Functional analysis of nuclear localization signals in VP1-2 homologues from all herpesvirus subfamilies

Abstract

The herpes simplex virus (HSV) tegument protein VP1-2 contains an N-terminal nuclear localization signal (NLS) that is critical for capsid routing to the nuclear pore. Here we analyzed positionally conserved determinants in VP1-2 homologues from each of the alpha, beta, and gamma classes of human herpesviruses. The overall architectures of the VP1-2s were similar, with a conserved N-terminal ubiquitin-specific protease domain separated from an internal region by a linker that was quite poorly conserved in length and sequence. Within this linker region all herpesviruses contained a conserved, highly basic motif which nevertheless exhibited distinct class-specific features. The motif in HSV functioned as a monopartite NLS, while in varicella-zoster virus (VZV) activity required an adjacent basic section defining the motif as a bipartite NLS. Neither the beta- nor gammaherpesvirus VP1-2 motifs were identified by prediction algorithms, but they nevertheless functioned as efficient NLS motifs both in heterologous transfer assays and in HSV VP1-2. Furthermore, though with different efficiencies and with the exception of human herpesvirus 8 (HHV-8), these chimeric variants rescued the replication defect of an HSV mutant lacking its NLS motif. We demonstrate that the lysine at position 428 of HSV is critical for replication, with a single alanine substitution being sufficient to abrogate NLS function and virus growth. We conclude that the basic motifs of each of the VP1-2 proteins are likely to confer a similar function in capsid entry in the homologous setting and that while there is flexibility in the exact type of motif employed, specific individual residues are critical for function.

Importance: To successfully infect cells, all herpesviruses, along with many other viruses, e.g., HIV, hepatitis B virus, and influenza virus, must navigate through the cytoplasmic environment and dock with nuclear pores for transport of their genomes into the nucleus. However, we still have a limited understanding of the detailed mechanisms involved. Insight into these events is needed and could offer opportunities for therapeutic intervention. This work investigated the role of a specific determinant in the structural protein VP1-2 in herpesvirus entry. We examined this determinant in representative VP1-2s from all herpesvirus subfamilies, demonstrated NLS function, dissected key residues, and showed functional relevance in rescuing replication of the mutant blocked in capsid navigation to the pore. The results are important and strongly support our conclusions of the generality that these motifs are crucial for entry of all herpesviruses. They also facilitate future analysis on selective host interactions and possible routes to disrupt function.

FIG 1

Summary of broad features of organization of VP1-2s from representatives of the alpha-, beta-, and gammaherpesvirus groups. VZV, varicella-zoster virus; PRV, pseudorabies virus; BHV, bovine herpesvirus 1; HCMV, human cytomegalovirus; HHV-6, human herpesvirus 6; EBV, Epstein-Barr virus; CalHV3, callitrichine herpesvirus 3; HHV-8, human herpesvirus 8; HVS, herpesvirus saimiri. Organization and shading are as discussed in the text. Regions involved in pUL37, VP16, or pUL25 interaction are indicated based on studies in either HSV or PRV (4, 5, 12, 26–29).

FIG 2

Sequence analysis and NLS function of the basic regions in VP1-2 in distinct herpesvirus families. (a) Alignment of regions termed R1, encompassing the basic region within the linker region downstream of the USP domain. Basic residues K and R are indicated with a red background. Other color shading is related to chemical property to best highlight conserved and nonconserved residues. Actual numbering of the R1 regions in relation to the parent VP1-2s is listed in full in Materials and Methods. In the alphaherpesvirus alignment, the extents of regions R4 and R5 are indicated by the wavy lines. The minimal regions tested for the beta- and gammaherpesvirus groups to examine sufficiency for NLS activity are indicated by the filled circles. (b) Schematic indicating the test for NLS function, wherein R1 or R4 was inserted in frame into the N terminus of β-galactosidase. (c) Higher-magnification images of typical fields of the parental β-galactosidase and a variant containing the complete R1 from HSV. (d) Lower-magnification images of the various test constructs as indicated.

FIG 3

Comparison of the HSV and VZV NLS motifs. (a) Detailed sequence comparison of the alphaherpesvirus R4 regions, highlighting the strong conservation of the region but also indicating differences at proline 427 and lysine 428 (asterisks), which are changed to arginines in VZV. (b) NLS function of variants of the HSV or VZV R4 after insertion into β-galactosidase as discussed in the text. (c) Conversion of R332 of VZV to the K of HSV is sufficient to convert the motif into a strong NLS.

FIG 4

Comparison of the basic motif in gamma-2-herpesviruses. (a) Schematic representation of the HHV-8 VP1-2, with each vertical red line indicating a basic arginine or lysine in the region downstream of the HHV-8 USP domain. This is expanded below, where the boundaries of the various regions tested are indicated. The only positive region for NLS function, termed R8, is further expanded, showing the sequence comparison with other gamma-2-herpesvirus VP1-2s. (b) Comparison of HHV-8 R1, R5, R8, and R9 for NLS function in the context of β-galactosidase.

FIG 5

Insertion of the cellular NLS from myopodin in place of R4 and analysis of R5 in HSV. (a) Schematic representation of the test regions wherein a confirmed cellular monopartite NLS from the protein myopodin was inserted in place of the core R4 NLS from HSV. This variant R1 region was termed R1.HSV.Myo. Arrows indicate the extent of the residues swapped out, which included part of the P/T/S-rich linker, with the resulting overall R1 10 residues shorter. In another test construct, the whole of R5 from residue 448 to 475 was deleted in R1.HSV.Δ448. (b) Test of NLS activity in the context of β-galactosidase, showing efficient function of the myopodin chimeric region and that the entire R5 could be deleted without significant impact on NLS activity. The latter result is consistent with data shown here and previously (32) that in HSV, R4 is sufficient for NLS activity.

FIG 6

Analysis of NLS function in the context of VP1-2. (a) Schematic representation of the expression constructs for the N-terminal half of VP1-2 (NT6), the NT6 variant lacking 7 residues within R4 (NT6ΔNLS), and the candidate NLS motifs or swapped in place of the natural R1 region. (b) Localization was scored into one of three categories (categories 1, 2, and 3). Typical examples are shown at higher magnification. (c) Localization of the parental NT6 and NT6ΔNLS in comparison to versions containing the R1 regions from VZV, HCMV, EBV, HHV-8, and the R8 region from HHV-8 as discussed in the text. (d) Localization of NT6 variants as indicated, where K428 was changed to either an arginine or alanine (NT6.K428R and NT6.K428A), R5 was deleted NT6ΔR5, each basic residue was mutated to alanine (NT6.MutR5), and the core R4 region was replaced with a cellular NLS in the context of R1 (NT6.MyoNLS). Representative fields are shown for each variant as discussed in the text.

FIG 7

The NLS region is not required for VP1-2 interaction with pUL37. (a) NT6 variants as indicated were transfected with HA-tagged pUL37 and immunoprecipitated by virtue of the epitope tag (V5) on the NT6 construct. Controls in lanes 1, 2, and 5 are as discussed in the text. Each of the variants interacted with pUL37, with no discernible difference in efficiency (lanes 6 to 10). (b) As for panel a, with results indicating that none of the substitution or deletion variants affected VP1-2-pUL37 interaction (c.f. lane 4 with lanes 6 to 11).

FIG 8

Rescue of viable virus using VP1-2 with heterologous NLS regions. (a) Schematic showing the experimental protocol wherein purified genomic DNA from the mutant virus KOS.VP1-2ΔNLS was cotransfected with plasmid NT6 DNA containing the N-terminal region of VP1-2 encompassing the R1 regions as indicated. (b) RSC cells were transfected as described in Materials and Methods and plates stained 3 days later for CPE and plaque formation. Representative results from several independent analyses are shown. (c) Quantitative analysis by titration on RSC cells of the yield of viable virus from transfection assays, with results discussed in the text. The asterisk over NT6.Mut5R indicates that recovered viruses were wt revertants as discussed in the text.

FIG 9

Rescue and replication of viruses containing heterologous NLS motifs. Purified virus isolates with sequence-confirmed insertion of variant NLS motifs (as indicated in the key) were inoculated at low MOI (0.002 PFU/cell) into each of several lines (RSC, RSCUL36, RPE, and Vero), and total yields on infectious progeny were assayed at various times thereafter on RSCUL36 cells. While in RPE and RSC cells each of the viruses grew normally (RPE cells) or with at most a modest reduction in yield (RSC cells), in Vero cells growth of the chimeric viruses, in particular that containing the VZV R1 NLS region, was significantly reduced.

FIG 10

Pronounced defect in spread of viruses with a variant NLS motif in Vero cells. A comparison of plaque formation by K.VP1-2ΔNLS.R1VZV and the revertant virus containing the homologous HSV sequence is shown. Plaques were fixed and examined by immunoperoxidase staining as described in Materials and Methods. Extremely small clusters of infection in Vero cells for K.VP1-2ΔNLS.R1VZV are indicated by outlined arrowheads (panel iv). Scale bars equal 1 mm.