Small Capsid Protein pORF65 Is Essential for Assembly of Kaposi's Sarcoma-Associated Herpesvirus Capsids

Abstract

Kaposi's sarcoma-associated herpesvirus (KSHV) is the etiologic agent for KS tumors, multicentric Castleman's disease, and primary effusion lymphomas. Like other herpesvirus capsids, the KSHV capsid is an icosahedral structure composed of six proteins. The capsid shell is made up of the major capsid protein, two triplex proteins, and the small capsid protein. The scaffold protein and the protease occupy the internal space. The assembly of KSHV capsids is thought to occur in a manner similar to that determined for herpes simplex virus type 1 (HSV-1). Our goal was to assemble KSHV capsids in insect cells using the baculovirus expression vector system. Six KSHV capsid open reading frames were cloned and the proteins expressed in Sf9 cells: pORF25 (major capsid protein), pORF62 (triplex 1), pORF26 (triplex 2), pORF17 (protease), pORF17.5 (scaffold protein), and also pORF65 (small capsid protein). When insect cells were coinfected with these baculoviruses, angular capsids that contained internal core structures were readily observed by conventional electron microscopy of the infected cells. Capsids were also readily isolated from infected cells by using rate velocity sedimentation. With immuno-electron microscopy methods, these capsids were seen to be reactive to antisera to pORF65 as well as to KSHV-positive human sera, indicating the correct conformation of pORF65 in these capsids. When either virus expressing the triplex proteins was omitted from the coinfection, capsids did not assemble; similar to observations made in HSV-1-infected cells. If the virus expressing the scaffold protein was excluded, large open shells that did not attain icosahedral structure were seen in the nuclei of infected cells. The presence of pORF65 was required for capsid assembly, in that capsids did not form if this protein was absent as judged by both by ultrastructural analysis of infected cells and rate velocity sedimentation experiments. Thus, a novel outcome of this study is the finding that the small capsid protein of KSHV, like the major capsid and triplex proteins, is essential for capsid shell assembly.

Kaposi's sarcoma-associated herpesvirus (KSHV) is the causative agent of KS tumors, an angiogenic skin malignancy, as well primary effusion lymphomas and multicentric Castleman's disease (7, 9, 40). Malignancies due to KSHV are a major clinical problem for AIDS patients and consequently represent the predominant cancer found in southern Africa. KSHV (or human herpesvirus 8), like the other herpesviruses, assembles icosahedral capsids that display a T=16 shell. For all herpesviruses studied to date the capsids are comprised of six polypeptides (reviewed in references 17, 34, and 42). The capsid shell subunits or capsomeres are composed of the major capsid protein (51). There are 150 hexons of the major capsid protein and 11 pentons, which are located at the vertices of the icosahedral structure. The portal protein occupies the 12th vertex (4, 8, 11) and forms a dodecameric ring structure (48), and it is the portal for DNA entry into and probably exit from the capsid shell. A hetero-trimeric complex of two proteins called the triplex structure connects the capsomeres together and ensures stability of the shell structure (28, 47, 55). A small protein decorates the capsid shell by virtue of its interaction with the major capsid protein (2, 12, 50, 52, 56). The internal space is occupied by the scaffold protein and to a lesser extent by the maturational protease. During DNA packaging, proteolysis of the scaffold protein mediated by the maturational protease allows release of the scaffold proteins from the interior, which is subsequently occupied by the viral genome (reviewed in references 17, 34, and 42).

Most of the biochemical, structural, and genetic studies of herpesvirus capsid assembly have been done on herpes simplex virus (HSV) capsids. For HSV type 1 (HSV-1), four closed structures have been detected in infected cells (14, 29, 36). The earliest structure detected is the spherical procapsid, which matures into an angular stable B capsid (26, 29, 36, 47). Both these capsids contain the internal scaffold protein, which is visualized as a circular internal core in ultrastructural analyses (1). Similar to bacteriophages, DNA is packaged into a preassembled capsid, the products of which are the C capsid, which contains the viral genome and which matures into an infectious particle. The other by-product is the A capsid, which is an empty shell and is thought to be a result of abortive DNA packaging events (reviewed in references 17, 34, and 42).

The assembly of KSHV capsids probably utilizes the same pathway as that shown for HSV-1. Thus, both the proteins involved and the structures that form are expected to be the same. Studies have elucidated the different capsid types formed in KSHV-infected cells and their polypeptide composition as well as the three-dimensional image reconstruction of KSHV capsids isolated from virions (25, 49, 53). Comparison of this gammaherpesvirus capsid with those of the alpha- (HSV-1) and beta- (cytomegalovirus) herpesviruses has revealed many similarities but also some key differences. One striking difference was in the shape of the hexon protrusions. The alphaherpesvirus capsid hexon has a “cog-wheel” shape due to the presence of the small capsid protein (VP26) on the tips of this structure. The hexon protrusions of KSHV and cytomegalovirus capsomeres are different from those of HSV-1, indicating that the small capsid protein in these viruses displays a different configuration in the shell (23, 49).

The goal of this study was to assemble KSHV capsids in insect cells by using recombinant baculoviruses. The rationale for this is that assembly in insect cells is more amenable to manipulation. The replication properties of KSHV make it especially hard to isolate capsids from infected cells in sufficient quantities for biochemical analysis. At present KSHV capsids are isolated from virions that are excreted into the supernatant of primary effusion lymphoma cell lines that have been induced to activate the lytic cycle of replication (25, 53). This induction occurs at a very low frequency, and thus large numbers of cells are needed to isolate significant quantities of capsids. This together with the lack of a system to manipulate this virus genetically makes it an especially intractable problem, and thus extremely difficult to study an essential process such as capsid assembly in KSHV-infected cells.

The open reading frames encoding the KSHV capsid proteins and their homologs in HSV-1 are shown in Table 1. Of the capsid shell proteins, the major capsid protein is encoded by ORF25, the two triplex proteins are encoded by ORF62 and ORF26, and the small capsid protein is encoded by ORF65. The internal proteins, the protease, and the scaffold are encoded by the overlapping genes ORF17 and ORF17.5, respectively. Using the baculovirus expression vector (BEV) system we cloned and expressed in Sf9 cells all six capsid proteins. Using coinfection methods and ultrastructural analysis of infected cells, we readily observed icosahedral capsids in the nucleus. Capsids could also be isolated and purified by rate velocity sedimentation. This BEV self-assembly system can now be used for the analysis of KSHV capsid assembly at the biochemical, genetic, and structural levels because it is easier to manipulate than KSHV-infected cells. Using electron microscopy (EM) and sedimentation methods the requirement of the small capsid protein for assembly of KSHV capsids in the BEV system was demonstrated visually. This is a new and an unexpected finding for gammaherpesvirus assembly and could potentially reveal the stabilizing functions of this small but essential protein.

TABLE 1.

KSHV capsid protein ORFs

KSHV ORF	KSHV capsid protein	Predicted mass (kDa)	HSV-1 homolog
ORF25	MCP	153	VP5
ORF62	Triplex 1	36	VP19C
ORF26	Triplex 2	34	VP23
ORF17.5	Scaffold	31	p22a
ORF17	Protease	58	VP24/p21
ORF65	SCP	18	VP26

MATERIALS AND METHODS

Cell lines and antibodies.

Spodoptera frugiperda (Sf9 and Sf21) cells were grown in Grace's insect cell medium, supplemented with 10% fetal calf serum (FCS; Gibco-Invitrogen). Sf9 and Sf21 cells were grown in spinner flasks at 27°C with constant agitation (85 rpm). Cells were seeded at 0.5 × 10⁶ cells/ml and harvested when the cell densities reached between 1.5 × 10⁶ and 2 × 10⁶ cells/ml. Cell viability was consistently in the 90 to 98% range. Antibodies to ORF65 were kindly provided by S.-J. Gao and Bala Chandran. KSHV-positive human sera were obtained from the AIDS Malignancy Bank (R. F. Ambinder, Principal Investigator). Secondary antibodies conjugated to 6-nm gold particles were purchased from Jackson ImmunoResearch Laboratories.

Plasmids.

For this study the open reading frames for ORF25, ORF62, ORF26, ORF17, ORF17.5, and ORF65 were amplified using the high-fidelity polymerase Pfu Ultra (Stratagene). The template for all amplifications was KSHV BAC36 DNA (54). The primers used for all amplifications are shown in Table 2, and the sequence of strain BC2 was used in primer design (37). The ORFs were cloned in the baculovirus transfer vector pFastBac 1 (pFB1; Invitrogen) (24). ORF25 was cloned as an EcoR1-Spe1 fragment, ORF62, ORF26, ORF17, and ORF17.5 were cloned as EcoR1-HindIII fragments, and ORF65 was cloned as a Spe1-HindIII fragment into pFB1. Confirmed plasmids were designated with the transfer plasmid abbreviation and gene name, for example, pFB1-ORF25. All PCR-generated genes were sequenced for authentic amplification.

TABLE 2.

KSHV ORF primer sequences

KSHV ORFa	Primer sequenceb
ORF25/F	GGAATTCACCATGGAGGCGACCTTGGAGCAAC
ORF25/R	GGACTAGTCTAATACACCACCTTGTTTC
ORF62/F	GGAATTCACCATGAAGGTGCAGGCTGAAAATG
ORF62/R	GGGAAGCTTTTACAGAAACACAGTCCAGGG
ORF26/F	GGAATTCACCATGGCACTCGACAAGAGTATAG
ORF26/R	GGGAAGCTTTTAGCGTGGGGAATACCAACAGG
ORF17/F	GGAATTCACCATGGCACAGGGCCTGTACGTC
ORF17.5/F	GGAATTCACCATGAACAGCTCTGGTCAAGAG
ORF17-17.5/R	GGGAAGCTTTTAACCCTCACTAAAGGGAAC
ORF65/F	GGACTAGTGACCATGTCCAACTTTAAGGTGAGA
ORF65/R	GGGAAGCTTCTATTTCTTTTTGCCAGAGGG

^aF, forward primer; R, reverse primer.

^bThe translation start sequence is shown in bold in the forward primers, and restriction enzyme sites are italicized.

Generation of recombinant baculoviruses using the BAC-to-BAC system.

The Escherichia coli cell line DH10BAC, which carries the baculovirus genome cloned into a bacterial artificial chromosome (BAC), was used to introduce each pFB:KSHV gene expression cassette using transposition methods according to the manufacturer's protocol (Invitrogen) (24). PCR analysis of the bacmid DNA was used to confirm the introduction of the KSHV gene into the baculovirus, and positive bacmid clones were transfected into low-passage Sf9 cells as described by Okoye et al. (30) to reconstitute infectious virus.

Baculovirus transfection.

Sf9 cells (9 × 10⁵ cells in 35-mm dishes) were transfected with bacmid DNA using Cellfectin reagent (Invitrogen) as previously described by Okoye et al. (30). Generally, two independent bacmid clones were transfected for each gene. The supernatant of the transfected Sf9 cells, which contained the recombinant virus, was harvested 72 h after transfection, clarified, and designated as the P1 stock.

Baculovirus amplification.

Viruses were amplified in low-passage Sf9 cells as described by Okoye et al. (30). P2 and P3 stocks were typically 10⁸ PFU/ml.

Radiolabeling and SDS-PAGE.

Sf9 cells (1 × 10⁶ per well of a 12-well tray) were infected with 50 μl of the P2 or P3 stocks. Radiolabeling of infected cells was performed at 72 h postinfection. Two hours before that, the cells were overlaid with methionine-free medium (Invitrogen). Cells were incubated with 50 μCi [³⁵S]methionine Express (Perkin-Elmer) in methionine-free medium for 2 h and then harvested for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis. SDS-PAGE analysis was performed as described by Person and Desai (32). In some experiments NuPage gels (Invitrogen) were used, and the manufacturer's protocol was followed for electrophoresis of proteins using morpholineethanesulfonic acid buffer (Invitrogen).

Sample preparation for TEM.

Samples were processed for conventional thin-section transmission electron microscopy (TEM) essentially as described by Huang et al. (18) and Perkins and McCaffery (31). Sf9 cells (1 × 10⁷) in 100-mm dishes were coinfected with 250 μl of each of the six viruses. The infected cells were harvested 68 h postinfection and processed for electron microscopy.

Negative staining of nonimmunolabeled capsids was performed by adsorbing purified capsids for 15 min to freshly ionized carbon and Formvar-coated nickel grids. Each grid was washed briefly by floating it on 5 drops of deionized water, stained on a drop of 2% uranyl acetate, partially blotted on filter paper, and air dried.

Immunolabeling, negative staining, and trehalose embedding of capsids for electron microscopy.

A lower-contrast method of negative staining was adapted from the methods of Harris and Scheffler (16) for sensitive detection of small (6-nm) gold on the capsids, which might normally be obscured by dense staining. Capsids were adsorbed for 15 to 20 min to freshly ionized Formvar- and carbon-coated nickel grids. Grids were floated on 8 consecutive drops of wash buffer (phosphate-buffered saline [PBS] containing 2.5% FCS and 10 mM glycine, pH 7.4) and blocked (PBS with 10% FCS) for 30 min. Each grid was floated on 10 μl of primary antiserum at the following concentrations: mouse anti-ORF65 (1:10), human normal serum (1:40), human KSHV-positive serum (1:40), or 10% FCS in PBS (negative control for nonspecific binding of the anti-mouse secondary antibody) overnight at 4°C. Each grid was then washed through 8 drops of wash buffer and incubated on 10 μl of 6-nm gold conjugated to secondary antibody (donkey anti-mouse, 1:20, or donkey anti-human, 1:40) for 2 h at room temperature. Grids were washed eight times through wash buffer and five times through deionized water, floated briefly on a solution of 1% uranyl acetate, 1% trehalose, partially blotted with filter paper, and air dried.

TEM.

Samples were examined using Phillips EM 410, EM 420, and Tecnai 12 transmission electron microscopes (FEI, Hillsboro, OR); images were captured with an SIS Megaview III or FEI Eagle 2k camera.

Sedimentation analysis.

Sf9 cells (1 × 10⁷ cells) in 100-mm dishes were infected with 250 μl of either the P2 or P3 stock of each virus. Two 100-mm dishes were used for each gradient. Sixty-eight hours after infection the cells were harvested in PBS, washed again with PBS, and lysed in 1 ml of 2× capsid lysis buffer (2 M NaCl, 10 mM Tris [pH 7.5], 2 mM EDTA, and 2% Triton X-100). The lysate was sonicated and layered on 20 to 50% sucrose gradients. Sucrose gradients were made using a BioComp Gradient Mate, and sedimentation was performed in a Beckman SW41 rotor at 39,000 rpm for 55 min. The light-scattering band was visualized and harvested by side puncture using an 18-gauge needle.

Data and figure preparation.

For figure preparation, autoradiographs were scanned at 300 dots per inch in Adobe Photoshop. Electron micrographs were captured as 12- or 16-bit images and exported as 16-bit .tiff files, linearly adjusted for brightness/contrast, and compiled using Canvas 9.0. Digital images of the sucrose gradients were obtained by John Letos (School of Medicine, Johns Hopkins University) using a Canon EOS camera and imported as .tiff files into Adobe Photoshop for figure compilation.

RESULTS

Cloning and expression of the ORFs encoding the KSHV capsid proteins.

Previously we used the BAC-to-BAC system from Gibco-Invitrogen for the introduction of HSV-1 genes encoding capsid proteins into the baculovirus genome (30). This method utilizes a baculovirus genome cloned in a BAC, propagated and manipulated in E. coli (24). The same method was used for expression of the KSHV capsid genes. All six KSHV capsid ORFs were amplified by PCR using the primers shown in Table 2 and cloned into the transfer vector pFB1. The baculovirus polyhedrin promoter is used to drive the expression of the foreign proteins. The genes were introduced into the baculovirus genome using recombineering methods in the host E. coli cell DH10BAC (24). Infectious baculovirus was reconstituted in insect cells following transfection of Sf9 cells. Expression of some proteins was evident by Coomassie-stained gels of infected cell lysates harvested 72 h after infection and confirmed by Western blot methods using specific antisera (data not shown). Because antibodies for all six proteins were not available, infected Sf9 cells were metabolically labeled with [³⁵S]methionine and analyzed by SDS-PAGE to demonstrate expression of all the capsid proteins. As shown in Fig. 1, radioactivity corresponding to the different capsid proteins was evident in the gel at the correct mobility. For both pORF17 and pORF17.5 a proteolytic processed polypeptide was observed during this short labeling period (2 h). This was not observed in another insect cell, Sf21 cells (Fig. 2). For example, compare lanes 1 and 4 for pORF17 and lanes 3 and 6 for pORF17.5. Therefore, this proteolysis appears to be host cell specific. This was also observed to a lesser extent with pORF25 and pORF26 (Fig. 1). The scaffold protein (pORF17.5) has a predicted mass of 31 kDa (Table 1), but in the gel the scaffold protein (Fig. 1, lane p17.5) had a much slower mobility than either of the triplex proteins (Fig. 1, lanes p62 and p26). This is probably due to posttranslational modifications, such as phosphorylation, which has been reported for other herpesvirus scaffold proteins (5, 15).

FIG. 1.

Expression of KSHV capsid proteins in insect cells. Sf9 cells were infected with recombinant baculoviruses expressing the six KSHV capsid proteins (designated by number only), a baculovirus (Bac) expressing a nonrelated HSV-1 gene, or mock infected (MI). The infected cells were labeled with [³⁵S]methionine 72 h after infection for a duration of 2 h. Lysates were analyzed by SDS-PAGE (15% acrylamide) followed by autoradiography. The KSHV capsid polypeptides in the autoradiograph are indicated by an asterisk on the right of the lanes. A black arrow indicates proteolytic products of pORF17 and pORF17.5. Molecular mass standards are in lane M and are 220, 97.4, 66, 46, 30, and 14.3 kDa.

FIG. 2.

Sf9 cell proteolytic cleavage of pORF17 and pORF17.5. Sf9 (lanes 1 to 3) and Sf21 (lanes 4 to 6) cells were infected with baculoviruses expressing pORF17 (lanes 1 and 4), pORF26 (lanes 2 and 5), and pORF17.5 (lanes 3 and 6). The cells were metabolically labeled with [³⁵S]methionine at 72 h after infection for a duration of 2 h. Cell extracts were analyzed by SDS-PAGE (15% acrylamide). The radioactive bands corresponding to the full-length proteins are indicated on the right with an asterisk, and the proteolytic cleavage products are shown by a black arrow.

Ultrastructural analysis of insect cells coinfected with baculoviruses expressing the KSHV capsid proteins.

Previous studies by Tatman et al. (44) and Thomsen et al. (46) showed that when insect cells were coinfected with baculoviruses expressing the HSV-1 capsid proteins, icosahedral capsids were detected as judged by conventional electron microscopy methods. Our aim was to determine if self- assembly of KSHV capsids could also occur in the BEV system. Sf9 cells were coinfected with baculoviruses expressing pORF25, pORF62, pORF26, pORF17, pORF17.5, and pORF65, and the infected cells were processed for conventional EM at 68 h postinfection. Spherical capsid structures were evident in the cells infected with all six viruses (Fig. 3A and B). When Sf9 cells were similarly infected but the virus expressing the major capsid protein (pORF25) was omitted (Fig. 3C), no assembled capsids were evident in the cells, indicating the importance of the major capsid protein for assembly. The only structures present were baculovirus particles.

FIG. 3.

Capsid assembly in insect cells using recombinant baculoviruses expressing KSHV capsid proteins. (A and B) Sf9 cells were coinfected with baculoviruses expressing pORF25, pORF62, pORF26, pORF17, pORF17.5, and pORF65 (A and B; BAC-ALL). (C and D) Other cells were similarly infected but the virus expressing pORF25 was omitted (-ORF25) (C) or mock infected (D). The cells were fixed and processed for conventional EM 68 h after infection. KHSV icosahedral capsids (marked by white arrows) were evident in the nuclei of the BAC-ALL infected cells (A and B). The black arrowheads in panels A to C indicate baculovirus particles. KSHV capsids were not evident in -ORF25-infected cells (C). Bar, 200 nm (B) or 1 μm (A, C, and D). The nuclear envelope (ne) and mitochondria (m) are indicated where visible.

Purification of KSHV capsids from insect cells.

Next, rate velocity sedimentation was used to isolate and purify capsids from coinfected insect cells. Sf9 cells were infected with all six bacuolviruses carrying the KSHV ORFs. Sixty-eight hours after infection the cells were harvested and lysates were prepared and sedimented through 20 to 50% sucrose gradients (Fig. 4A). A light-scattering band was visualized in the gradient at the position where HSV-1 B capsids sediment. Material from this band was harvested by side puncture and analyzed by EM following negative staining (Fig. 4B). Numerous icosahedral capsid structures were evident by this method. Similar capsid preparations purified from sucrose gradients as shown in Fig. 4A were concentrated by ultracentrifugation and the pellets resuspended in Laemmli sample buffer. The polypeptide composition of these capsids was analyzed using 4 to 12% Nu-PAGE (Invitrogen) gels followed by Coomassie staining (Fig. 4C). The five KSHV capsid proteins detected in the gels are indicated on the right (see also Table 1). There is always small discrepancy for our molecular weight standards in Nu-PAGE gels. In addition to the capsid polypeptides, there are other bands present in this preparation which are probably baculovirus polypeptides or insect cell proteins that cosediment at the same position. The predicted molecular weight of the scaffold protein is smaller than the two triplex proteins (Table 1); however, due to posttranslational modifications, the scaffold proteins present in angular capsids tend to migrate at a position between the two triplex proteins (25). The presence of the catalytic domain of the protease in these angular capsids could not be determined due to the lack of specific antisera; however, it has been detected in capsids isolated from virions (33).

FIG. 4.

Biochemical and structural analyses of KSHV capsids. Sf9 cells coinfected with all six baculoviruses expressing the KSHV capsid proteins were harvested 68 h after infection, and the lysates were prepared and sedimented on sucrose gradients. A light-scattering band visualized by reflected light that migrated just below the half-way mark was evident in the digital picture (A). The direction of sedimentation is indicated by the arrow. The capsids harvested from the gradient were examined by negative staining followed by EM (B). Bars, 250 nm. Capsids were also concentrated and the polypeptide composition analyzed by 4 to 12% Nu-PAGE gel (C). Different amounts (20, 10, and 5 μl) of the capsid preparation were analyzed in the protein gel. The polypeptide bands corresponding to the five abundant capsid proteins are indicated. Molecular mass standards are shown in the left lane and correspond to 150, 100, 75, 50, 37, 25, 20, 15, and 10 kDa.

Immuno-EM detection of the small capsid protein pORF65.

Because pORF65 binds to the outer surface of the capsid shell (23), we used immuno-EM methods to detect this protein in the capsids isolated from insect cells. KSHV capsids were purified from coinfected insect cells as described above and used in immuno-EM analysis for pORF65 reactivity. A monoclonal antibody to pORF65 was highly reactive with the capsids as judged by the large numbers of specifically bound gold particles (Fig. 5B). This decoration of the capsids with gold particles was also observed by Nealon et al. (25) on their capsids isolated from purified virions using similar antisera. There were few or no gold particles bound to these capsids when a control serum was used (Fig. 5A). Because pORF65 is used for serology of KSHV (39, 45) we also used human sera obtained from both a KSHV-positive and a control (KSHV-negative) individual and performed a similar analysis. The KSHV-positive antiserum was tested in Western blot assays to confirm detection of pORF65 produced in insect cells (data not shown). This serum did not react with the other capsid shell proteins (pORF25, pORF62, or pORF26) in similar assays. Numerous gold particles were observed bound to the capsids when KSHV-positive serum (Fig. 5D) was used, indicating that this serum recognized pORF65 in these structures. There was nonspecific reactivity of the control human serum (Fig. 5C) to these capsids, as judged by the few gold particles associated with capsid shell.

FIG. 5.

Immuno-EM detection of pORF65. Sf9 cells were infected with all six viruses carrying the KSHV capsid ORFs; the cells were harvested 68 h postinfection, and extracts were prepared and layered onto sucrose gradients. The capsid band was harvested by side puncture, and the capsids were incubated with fetal calf serum (A), a monoclonal antibody to pORF65 (B), KSHV-positive human sera (D), and normal human sera (C) followed by 6-nm gold conjugated to donkey anti-mouse or goat anti-human secondary antibody. Capsids were embedded in 1% trehalose solution and negatively stained with 1% uranyl acetate to allow visualization of small (6-nm) gold particles. Bars, 200 nm.

Role of the individual KSHV capsid proteins in self-assembly.

Experiments were carried out to determine the contribution of each individual KSHV capsid protein for capsid assembly. Insect cells were infected with all six recombinant baculoviruses expressing the KSHV ORFs. Similar infections were also carried out except that the viruses expressing the protease (-ORF17), the scaffold protein (-ORF17.5), or both were omitted from the infections. Also, in other infections the viruses expressing either pORF25 (-ORF25), pORF62 (-ORF62), pORF26 (-ORF26), or small capsid protein (-ORF65) were omitted from the infection. Cells were processed for TEM 68 h after infection, and the results are shown in Fig. 6. Capsids were detected in cells infected with all six viruses (Fig. 6A) but not in -ORF25 (Fig. 6E), -ORF62 (Fig. 6H), or -ORF26 (Fig. 6F) infected cells, indicating the importance of the triplex proteins and the major capsid protein for shell assembly. In the absence of protease, closed structures were evident (Fig. 6B). In the absence of the scaffold protein many open shell structures were observed in the cells (Fig. 6C), indicating the importance of this protein for icosahedral capsid formation. The open shell structures were also evident in cells in which both the protease- and scaffold protein-expressing viruses were not included (Fig. 6D). From the data shown in Fig. 6C it appears that the KSHV protease cannot act as a scaffold for icosahedral capsid assembly in this system. Interestingly, capsids were not detected in cells in which the virus expressing the small capsid protein was left out (Fig. 6G). These results demonstrate that the small capsid protein is required for assembly of KSHV capsids.

FIG. 6.

Role of individual proteins for capsid shell and icosahedral capsid assembly. Sf9 cells were infected as described in the legend of Fig. 3 and processed for conventional EM. Examples of KSHV icosahedral capsids (indicated by white arrows) are shown for cells infected with the viruses expressing all six proteins (A) as well as for those for which the virus expressing pORF17 was not added (B). Open capsid shells that did not attain closure are indicated with white arrowheads in sections of infected cells where the virus expressing the scaffold protein (pORF17.5) was omitted (C) and also in cells where both the protease and scaffold protein were not included (D). Capsids were not detected in sections examined for cells in which the virus expressing pORF25 (E), pORF26 (F), pORF65 (G), or pORF62 (H) was omitted. Bars, 500 nm (A, F, and G), 1 μm (B, C, and E), or 2 um (D and H). Baculovirus particles are indicated by black arrowheads. The nuclear envelope (ne) and plasma membrane (pm) are marked.

pORF65 is required for KSHV capsid assembly.

To further demonstrate the important role of the small capsid protein for self-assembly, rate velocity sedimentation of infected cell lysates was performed. Sf9 cells were infected with all six baculoviruses, and similar infections were done but the virus expressing the small capsid protein was omitted (-ORF65). The infected cells were harvested 68 h postinfection, and cell lysates prepared from these cultures were sedimented through sucrose gradients (20 to 50%). In the BAC-ALL gradient, a light-scattering band was detected half-way in the gradient (Fig. 7). When material harvested from this band was examined by EM, icosahedral capsids were detected (data not shown). HSV-1-infected cell lysates were similarly sedimented to show the positions of A, B, and C capsids, of which the first two were marked (Fig. 7). Relative to KSHV capsids, HSV-1 B capsids sediment a bit further down the gradient. In the -ORF65 gradient there was no light-scattering band, a visual demonstration that pORF65 is essential for KSHV capsid assembly.

FIG. 7.

KSHV small capsid protein is required for capsid assembly. Sf9 cells were infected with all six viruses expressing the KSHV capsid proteins (BAC-ALL). In similar infections, the virus expressing pORF65 was excluded (-ORF65). Sixty-eight hours after infection the cells were lysed and the extracts layered on 20 to 50% sucrose gradients and sedimented. Reflected light was used to visualize the light-scattering bands, and digital pictures were taken. HSV-1-infected cell extracts were similarly sedimented, and the two visible capsid types (A and B capsids) are indicated. The arrow indicates the direction of sedimentation.

DISCUSSION

Herpesviruses encode six proteins that come together to form a protective coat around the virus genome. Capsid shell proteins have an inherent ability to self-assemble into higher-order structures if expressed together. Protein-protein interactions, as well as protein-protein contacts, drive the self-assembly of these structures. This property is evident in the numerous reports of self-assembling structures, first of the bacteriophages and then with animal viruses (6, 21, 22, 38). Using BEV technology Homa et al. (46) and Rixon et al. (44) demonstrated self-assembly of HSV-1 capsids into icosahedral shells, both in insect cells, and subsequently Newcomb and Brown demonstrated capsid assembly using purified components isolated from Sf9 cells (27).

Our goal was to determine whether self-assembly of the gammaherpesvirus KSHV capsid could be accomplished in insect cells using BEV technology. All six proteins were expressed in insect cells using recombinant baculoviruses. Coinfection of insect cells with all six viruses gave assembled capsid structures in these cells as judged by ultrastructural and sedimentation methods. This is the first demonstration of self-assembly of a gammaherpesvirus capsid using the baculovirus expression method. For gammaherpesviruses and KSHV in particular, the lack of an “easy to use” cell culture system and the challenges of genetically manipulating the virus genome make it difficult to study virus assembly and morphogenesis in infected cells. Furthermore, the ability to easily purify large quantities of capsids should allow one to perform a more detailed biochemical analysis of KSHV capsids and hence to determine the genetic and structural similarities and differences between herpesvirus capsids and their assembly pathways.

The similarities between the KSHV and HSV-1 assembly pathways are several. Thus, both KSHV triplex proteins and major capsid protein are needed for shell construction, and the scaffold protein is required for icosahedral capsid assembly. In addition, the protease was not required for capsid assembly. The major difference between HSV-1 and KSHV was the essential role of the small capsid protein for KSHV capsid assembly. From this study the small capsid protein should be classified as essential for KSHV shell assembly. The phenotype seen in cells when pORF65 was absent (-pORF65) was the same as when the major capsid protein or the two triplex proteins were absent.

It is possible that pORF65 is required for nuclear translocation of the other capsid shell proteins, which could also explain the lack of capsids in its absence. However, for HSV-1, Rixon and colleagues (35) showed that VP26 localizes to the nucleus only under conditions when VP5 is also nuclear, that is, in the presence of scaffold protein or VP19C. We have demonstrated in infected cells that VP5 is required for nuclear concentration of a green fluorescent protein (GFP)-tagged VP26 polypeptide (12). To address the nuclear transport of KSHV pORF65, a baculovirus was made that expresses a GFP-tagged pORF65. The fluorescence in Sf9 cells infected with this virus was distributed throughout the cell. Coinfection of cells with this virus and baculoviruses expressing the other capsid proteins was performed, and it was observed that fluorescence relocalized to large nuclear puncta only in the presence of pORF25 and pORF17.5 (data not shown). Thus, the scaffold protein is the major determinant of nuclear transport of pORF25 and, consequently, pORF65 is similarly translocated into the nucleus by virtue of its association with the major capsid protein.

For HSV-1 in the baculovirus expression system (44, 46) VP26 was not required for capsid assembly. Furthermore, in an HSV-1 mutant virus in which the small capsid protein (VP26) was mutated, capsids assembled normally, were packaged with the virus genome, and matured into infectious virions (13). There was, however, a pronounced effect of the VP26 null mutation on the replication properties of this virus in the mouse central nervous system, in which infectious virus yields were reduced by at least 100-fold (13). Structurally, the absence of HSV-1 VP26 in the virion does not appear to affect the association of certain tegument proteins with the capsid shell (10). For human cytomegalovirus, the data indicate a role for the small capsid protein in a function postassembly, although this has yet to be formally proved (3). For KSHV it still remains to be determined whether the findings with the BEV system are similar to those in KSHV-infected cells.

The gammaherpesvirus small capsid proteins are the largest in terms of size relative to the alpha- or betaherpesvirus proteins (49). Although there is very little homology between these proteins, it is possible that gammaherpesviruses have evolved an additional stabilization function important for assembly. These small capsid proteins interact with the major capsid protein; in fact, their localization to nuclear sites of assembly is dependent on this interaction (12, 35). Both the HSV-1 and KSHV small capsid proteins are found on the hexon configuration of the major capsid protein and not on the penton conformation (23, 52, 56). Because of this difference it has been suggested that HSV-1 VP26 forms a hexamer (56), although in vitro-produced VP26 can also be found in a monomer or dimer conformation (52). Purified VP26 was also shown to be predominantly in a β-sheet (80%) conformation with some α-helical segments (15%) (52). Based on results from secondary structure prediction programs, the KSHV pORF65 appears to contain similar percentages of α-helical and β-sheet conformations with a single hydrophobic region in the N terminus (amino acids 45 to 65).

Capsid decoration proteins similar to the herpesvirus small capsid protein are found in many phage structures. These proteins, such as the Soc protein of phage T4 (19, 20, 43) or the gpD of phage lambda (41), are small in size and present in several copies on the shell. They act to stabilize the capsid structure, but in many cases they are not essential to make a capsid. Herpesviruses utilize the triplex protein complex to stabilize the shell, and this structure is unique and essential for capsid assembly for this group of viruses. The small capsid protein gene is conserved in all three herpesvirus families, indicating the importance of this decoration protein for assembly and subsequent morphogenesis. Whereas in HSV-1 this protein is not required to assemble an icoshedral capsid, the KSHV small capsid protein is, from the evidence presented here, important for capsid assembly. This small protein, which is an important immunological reporter for KSHV infection (39, 45), could potentially also serve as a new antiviral target.