD-17934, EMD-17935, EMD-17936, EMD-17937, EMD-17938, EMD-17939, EMD-17940, EMD-17941, EMD-17942 and EMD-17943 and the PDB under accession codes 8PU6, 8PU7, 8PU8, 8PU9, 8PUA, 8PUB, 8PUC, 8PUD, 8PUE, 8PUF, 8PUG and 8PUH. PDB 1QRJ was used as starting model to derive a refined model of the HTLV-1 CA-NTD. The models used to generate Fig. 3 and Extended Data Fig. 3 were PDB 5L93 (HIV-1), PDB 5A9E (RSV), PDB 6HWI (M-PMV) and PDB 6HWW (MLV). The UniProt codes for the protein sequences used for generating Extended Data Fig. 9 were P03345 (GAG_HTL1A), P03346 (GAG_HTLV2), Q0R5R4 (GAG_HTL32), P25058 (GAG_BLVAU), P04585 (POL_HV1H2), P15832 (GAG_HV2D2), P16087 (GAG_FIVPE), P69732 (GAG_EIAVY), P03322 (GAG_RSVP), P0C776 (GAG_ALVA), P07567 (GAG_MPMV), P31622 (GAG_JSRV), P10258 (GAG_MMTVB), P03355 (POL_MLVMS), P10262 (GAG_FLV) and Q9TTC2 (GAG_KORV). Source data are provided with this paper.
References
Gessain, A., Gessain, A. & Cassar, O. Epidemiological aspects and world distribution of HTLV-1 infection. Front. Microbiol. 3, 388 (2012).
Cook, L. B., Elemans, M., Rowan, A. G. & Asquith, B. HTLV-1: persistence and pathogenesis. Virology 435, 131–140 (2013).
Mahieux, R. & Gessain, A. Adult T-cell leukemia/lymphoma and HTLV-1. Curr. Hematol. Malig. Rep. 2, 257–264 (2007).
Freed, E. O. HIV-1 assembly, release and maturation. Nat. Rev. Microbiol. 13, 484–496 (2015).
Ako-Adjei, D., Johnson, M. C. & Vogt, V. M. The retroviral capsid domain dictates virion size, morphology, and coassembly of gag into virus-like particles. J. Virol. 79, 13463–13472 (2005).
Mattei, S., Schur, F. K. M. & Briggs, J. A. Retrovirus maturation—an extraordinary structural transformation. Curr. Opin. Virol. 18, 27–35 (2016).
Bush, D. L. & Vogt, V. M. In vitro assembly of retroviruses. Annu. Rev. Virol. 1, 561–580 (2014).
Schur, F. K. M. et al. Structure of the immature HIV-1 capsid in intact virus particles at 8.8 Å resolution. Nature 517, 505–508 (2015).
Schur, F. K. M. et al. An atomic model of HIV-1 capsid-SP1 reveals structures regulating assembly and maturation. Science 353, 506–508 (2016).
Zhao, G. et al. Mature HIV-1 capsid structure by cryo-electron microscopy and all-atom molecular dynamics. Nature 497, 643–646 (2013).
Schirra, R. T. et al. A molecular switch modulates assembly and host factor binding of the HIV-1 capsid. Nat. Struct. Mol. Biol. 30, 383–390 (2023).
Khorasanizadeh, S., Campos-Olivas, R., Clark, C. & Summers, M. Sequence-specific 1H, 13C and 15N chemical shift assignment and secondary structure of the HTLV-I capsid protein. J. Biomol. NMR 14, 199–200 (1999).
Tang, C., Ndassa, Y. & Summers, M. F. Structure of the N-terminal 283-residue fragment of the immature HIV-1 Gag polyprotein. Nat. Struct. Biol. 9, 537–543 (2002).
Bartonova, V. et al. Residues in the HIV-1 capsid assembly inhibitor binding site are essential for maintaining the assembly-competent quaternary structure of the capsid protein. J. Biol. Chem. 283, 32024–32033 (2008).
Kingston, R. L. et al. Structure and self-association of the Rous sarcoma virus capsid protein. Structure 8, 617–628 (2000).
Campos-Olivas, R., Newman, J. L. & Summers, M. F. Solution structure and dynamics of the Rous sarcoma virus capsid protein and comparison with capsid proteins of other retroviruses. J. Mol. Biol. 296, 633–649 (2000).
Cornilescu, C. C., Bouamr, F., Yao, X., Carter, C. & Tjandra, N. Structural analysis of the N-terminal domain of the human T-cell leukemia virus capsid protein. J. Mol. Biol. 306, 783–797 (2001).
Qu, K. et al. Structure and architecture of immature and mature murine leukemia virus capsids. Proc. Natl Acad. Sci. USA 115, E11751–E11760 (2018).
Mammano, F., Ohagen, A., Höglund, S. & Göttlinger, H. G. Role of the major homology region of human immunodeficiency virus type 1 in virion morphogenesis. J. Virol. 68, 4927–4936 (1994).
Bharat, T. A. M. et al. Structure of the immature retroviral capsid at 8 Å resolution by cryo-electron microscopy. Nature 487, 385–389 (2012).
Strohalmová-Bohmová, K. et al. Role of mason-pfizer monkey virus CA–NC spacer peptide-like domain in assembly of immature particles. J. Virol. 88, 14148–14160 (2014).
Pornillos, O. & Ganser-Pornillos, B. K. Maturation of retroviruses. Curr. Opin. Virol. 36, 47–55 (2019).
Accola, M. A., Strack, B. & Göttlinger, H. G. Efficient particle production by minimal Gag constructs which retain the carboxy-terminal domain of human immunodeficiency virus type 1 capsid–p2 and a late assembly domain. J. Virol. 74, 5395–5402 (2000).
Borsetti, A., Öhagen, Å. & Göttlinger, H. G. The C-terminal half of the human immunodeficiency virus type 1 Gag precursor is sufficient for efficient particle assembly. J. Virol. 72, 9313–9317 (1998).
Schur, F. K. M., Dick, R. A., Hagen, W. J. H., Vogt, V. M. & Briggs, J. A. G. The structure of immature virus-like Rous sarcoma virus Gag particles reveals a structural role for the p10 domain in assembly. J. Virol. 89, 10294–10302 (2015).
Dick, R. A. et al. Structures of immature EIAV Gag lattices reveal a conserved role for IP6 in lentivirus assembly. PLoS Pathog. 16, e1008277 (2020).
Temple, J., Tripler, T. N., Shen, Q. & Xiong, Y. A snapshot of HIV-1 capsid–host interactions. Curr. Res. Struct. Biol. 2, 222–228 (2020).
Rayne, F., Bouamr, F., Lalanne, J. & Mamoun, R. Z. The N-terminal domain of the human T-cell leukemia virus type 1 capsid protein is involved in particle formation. J. Virol. 75, 5277–5287 (2001).
Cao, S., Maldonado, J. O., Grigsby, I. F., Mansky, L. M. & Zhang, W. Analysis of human T-cell leukemia virus type 1 particles by using cryo-electron tomography. J. Virol. 89, 2430–2435 (2015).
Maldonado, J. O., Cao, S., Zhang, W. & Mansky, L. M. Distinct morphology of human T-cell leukemia virus type 1-like particles. Viruses 8, 132 (2016).
Martin, J. L., Cao, S., Maldonado, J. O., Zhang, W. & Mansky, L. M. Distinct particle morphologies revealed through comparative parallel analyses of retrovirus-like particles. J. Virol. 90, 8074–8084 (2016).
Carnes, S. K., Sheehan, J. H. & Aiken, C. Inhibitors of the HIV-1 capsid, a target of opportunity. Curr. Opin. HIV AIDS 13, 359 (2018).
Wright, E. R. et al. Electron cryotomography of immature HIV-1 virions reveals the structure of the CA and SP1 Gag shells. EMBO J. 26, 2218–2226 (2007).
Tegunov, D., Xue, L., Dienemann, C., Cramer, P. & Mahamid, J. Multi-particle cryo-EM refinement with M visualizes ribosome–antibiotic complex at 3.5 Å in cells. Nat. Methods 18, 186–193 (2021).
Wagner, J. M. et al. Crystal structure of an HIV assembly and maturation switch. eLife 5, e17063 (2016).
Bharat, T. A. M. et al. Cryo-electron microscopy of tubular arrays of HIV-1 Gag resolves structures essential for immature virus assembly. Proc. Natl Acad. Sci. USA 111, 8233–8238 (2014).
Martin, J. L. et al. Critical role of the human T-cell leukemia virus type 1 capsid N-terminal domain for Gag–Gag interactions and virus particle assembly. J. Virol. 92, e00333-18 (2018).
Real-Hohn, A., Groznica, M., Löffler, N., Blaas, D. & Kowalski, H. nanoDSF: in vitro label-free method to monitor picornavirus uncoating and test compounds affecting particle stability. Front. Microbiol. 11, 1442 (2020).
Llorens, C., Fares, M. A. & Moya, A. Relationships of Gag–Pol diversity between Ty3/Gypsy and Retroviridae LTR retroelements and the three kings hypothesis. BMC Evol. Biol. 8, 276 (2008).
Madeira, F. et al. Search and sequence analysis tools services from EMBL-EBI in 2022. Nucleic Acids Res. 50, W276–W279 (2022).
Letunic, I. & Bork, P. Interactive Tree Of Life (iTOL) v5: an online tool for phylogenetic tree display and annotation. Nucleic Acids Res. 49, W293–W296 (2021).
Martin, J. L. et al. Disparate contributions of human retrovirus capsid subdomains to Gag–Gag oligomerization, virus morphology, and particle biogenesis. J. Virol. 91, e00298–17 (2017).
Kun, Q. et al. Maturation of the matrix and viral membrane of HIV-1 Structural changes in HIV. Science 373, 700–704 (2021).
Briggs, J. A. G., Johnson, M. C., Simon, M. N., Fuller, S. D. & Vogt, V. M. Cryo-electron microscopy reveals conserved and divergent features of Gag packing in immature particles of Rous sarcoma virus and human immunodeficiency virus. J. Mol. Biol. 355, 157–168 (2006).
Inlora, J., Chukkapalli, V., Derse, D. & Ono, A. Gag localization and virus-like particle release mediated by the matrix domain of human T-lymphotropic virus type 1 Gag are less dependent on phosphatidylinositol-(4,5)-bisphosphate than those mediated by the matrix domain of HIV-1 Gag. J. Virol. 85, 3802–3810 (2011).
Inlora, J., Collins, D. R., Trubin, M. E., Chung, J. Y. J. & Ono, A. Membrane binding and subcellular localization of retroviral gag proteins are differentially regulated by MA Interactions with phosphatidylinositol-(4,5)-bisphosphate and RNA. mBio 5, e02202 (2014).
Herrmann, D. et al. Structural insights into the mechanism of human T-cell leukemia virus type 1 Gag targeting to the plasma membrane for assembly. J. Mol. Biol. 433, 167161 (2021).
Accola, M. A., Höglund, S. & Göttlinger, H. G. A putative α-helical structure which overlaps the capsid–p2 boundary in the human immunodeficiency virus type 1 Gag precursor is crucial for viral particle assembly. J. Virol. 72, 2072–2078 (1998).
Cheslock, S. R. et al. Charged assembly helix motif in murine leukemia virus capsid: an important region for virus assembly and particle size determination. J. Virol. 77, 7058–7066 (2003).
von Schwedler, U. K., Stray, K. M., Garrus, J. E. & Sundquist, W. I. Functional surfaces of the human immunodeficiency virus type 1 capsid protein. J. Virol. 77, 5439–5450 (2003).
Füzik, T. et al. Nucleic acid binding by Mason-Pfizer monkey virus CA promotes virus assembly and genome packaging. J. Virol. 90, 4593–4603 (2016).
Qualley, D. F. et al. C-terminal domain modulates the nucleic acid chaperone activity of human T-cell leukemia virus type 1 nucleocapsid protein via an electrostatic mechanism. J. Biol. Chem. 285, 295–307 (2010).
Dick, R. A. et al. Inositol phosphates are assembly co-factors for HIV-1. Nature 560, 509–512 (2018).
Bester, S. M. et al. Structural and mechanistic bases for a potent HIV-1 capsid inhibitor. Science 370, 360–364 (2020).
Link, J. O. et al. Clinical targeting of HIV capsid protein with a long-acting small molecule. Nature 584, 614–618 (2020).
Adamson, C. S., Salzwedel, K. & Freed, E. O. Virus maturation as a new HIV-1 therapeutic target. Expert Opin. Ther. Targets 13, 895–908 (2009).
Mastronarde, D. N.SerialEM: a program for automated tilt series acquisition on Tecnai microscopes using prediction of specimen position. Microsc. Microanal. 9, 1182–1183 (2003).
Hagen, W. J. H., Wan, W. & Briggs, J. A. G. Implementation of a cryo-electron tomography tilt-scheme optimized for high resolution subtomogram averaging. J. Struct. Biol. 197, 191–198 (2017).
Grant, T. & Grigorieff, N. Measuring the optimal exposure for single particle cryo-EM using a 2.6 Å reconstruction of rotavirus VP6. eLife 4, e06980 (2015).
Kremer, J. R., Mastronarde, D. N. & McIntosh, J. R. Computer visualization of three-dimensional image data using IMOD. J. Struct. Biol. 116, 71–76 (1996).
Obr, M. et al. Structure of the mature Rous sarcoma virus lattice reveals a role for IP6 in the formation of the capsid hexamer. Nat. Commun. 12, 3226 (2021).
Turoňová, B., Schur, F. K. M., Wan, W. & Briggs, J. A. G. Efficient 3D-CTF correction for cryo-electron tomography using NovaCTF improves subtomogram averaging resolution to 3.4 Å. J. Struct. Biol. 199, 187–195 (2017).
Castano-Diez, D., Kudryashev, M., Arheit, M. & Stahlberg, H. Dynamo: a flexible, user-friendly development tool for subtomogram averaging of cryo-EM data in high-performance computing environments. J. Struct. Biol. 178, 139–151 (2012).
Tegunov, D. & Cramer, P. Real-time cryo-electron microscopy data preprocessing with Warp. Nat. Methods 16, 1146–1152 (2019).
Scheres, S. H. W. RELION: implementation of a Bayesian approach to cryo-EM structure determination. J. Struct. Biol. 180, 519–530 (2012).
Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nat. Methods 19, 679–682 (2022).
Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).
Obr, M. et al. Exploring high-resolution cryo-ET and subtomogram averaging capabilities of contemporary DEDs. J. Struct. Biol. 214, 107852 (2022).
Croll, T. I. ISOLDE: a physically realistic environment for model building into low-resolution electron-density maps. Acta Crystallogr. D Struct. Biol. 74, 519–530 (2018).
Goddard, T. D. et al. UCSF ChimeraX: meeting modern challenges in visualization and analysis. Protein Sci. 27, 14–25 (2018).
Adams, P. D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D Biol. Crystallogr. 66, 213–221 (2010).
Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132 (2004).
Heydenreich, F. M. et al. High-throughput mutagenesis using a two-fragment PCR approach. Sci. Rep. 7, 6787 (2017).
Yang, H., Talledge, N., Arndt, W. G., Zhang, W. & Mansky, L. M. Human immunodeficiency virus type 2 capsid protein mutagenesis reveals amino acid residues important for virus particle assembly. J. Mol. Biol. 434, 167753 (2022).
Acknowledgements
This work was funded by the Institute of Science and Technology Austria (ISTA) and the Austrian Science Fund (grant P31445 to F.K.M.S.). Access to high-resolution cryo-ET data acquisition at European Molecular Biology Laboratory (EMBL) Heidelberg was supported through the EMBL cryo-EM platform. We thank V.-V. Hodirnau at ISTA and W. Hagen and F. Weis at EMBL Heidelberg for support in cryo-ET data acquisition. This research was also supported by the scientific service units of ISTA through resources provided by Scientific Computing, the Life Science Facility, and the EM Facility. L.M.M. was supported by National Institutes of Health grants R01 GM151775 and R21 DE032878 and by the University of Minnesota Masonic Cancer Center. D.P. was supported by the DOC doctoral fellowship program of the Austrian Academy of Sciences. R.A.D was supported by the National Institute of Allergy and Infectious Diseases (grant R01AI147890). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. Specifically, we also want to thank A. Schlögl for computational support and J. Hansen and V. Vogt for critical comments on the manuscript. We also thank the other members of the Schur lab for helpful discussions and experimental advice.
Author information
Authors and Affiliations
Contributions
Project administration, F.K.M.S. Supervision and funding acquisition, L.M.M., R.A.D. and F.K.M.S. Conceptualization, M.O. and F.K.M.S. Methodology, M.O. and F.K.M.S. Investigation, M.O., M.P., D.C., H.Y., A.T., G.P., D.P., R.A.D. and F.K.M.S. Software, M.O. Validation, formal analysis and visualization, M.O., M.P., H.Y. and F.K.M.S. Data curation, M.O. and F.K.M.S. Writing—original draft, M.O. and F.K.M.S. Writing—review and editing, M.O., M.P., D.C., H.Y, A.T., G.P., D.P., L.M.M., R.A.D. and F.K.M.S.
Corresponding author
Ethics declarations
Competing interests
M.O. is currently an employee of Thermo Fisher Scientific. Thermo Fisher Scientific had no role in study design or experimental aspects of the submitted work and did not provide any financial support for this project. The other authors declare no competing interests.
Peer review
Peer review information
Nature Structural & Molecular Biology thanks Luiza Mendonça and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. Primary Handling Editor: Katarzyna Ciazynska, in collaboration with the Nature Structural & Molecular Biology team.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data
Extended Data Fig. 1 HTLV-1 Gag/CA sequence and Gag VLP purification.
a) Schematic depiction of the domain architecture of HTLV-1 Gag and the MA126CANC truncation protein used for in vitro assembly and structure determination. Residue numbers and black triangles denote protease cleavage sites. Note the absence of non-canonical Gag domains in HTLV-1. b) Sequence of HTLV-1 MA126CANC. The positions of α-helices are shown as cyan and yellow bars. Regions for which structural data is lacking are dashed. Triangles denote protease cleavage sites. Please note that due to the used expression and purification system of our 6xHIS-SUMO construct, an ectopic Serine residue remains at the very N-terminus of the purified protein (not shown in this presented sequence). The underlined residues in Helix 9 and 10 denote positively charged residues which could form potential interactions with the NC-RNA. Tryptophan residues (W) are highlighted in green. c) Samples of Optiprep gradient (OP)-purified particles were separated by SDS–polyacrylamide gel electrophoresis. Proteins were visualized by immunoblot using an anti-HTLV-1 p24 antibody. Sucrose pellet (SP), Optiprep bottom (OP), Final Optiprep pellet (F). Positions of molecular mass standards (in kiloDaltons, kDa) are indicated. The molecular weight (MW) of HTLV-1 Gag of 47.5 kDa is annotated. Cell culture production of HTLV-1 Gag VLPs was first established in small scale experiments. The expression was repeated at least 3x with consistent results. The purification blot shown in (c) was performed once to produce the sample in a sufficient quantity to be analysed using cryo-ET.
Extended Data Fig. 2 Image processing of HTLV-1 Gag VLPs.
a) Image processing workflow for the subtomogram averaging and multiparticle refinement of the CA layer from HTLV-1 Gag-based VLPs. See Methods for more details. b) XZ-slices through bin4 averages of the immature HTLV-1 Gag lattice, showing the varying distance of the CA-layer with respect to the viral membrane (VM). The distance was measured from the outer VM layer to the electron-lucent layer between CA-NTD and CA-CTD. c) Histogram showing the measured distances of the CA-layer to the viral membrane. The distance was determined by re-aligning particles to viral membrane layers. The individual CA-membrane distances then correspond to the mean value + particle Z-shift. N = 51,700. d) Fourier-shell correlation (FSC) for the independently aligned CA-NTD (blue line) and CA-CTD layer (red line), showing a resolution of 5.9 Å and 6.2 Å at the 0.143 criterion, respectively.e) Local resolution measurements for the two cryo-EM density maps generated by focusing either on the CA-NTD (top) or the CA-CTD (bottom). The color code for the local resolution in Å is shown below. The maps are shown in two views, once as seen from the outside of the VLP (left) and in a side view (right). The viral membrane (VM), the CA-NTD, and CA-CTD are annotated in the side view. The CA hexamer center is indicated with a small black hexagon. f) Model of HTLV-1 CA-NTD (cyan) and CA-CTD (orange) rigid-body fitted into a composite map of the immature HTLV-1 CA lattice and shown from two views (as seen from the outside of the VLP on the top, and a side view on the bottom). The density of the CA-NTD is from the cryo-electron microscopy density map refined on the CA-NTD (EMD-17942) and the density for the CA-CTD from EMD-17943). The CA hexamer center is indicated with a small black hexagon.
Extended Data Fig. 3 CA-CTD interactions in immature retroviral lattices.
a) Comparison of different retroviral immature CA-CTD dimer structures. The pdb accession codes for the different retroviruses are indicated. The same models have been used to generate the data shown in panel (b). b) Comparison of the immature CA-CTD hexamer assembly. The main interaction interfaces described previously to stabilize the hexamer are annotated. The CA-CTD assembly unit in HIV-1 is described in9. For RSV and M-PMV the electron microscopy densities for the RSV SP and the M-PMV spacer peptide-like region are shown, as for these regions no molecular model has yet been obtained.
Extended Data Fig. 4 Expression and purification of retrovirus CA proteins for in vitro experiments.
a) Samples of purified HTLV-1 CA protein variants were separated by SDS–PAGE (15%) and stained via Coomassie brilliant blue. Positions of molecular mass standards (in kiloDaltons, kDa) are indicated. HTLV-1 CA is full-length from residues 131–344 (using Gag-numbering), CA-NTD includes residues 131–258 and CA-CTD includes residues 259–344. b) Negative Stain TEM micrographs of HTLV-1 CA tubes. Scale bars are annotated in the figure. c) Samples of purified HIV-1 CA protein variants were separated by SDS–PAGE (15%) and stained via Coomassie brilliant blue. Positions of molecular mass standards (kDa) are indicated. HIV-1 CA is full-length from residues 133–363 (using Gag-numbering), CA-NTD includes residues 133–278 and CA-CTD includes residues 279–363. d) Negative Stain TEM micrographs of HIV-1 CA assemblies. Scale bars are annotated in the figure. The shown gels and negative stain micrographs are representative for at least 3 repetitions of the same experiment which yielded similar results.
Extended Data Fig. 5 Image processing and classification of HTLV-1 MA126CANC tubes.
a) Image processing workflow for the subtomogram averaging and multiparticle refinement of the CA-NTD layer from HTLV-1 MA126CANC tubes. Please see Materials and Methods for more details. b) Classification of tubes based on geometry. Different tubes have different geometries, with varying helical rise and pitch. For improved structure determination, tubes were grouped into classes according to their hexamer orientation (blue arrow) with respect to the tubes axis (black arrow). c) Histogram of abundance of different tube groups. d) Refined EM-density maps for the individual tube groups. Upper left corner – angle group; lower left corner – number of particles; lower right corner – resolution.
Extended Data Fig. 6 High-resolution cryo-ET structure of HTLV-1 MA126CANC tubes.
a) FSC for the CA-NTD from MA126CANC tubes, showing a resolution of 3.4 Å at the 0.143 criterion. b) Local resolution measurements for the cryo-EM density map of the CA layer of HTLV-1 MA126CANC tubes tubes. The CA-NTD and CA-CTD are annotated in the side view, highlighting again the weak density for the CA-CTD. The CA hexamer center is indicated with a small black hexagon. c) Model of HTLV-1 CA-NTD (cyan) refined into the 3.4 Å cryo-EM density map. The CA hexamer center is indicated with a small black hexagon. Below, representative EM densities for two helices of the CA-NTD are shown with the respective fit. At this resolution, the helical pitch and densities for larger side chains are visible. d) Low-pass-filtered isosurface representation of the subtomogram average of the C2-symmetric CA hexamer from HTLV-1 MA126CANC tubes, as seen in a side view (left) or from within the tube, showing the CA-CTD. This panel relates to the Fig. 2d, which shows the same map at the unfiltered 3.4 Å resolution. The orange dashed rectangle indicates the CA-CTD layer, which is not visible at high resolution in Fig. 2d. e) Distances between adjacent CA-NTD trimers and CA-CTD dimers.
Extended Data Fig. 7 Mutational analysis of the HTLV-1 CA-NTD interfaces.
Site-directed mutagenesis of amino acid residues at the CA-NTD interfaces were analyzed for their ability to produce particles relative to that of the WT. Transfection of 293 T/17 cells with plasmids expressing the HTLV-1 Gag (WT or alanine-scanning mutant) was done. Cell culture supernatants and cell lysates were harvested 48 h post-transfection. a) Gag expression in cells and particle production of L43A mutant. Immunoblot analysis was conducted to determine particle production relative to that of WT HTLV-1 Gag. Data was collected from three independent experiments. Error bars represent the standard error of the mean. Significance relative to WT was determined by using an unpaired two-sided t-test. ****, P < 0.0001. One representative set of immunoblots of the Gag protein detected from cell culture supernatants and from cell lysates by using an anti-HTLV p24 (CA) antibody is shown. The locations of the molecular weight markers (in kilodalton) are indicated along the left side of the immunoblots. b) Representative cryo-EM images showing the morphology of the L43A mutant particles. Scale bar = 100 nm. The shown examples are representative of at least 250 individual immature L43A mutant particles imaged. c) Gag expression in cells and particle production of the indicated alanine double mutants. Immunoblot analysis was conducted to determine particle production relative to that of WT HTLV-1 Gag. Data was collected from three independent experiments. Error bars represent the standard error of the mean. Significance relative to WT was determined by using an unpaired two-sided t-test. ****, P < 0.0001; ***, P < 0.001; **, P < 0.01. One representative set of immunoblots of the Gag protein detected from cell culture supernatants and from cell lysates by using an anti-HTLV p24 (CA) antibody is shown. The locations of the molecular weight markers (in kilodalton) are indicated along the left side of the immunoblots. Detailed statistical information is given in Supplementary Table 1.
Extended Data Fig. 8 Biophysical analysis of HTLV-1 and HIV-1 CA constructs.
Biophysical analysis of HTLV-1 (panels a-c) and HIV-1 (panels d-f) CA and their N-terminal and C-terminal domains. Measurements are always performed in two conditions, termed pre-assembly (left side of respective panel) and post-assembly (right side of respective panel). a,d) First derivative of the fluorescence ratio over a temperature ramp, measured by nanoDSF, for HTLV-1 (a) and HIV-1 (d) CA (green), CA-NTD (cyan), and CA-CTD (orange) in pre-assembly and post-assembly conditions. Both full-length HTLV-1 CA and HTLV-1 CA-NTD show a drop in Tm upon assembly. This drop in Tm might reflect the changed reactivity of CA and CA-NTD in assembly conditions. Buffer conditions selected for assembly are chosen to stimulate protein-protein interactions (for example rendering the proteins to be more assembly reactive). This could make them more dynamic and also flexible in order to allow forming such interactions. Virus assembly is not necessarily representing a stable end state, as viruses need to be metastable to eventually release the viral genome. A Tm drop might therefore reflect the propensity of the proteins to allow assembly and disassembly. b,e) Turbidity change as a function of temperature, measured by back-reflection, for HTLV-1 (b) and HIV-1 (e) CA, CA-NTD, and CA-CTD in pre-assembly and post-assembly conditions. Tm for all constructs as well as the turbidity inflection points are given in Supplementary Table 2. c,f) Cumulant radius over a temperature ramp, measured by DLS, for HTLV-1 (c) and HIV-1 (f) CA, CA-NTD, and CA-CTD in pre-assembly and post-assembly conditions. CA and CA-NTD show an increase in cumulant radius coinciding with the observed turbidity increase and respective melting temperatures. CA-CTD shows no significant cumulant radius increase across all conditions. The color code is indicated on the bottom of the figure. The shown data is from one biological replicate measured three times, and representative for all three replicates performed at least three times. Data for the fluorescence ratio and turbidity measurements are presented as mean values +/- standard deviation and for the cumulant radius as mean values with smoothed curve using a Savitzky-Golay filter with a 5-point window and 0-order polynomial. Source data is provided alongside the manuscript.
Extended Data Fig. 9 Phylogenetic tree of retrovirus CA.
Phylogenetic tree derived from 16 different retrovirus CA sequences, from 5 different retrovirus genera. Multiple sequence alignment (MSA) was performed using Clustal Omega (https://www.ebi.ac.uk/jdispatcher/msa/clustalo) and the phylogenetic tree was generated using the webserver implementation of iTOL (https://itol.embl.de/). The UniProt accession IDs and the CA residues are indicated in the FASTA header on the right side. The boxed numbers indicate the branch length. EIAVY – Equine infectious anemia virus, FIVPE – Feline immunodeficiency virus, HV2 – HIV2, HV1 – HIV1, ALVA – Avian leukosis virus, RSVP – Rous sarcoma Virus, MMTV – Mouse mammary tumor virus, JSRV - Jaagsiekte sheep retrovirus, MPMV – Mason-Pfizer monkey virus, KORV – Koalaretrovirus, FLV – Feline leukemia virus, MLVMS – Murine leukemia virus, BLVAU – Bovine leukemia virus, HTL1A – HTLV1, HTLV2, HTL32 – HTLV3.
Supplementary information
Supplementary Information
Supplementary Fig. 1 and Tables 1 and 2.
Supplementary Video 1
Cryo-electron tomogram of HTLV-1 Gag-based VLPs.
Supplementary Video 2
Tour through the immature HTLV-1 CA lattice, annotating the CA-NTD and CA-CTD interfaces as described in Fig. 1.
Supplementary Video 3
Cryo-electron tomogram of HTLV-1 MA126CANC tubes.
Source data
Source Data Extended Data Fig. 4
Uncropped, nonannotated gels.
Source Data Extended Data Fig. 7
Uncropped, nonannotated blots.
Source Data Extended Data Fig. 8
Source data for nanoDSF analysis, including biological replicates.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Obr, M., Percipalle, M., Chernikova, D. et al. Distinct stabilization of the human T cell leukemia virus type 1 immature Gag lattice. Nat Struct Mol Biol (2024). https://doi.org/10.1038/s41594-024-01390-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41594-024-01390-8
