Previous Article | Next Article ![]()
Journal of Virology, October 2005, p. 12999-13006, Vol. 79, No. 20
0022-538X/05/$08.00+0 doi:10.1128/JVI.79.20.12999-13006.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Department of Virology II, National Institute of Infectious Diseases, Gakuen 4-7-1, Musashi-Murayama, Tokyo 208-0011,1 Department of Molecular Virology, Research Institute for Microbial Diseases, Osaka University, Suita-shi, Osaka 565-0871, Japan,2 Karolinska Institute, Department of Biosciences, 141 57 Huddinge, Sweden,3 Department of Molecular and Cellular Biology, University of California, Davis, California 956164
Received 21 April 2005/ Accepted 20 July 2005
|
|
|---|
|
|
|---|
The expression of foreign proteins in baculovirus systems opens the prospect of studying HEV capsid assembly, since virus-like particles (VLPs) of pronounced spikes on the surface can be formed with the recombinant protein expressed with this system (11, 25). This VLP is capable of inducing systemic and mucosal immune responses in experimental animals (9). With an oral inoculation of 10 mg of recombinant HEV VLPs, cynomolgus monkeys can develop anti-HEV immunoglobulin M (IgM), IgG, and IgA responses and protect against HEV infection (10). All these data suggest that VLPs are a candidate HEV vaccine.
The VLPs produced from Tn5 cells appear as T=1 icosahedral particles, which are composed of 60 copies of truncated pORF2 (25). The protein contains two distinctive domains: the shell (S) domain forms the semiclosed icosahedral shell, while the protrusion (P) domain interacts with the neighboring proteins to form the protrusion. The projection of T=1 recombinant HEV VLPs appears as spikes decorated with spherical rings (25), which fits with the morphology obtained from negatively stained HEV native virions. The diameter of these VLPs, 27 nm, is less than that reported for partially purified native virions (16). However, VLPs retain the antigenicity of the native HEV virion by designated antigenic sites at the P domain and by the capsid connection at the S domain. The particles appear empty, with no significant RNA-like density inside. The N-terminal region of pORF2 is rich in positively charged amino acid residues and may interact with RNA molecules (21). Thus, the deletion of the N-terminal 111 amino acid (aa) residues and the insufficient volume of the central cavity may lead to the failure of RNA encapsidation (25).
Cell type dependence in the VLP formation of the recombinant capsid protein was observed when aa residues 112 to 660 of ORF2 were expressed with a recombinant baculovirus in two insect cell lines, Tn5 and Sf9. In Tn5 cells, two major bands, having molecular masses of 58 kDa (58K) and 53 kDa (53K), were found in the cell lysate, while a peptide in the VLPs comprising a 53K protein was found in the culture medium. The 53K protein has been designated as either the 50K or 54K protein in previous studies (9, 11). In Sf9 cells, an additional peptide with a size between that of 58K and that of 53K was found in the cell lysate. However, no VLP was recovered from the culture medium. In Tn5 cells, terminal sequencing revealed that 58K and 53K proteins have the same first 15 aa in the N terminus and that a posttranslation cleavage by cellular protease(s) occurred at the pORF2 C termini and converted 58K into 53K. An independent but similar observation was obtained when pORF2 of the Pakistani strain was expressed in Sf9 cells (17) where several immunoactive proteins were detected in the cell lysate, and a 53K protein was secreted into the culture medium, but no VLP was found. Further investigation of pORF2 expression in Sf9 and Tn5 cells may allow us to understand the mechanism underlying the subunit assembly and particle formation of the recombinant HEV capsid.
We analyzed particle formation with pORF2 containing a series of truncated deletions at the N- and/or C-terminal region. In both Sf9 and Tn5 cells, amino acid residues 126 to 601 appeared to form the pORF2 core structure and were capable of self-assembling into VLPs. These results indicated that the cell dependence on particle formation is due to the difference between Sf9 and Tn5 cells in the modification process of pORF2.
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. Oligonucleotides used in the construction of baculovirus recombinants
|
![]() View larger version (20K): [in a new window] |
FIG. 1. Genome organization of HEV and schematic diagram of recombinant baculovirus vectors. DNA fragments encoding N- and C-terminal aa-truncated ORF2 were prepared by PCR with the primers listed in Table 1 and were used to construct 14 recombinant baculoviruses. Full-length pORF2 consisted of 660 aa. The N- and C-terminal aa numbers of the truncated protein are indicated.
|
Terminal amino acid sequence analysis.
The VLPs were further purified by 5 to
30% sucrose gradient centrifugation at 35,000 rpm for 2 h in a Beckman SW50.1 rotor. The visible white band was harvested as described above, diluted with EX-CELL 405, and again centrifuged at 45,000 rpm for 2 h in a Beckman TLA55 rotor to precipitate the VLPs. N-terminal aa microsequencing was carried out using 100 pmol of the protein by Edman automated degradation on an Applied Biosystems model 477 protein sequencer, and C-terminal aa sequencing was performed by Applied Biosystems.
SDS-PAGE and Western blot analysis. Dispersed insect cells were incubated for 20 min at room temperature to allow the cells to attach to culture flasks in TC-100 (Sf9 cells) or EX-CELL 405 (Tn5 cells) medium. The culture medium was removed, and the cells were infected with the recombinant baculoviruses at a multiplicity of infection of 10. Virus adsorption was carried out for 1 h at room temperature, and then the cells were incubated at 26.5°C. The proteins in the cell lysate and in the culture medium were separated by 10% SDS-PAGE and stained with Coomassie blue. For Western blotting, the proteins in the SDS-PAGE gel were electrophoretically transferred onto a nitrocellulose membrane. The membrane was then blocked with 5% skim milk in 50 mM Tris-HCl (pH 7.4)-150 mM NaCl and reacted with a patient's serum from an acute phase. Human IgG antibody was detected by using alkaline phosphatase-conjugated goat anti-human immunoglobulin (1:1,000 dilution) (DAKO A/S, Copenhagen, Denmark). Nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate P-toluidine were used as coloring agents (Bio-Rad Laboratories).
Cryo-electron microscopy (cryo-EM) and image reconstruction.
A 3-µl drop of purified HEV VLP (
1 mg/ml) was applied onto holey carbon film. After extra solution was wiped away with filter paper, the grid was rapidly plunged into liquid ethane surrounded by liquid nitrogen. Thus embedded in a thin layer of vitrified ice, the specimen was then transferred via a Gatan 626 cryo-transfer system to a Philips CM120 microscope. The specimen was observed at liquid nitrogen temperature and photographed at a magnification of 45,000. Each area was photographed twice, with defocus levels of 1 µm and 3 µm, respectively. The electron dose of each exposure was less than 10 electrons/Å2. The selected electron micrographs were digitized with a Zeiss scanner at a step size of 14 µm, corresponding to 3.1 Å at the specimen. The images were reconstructed according to icosahedral symmetry with Fourier-Basel procedures (4, 28). Briefly, the particle orientation and center of each image were estimated with the EMPFT program, where the structure of Tn5-produced HEV VLP was used as the initial model (1). The first reconstruction was generated from selected images and used as a model to refine the orientation and center parameters. After itinerant runs of EMPFT, the parameters were stable and appeared unchanged from one EMPFT run to another. The final reconstruction was computed by combining 353 images at a resolution of 23 Å. The surface-rendering map was generated with the NAG Explorer program combined with custom-created modules.
Mass spectrometry. The mass spectrometry experiment was done with a Reflex III mass spectrometer from Bruker, equipped with gridless delayed extraction. The samples were mixed with an equal volume of a saturated solution of sinapinic acid (Sigma Chemical Co., St. Louis, MO) in 33% (vol/vol) acetonitrile and 0.1% (vol/vol) trifluoroacetic acid. On the target plate, a thin layer was prepared with a saturated solution of sinapinic acid in ethanol. A sample volume of 0.5 µl was applied to a thin layer of sinapinic acid and allowed to crystallize. Data were acquired in the linear instrument mode. Data were processed and evaluated by XMASS software from Bruker.
|
|
|---|
58K and
53K, were detected in both cell lysates. The 53K protein was released into cell culture medium and assembled into VLPs in Tn5 cells but not in Sf9 cells (11).
Analysis of the N- and C-terminal aa sequences of the VLPs revealed that the N terminus was at aa residue 112 and the C terminus ended at aa residue 608, indicating that the C-terminal 52 aa of ORF2 were deleted. The protein that forms VLPs contains 497 amino acids (112 to
608), and its molecular mass was about 53K. An N-terminal 111 aa- and C-terminal 52 aa-truncated construct, Ac[n111c52], was generated, and the protein was expressed in both Sf9 and Tn5 cells. As expected, a single 53K protein was found in both Sf9 and Tn5 cell lysates (Fig. 2, Ac[n111c52] lanes in Sf9 and Tn5). Interestingly, these 53K proteins were released into both culture media as VLPs, as observed by electron microscopy (Fig. 3). The particle appeared empty and homogenous in size. Therefore, C-terminal truncation to aa residue 608 is crucial for particle formation and release into Sf9 cells.
![]() View larger version (28K): [in a new window] |
FIG. 2. Western blot assay of truncated pORF2 expressed in Sf9 and Tn5 cells. Eight recombinant baculoviruses, Ac[n111], Ac[n111c52], Ac[n111c58], Ac[n111c59], Ac[n111c60], Ac[n111c64], Ac[n111c72], and Ac[c52], were used to infect the insect cells. Ten microliters of the culture medium (right column) and 5 µl of the cell lysate (left column) were separated by 10% SDS-PAGE, and HEV-specific proteins were detected by Western blot analysis using the serum of a patient with acute hepatitis E. M, molecular weight markers; W, wild-type baculovirus-infected cells.
|
![]() View larger version (211K): [in a new window] |
FIG. 3. EM of the HEV VLPs generated in Tn5 (A) and Sf9 (B) cells with recombinant baculovirus Ac[n111c52]. The VLPs were stained with 2% uranyl acetate. Bar, 100 nm.
|
53K, were observed in infected Tn5 and Sf9 cell lysates 5 days postinfection (p.i.). However, these two proteins were not detected in their culture media (Fig. 2, Ac[c52] lanes in Sf9 and Tn5). These results indicated that aa residues before 601 were essential to the formation of VLPs. VLPs produced in Sf9 and Tn5 cells possess the same configurations and structures. The morphology of the VLPs generated in Sf9 cells appeared to be similar to that generated in Tn5 cells, as observed in the negatively stained particles (Fig. 3). To investigate the structural properties of these two released VLPs, we performed cryo-electron microscopy and image processing using VLPs produced in Tn5 cells. The electron cryomicrographs showed that the particle projected as a spiky hollow sphere, indicating that no RNA-like density was packed inside the capsid (Fig. 4A). The image processing was done according to the icosahedral procedure. The rotational symmetry of 522 was applied to reconstruct the final three-dimensional structure. The reconstructed VLP displayed a T=1 surface lattice with protruding density located at each of 30 twofold axes (Fig. 4B). The VLP was composed of 60 copies of pORF2, and the protruding density consisted of dimeric, projecting domains from twofold-related peptides. The particle diameter was 270 Å, measured from the three-dimensional reconstruction. The protein shell was 85 Å thick at the twofold axes. A channel can be observed under each protruding density. The protruding density was about 43 Å high, and the twofold platform was 56 Å in the long axes (data not shown). The threefold-related dimers formed a regular triangle, and the dimer-dimer distance was 76 Å measured from center to center (Fig. 4B). Molecular interactions at the icosahedral threefold region appeared much stronger than those at the fivefold region. There was no significant difference in radial density distribution between Tn5- and Sf9-produced VLPs (Fig. 4C).
![]() View larger version (54K): [in a new window] |
FIG. 4. HEV VLP structures determined by cryo-electron microscopy and image reconstruction. (A) Cryo-electron micrograph of ice-embedded HEV VLPs produced from Tn5 and Sf9 cells. The bar corresponds to 100 nm. (B) Surface-shaded representation of HEV VLP structures viewed along icosahedral twofold axes. VLPs from both Tn5 (left panel) and Sf9 (right panel) cells were color coded according to the radius, as indicated in the scale bar. The adjacent protruding spikes remain at equal distances of 76Å (white lines). The asterisks mark the positions of three adjacent icosahedral fivefold axes. (C) Sequential radial density projections generated from the twofold-oriented density map at corresponding radii. The protein density appears as the light color, while the background density is black.
|
![]() View larger version (20K): [in a new window] |
FIG. 5. Mass spectra from purified HEV VLPs displayed as the counts of isotope abundance (a.i.) versus mass/charge values (m/z). HEV VLPs produced from Tn5 (top panel) and Sf9 (bottom panel) cells with recombinant baculovirus Ac[n111c52] gave consistent mass spectra in which the abundant elements show similar m/z values at 53,000, 107,000, and 160,000.
|
57K) were thought to be the primary translation products encoded by N-terminal 123, 124, 125, 126, and 130 aa-truncated ORF2. In Tn5 cells, a C-terminal 52-aa-deleted product, about 51K protein, was the major protein to be efficiently released into the culture medium, where VLP formation occurred in Ac[n123]-, Ac[n124]-, and Ac[n125]-infected Tn5 cells (data not shown). Although the 51K protein was released into the culture medium, no VLP formation occurred in Ac[n126]- or Ac[n130]-infected Tn5 cells. In contrast, the 51K protein was not released into the culture medium in infected Sf9 cells (Fig. 6). These results demonstrated that aa residues after 125 were essential to the formation of VLPs.
![]() View larger version (23K): [in a new window] |
FIG. 6. Expression of N-terminally truncated pORF2 in Sf9 and Tn5 cells infected with Ac[n123], Ac[n124], Ac[n125], Ac[n126], Ac[n130], and Ac[n125c59]. A Western blot assay was carried out as described in the legend to Fig. 2. Ac[n111] was included for the expression of the 58K and 53K proteins. M, molecular weight markers; W, wild-type baculovirus-infected cells.
|
|
|
|---|
We previously found that when an N-terminal 111 aa-truncated ORF2 protein was expressed in Tn5 and Sf9 cells, two major peptides, having molecular masses of 58 and 53 kDa, were generated in both cells, and only the 53-kDa protein generated in Tn5 cells was released into culture medium and self-assembled into VLPs (11). The 58K protein presented the primary translation product, and the 53K protein is a processing product from the 58K protein. In this study, we examined the difference between Tn5 and Sf9 cells in HEV ORF2 gene expression and found that when a recombinant baculovirus (Ac[n111c52]) harboring a construct of the C-terminal 52-aa deletion was used, no difference between Sf9 and Tn5 cells in protein translation and particle formation was found. The observation that Ac[n111] failed to produce VLPs in Sf9 cells raised a question about the posttranslation modification in insect cells. In Tn5 cells, the levels of protein expression by Ac[n111] and Ac[n111c52] appeared to be similar. Therefore, it is likely that the 58K protein was incorrectly processed in Sf9 cells, thus affecting VLP assembly.
In addition, when Sf9 insect cells were infected with Ac[n111], the expressed proteins were localized in the cytoplasm and observed as inclusion-like bodies (one to four structures per cell) by EM (25). In contrast, when Sf9 cells were infected with Ac[n111c52], there were no inclusion-like bodies (data not shown), and the expressed proteins were localized evenly in the cytoplasm. Concomitantly, expressed protein was poorly detected in the culture medium from Ac[n111]-infected Sf9 cells at 3 days p.i., whereas a large amount of the 53K protein was detected in the culture medium from Ac[n111c52]-infected Sf9 cells. These findings suggest that the C-terminal aa of ORF2 might affect the localization, and subsequently the release, of the capsid protein from the insect cells. However, we do not yet know whether the VLPs form before release in infected cells or after release in culture medium.
The presence of Leu601 in pORF2 is important for the formation of HEV VLPs. A protein with a longer (580 to 610) deletion of aa residues was aberrant in protein folding; this may reduce the ORF2 homo-oligomerization (24). The prediction of the secondary structure based on protein sequence suggests two ß-strand motifs in the region between aa 580 and 601 (580 to
589 and 593 to
601). The failure in the particle assembly with Ac[n111c60] is due to incomplete formation of this ß-strand motif. Although aa 111 to 601 and aa 111 to 602 formed VLPs, the yield of each of these was about 10 to 20% of the yields of aa 111 to 660 (data not shown). This is in contrast to the fact that the levels of protein expression inside the cells were similar in these constructs. This observation further confirmed that stability of the C-terminal ß-strand motif is essential for VLP assembly.
The N-terminal 111-aa-deletion was found to be essential for cellular membrane dissociation of pORF2 expressed in insect cells (17, 24). We extended the N-terminal deletion up to Val125 without altering the ability to form HEV VLPs (Fig. 6). The ORF2 protein exhibits two-domain folding (25), with a domain organization similar to those of the norovirus (NV) capsid protein (15) and the tomato bushy stunt virus capsid protein (14). The N-terminal aa residues 112 to 125 may be the arm region extending from the S domain into the particle interior. In NV, the N-terminal region appeared to serve as a switch controlling the S domain configuration in the assembly process (3). Removal of the first 20 amino acids did not affect NV-like particle self-assembly, but a longer deletion at the N-terminal region did (3). Thus, residues 112 to 125 are putatively located in the HEV virion interior and may regulate VLP assembly.
Tn5 and Sf9 are insect cell lines that are commonly used in recombinant protein expression. The Tn5 cell is becoming more and more popular because it yields higher quantities of tissue factor than Sf9. Under optimum conditions, Tn5 cells produce 28-fold more secreted soluble tissue factor than Sf9 cells on a per-cell basis (23). In this paper, we report the difference between Tn5 and Sf9 cells in a protein synthesis system. The ORF2 protein underwent posttranslation cleavage, which is crucial for HEV VLP assembly. Although the HEV virion assembly mechanism remains unclear, our data indicate that the region consisting of ORF2 residues 126 to 601 is the kernel element for the monomer-monomer interaction and thus initiates VLP assembly.
Recombinant HEV VLPs themselves can be candidates for parenteral as well as oral hepatitis E vaccines (9, 10), and these VLPs have potential as mucosal vaccine carrier vehicles for the presentation of foreign antigenic epitopes through oral administration (13). Furthermore, HEV VLPs can be a vector for gene delivery to mucosal tissue for the purposes of DNA vaccination and gene therapy (20). The results of the present study provide the basic tool to construct VLPs having novel functions.
The study was supported in part by Health and Labor Sciences Research Grants, including Research on Emerging and Re-emerging Infectious Diseases, Research on Hepatitis, Research on Human Genome, Tissue Engineering, and Research on Food Safety, from the Ministry of Health, Labor and Welfare, Japan. This work was sponsored by grants from the Swedish Research Council to R.H.C. and L.X. A grant from the National Science Council, Taiwan, supported the work of J.C.Y.W.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»