Structural basis of membrane budding by the nuclear egress complex of herpesviruses

Abstract

During nuclear egress, herpesvirus capsids bud at the inner nuclear membrane forming perinuclear viral particles that subsequently fuse with the outer nuclear membrane, releasing capsids into the cytoplasm. This unusual budding process is mediated by the nuclear egress complex (NEC) composed of two conserved viral proteins, UL31 and UL34. Earlier, we discovered that the herpesvirus nuclear egress complex (NEC) could bud synthetic membranes in vitro without the help of other proteins by forming a coat-like hexagonal scaffold inside the budding membrane. To understand the structural basis of NEC-mediated membrane budding, we determined the crystal structures of the NEC from two herpesviruses. The hexagonal lattice observed in the NEC crystals recapitulates the honeycomb coats within the budded vesicles. Perturbation of the oligomeric interfaces through mutagenesis blocks budding in vitro confirming that NEC oligomerization into a honeycomb lattice drives budding. The structure represents the first atomic-level view of an oligomeric array formed by a membrane-deforming protein, making possible the dissection of its unique budding mechanism and the design of inhibitors to block it.

Keywords: UL31; UL34; herpesvirus; membrane budding; nuclear egress.

Figure 1. Crystal structures of NEC from HSV‐1 and PRV

1. UL31 and UL34 constructs from HSV‐1 or PRV used to obtain diffraction‐quality crystals are shown schematically next to the full‐length proteins. FL, full‐length protein; xtal, construct used for crystallization; NLS, nuclear localization signal; TM, transmembrane region.

2. HSV‐1 and PRV NEC crystal structures strongly resemble each other. UL31 is shown in slate and UL34 in pink.

Figure 2. Secondary structure assignment

The HSV‐1 and PRV sequences are shown in black and blue, respectively.

1. A, B

   UL31 sequences (A) and UL34 sequences (B). Unresolved residues are shown in gray and marked with dotted lines. Underlined residues are conserved in HSV‐1 and PRV. Secondary structure elements are shown as tubes for α‐helices and arrows for β‐sheets. Zn‐coordinating residues are boxed in black. Mutated residues are boxed in red, blue, or yellow, with red labeling mutants that show reduced budding, blue for no effect and yellow for an increase in budding.

2. C

   HSV‐1 UL31 and UL34 structures are shown separately. Structural elements are labeled and colored as in (A, B). The inlet shows the conserved UL31 Zn‐binding site with labeled coordinating residues.

Figure EV1. Secondary structure assignment in PRV UL31 and UL34

1. Topology diagram for the PRV UL31 and UL34.

2. PRV UL31 and UL34 structures are shown separately.

Data information: Colors are as in Fig 2C.

Figure EV2. Crystal structures of HSV‐1 and PRV UL31 and UL34

1. A, B

   Crystal structures of HSV‐1 (A) and PRV (B) UL31 and UL34, colored according to Fig 2A and B. The top of the NEC in the shown orientation represents the membrane‐distal end, whereas the bottom of the NEC represents the membrane‐proximal end.

Figure 3. UL31 binds to UL34 via two distinct interfaces

HSV‐1 UL31 is shown in slate and HSV‐1 UL34 in pink. Residues involved in interface 1 are colored deep teal and in interface 2 hot pink. A list of residues involved in the UL31/UL34 interactions can be found in Appendix Table S1. Similar interfaces can be observed in PRV NEC (Fig EV3).

Figure EV3. UL31 binds to UL34 using two distinct interfaces

PRV UL31 is colored in light blue and PRV UL34 in light pink. Residues involved in interface 1 are shown in deep teal, and residues involved in interface 2 are shown in hot pink.

Figure 4. The NEC forms hexameric lattices in the presence of membranes or at high concentrations

1. Hexameric lattice as observed by cryoEM (Bigalke et al, 2014). The diameter of the hexameric rings is ˜110 Å, while the spikes are ˜110 Å in length.

2. Hexameric lattice in the HSV‐1 NEC crystal. The lattice for NEC_CD is depicted. The diameter of each hexameric ring is 110 Å, while the length of the spikes is 78 Å. The difference in length can be accounted for by regions absent from the crystallization construct but present in the construct used in budding assays and cryoEM.

Figure 5. The two NCS mates in the HSV‐1 NEC crystal form two types of hexameric lattices

1. The hexameric contacts are largely the same in both NEC_AB and NEC_CD (Appendix Table S3), but inter‐hexameric contacts differ. Hexameric interfaces are colored green, trimeric interfaces yellow, and dimeric interfaces red and orange. The lattice is shifted by 10.5° in NEC_CD versus NEC_AB.

2. A detailed comparison of NEC_AB and NEC_CD and the oligomeric contacts. Color scheme is the same as in (A).

3. Previously described non‐functional mutations, shown in hot pink, are mapped onto NEC_CD. Mutations that map to the UL34 interior likely disrupt the structural stability of the protein. Mutations that map to the oligomeric interfaces probably interfere with proper lattice formation, which explains the non‐functional phenotype of these mutants.

4. Conserved residues in α‐herpesviruses are shown in red, and strictly conserved residues are shown in hot pink. Hexameric contact patches are outlined in yellow and inter‐hexameric patches in blue. Most conserved and surface‐exposed residues are located at the hexameric interface. A proposed conserved capsid‐binding site is located at the top of UL31 on the membrane‐distal side of the NEC.

Figure EV4. Crystal packing for HSV‐1 NEC

UL31 molecules are colored in shades of blue and UL34 molecules in shades of pink. The two different hexagonal lattices are stacked on top of each other, but because the individual NEC molecules are tilted toward the crystallographic c‐axis, the two different lattices can be accommodated. One asymmetric unit is shown in bold colors, visualizing that the only contacts mediating the NCS interface are provided by residues from helices α6 and α9 in UL31. The interactions are shown in detail next to the interface. The proposed disulfide bond is shown in yellow, hydrogen bonds in green, and salt bridges in firebrick. Other residues that are involved in hydrophobic interactions are not labeled, but the side chains are shown. The tail‐to‐tail interactions link the next layer (gray) to the lattice. These interactions are mediated by helix α1 from UL31. A detailed view is shown on the top right. Again, hydrogen bonds are colored green and the side chains of other residues involved in the interaction are shown.

Figure 6. Mutational analysis of hexameric lattice formation

1. Overview of mutations designed to perturb hexamer formation. Mutated residues that reduced budding are colored in firebrick while those that did not significantly affect budding are colored in blue.

2. Three mutants were designed to perturb the inter‐hexamer interface. Mutated residues that reduced budding are colored in firebrick.

3. Quantification of budding events. Budding efficiency is shown compared to wild‐type (wt). Mutants designed to interfere with hexamer formation are colored green while mutants designed to interfere with inter‐hexamer formation are colored orange. The reported values represent averages of the results of at least two individual experiments. Error bars represent the standard errors of measurement from at least two individual experiments, with a count of at least 75 GUVs per sample and experiment. The statistical analysis used is the Student's t‐test, indicating the significance compared to wt. *P‐value < 0.05, **P‐value < 0.005, ***P‐value < 0.0005, ****P‐value < 0.00005. DN budding data have been shown previously (Bigalke et al, 2014). DN, dominant‐negative non‐budding UL34 mutant containing D35A₃₄/E37A₃₄. DN/SUP is DN mutant that additionally contains mutation R222L₃₁ in UL31 and has a wt phenotype. Raw average values of all mutants are listed in Appendix Table S5.

Figure EV5. Cosedimentation assay with NEC mutants and multilamellar vesicles (MLVs)

Most mutants bind to membranes comparable to WT NEC220, except for L142E₃₁, which shows reduced binding. The reported values represent averages of the results of three individual experiments. Error bars represent the standard errors of measurement. The statistical analysis used is the Student's t‐test. ***P‐value < 0.005. The asterisks above L142E₃₁ represent the significance compared to WT. All other samples do not show significant changes in membrane binding compared to WT. Coloring corresponds to Fig 6C, with hexamer mutants in green and inter‐hexamer mutants in orange.

Figure 7. Model of NEC‐mediated budding

1. A–D

   The NEC is represented by rectangles with UL31 in blue and UL34 in pink. Upon membrane binding (B), individual NEC heterodimers assemble into hexameric rings. These represent the individual building blocks of the lattice. (C) Once the hexamers are linked to each other, a negative curvature is induced. (D) The NEC‐lattice forms a coat enabling budding independently of other factors. Flaws in the hexameric lattice are required to form a spherical object, but these have not yet been visualized.