INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Viruses belonging to the family Poxviridae are large, DNA-containing viruses whose replication cycle takes place in the cytoplasm of infected cells (27). Vaccinia virus, a representative of the genus Orthopoxvirus and the best-studied member of the family, is the model system of choice to study the morphogenesis and transmission of the poxvirus particles. Late steps in the replication cycle of the virus involve the assembly of infectious intracellular mature virions (IMV) that remain in the cytosol and can be released mechanically by breaking the cells. Transport of virions to the cell surface involves the wrapping of IMV particles by intracellular vesicles derived from the trans-Golgi network (TGN) (36) to form double-membrane-bound viruses called intracellular enveloped virions (IEV) that are transported to the cell periphery, where the outer membrane fuses with the plasma membrane. Intracellular transport of virus, which can occur by the induction of actin polymerization or by another mechanism(s), is dependent on the acquisition of the envelope (4, 5). The enveloped form of the virus found in the extracellular space has one more membrane than IMV and may remain cell associated (cell-associated enveloped virus [CEV]) or may be released from the cell (extracellular enveloped virus [EEV]). Much research has been directed to the biochemical and functional characterization of the EEV envelope, since the envelope plays a crucial role in virus dissemination (29) as well as in the establishment of immunological protection (1, 2, 8, 9, 29, 42). In addition, the EEV envelope may participate in virus evasion of the immune response (41, 43).

To date, six vaccinia virus proteins have been reported to be present in the EEV envelope. Four of these proteins (A33R, A34R, A56R, and B5R) are glycoproteins, with most of the protein being extracellular (10, 11, 18, 24, 31, 39). Protein A36R is a type Ib transmembrane protein with a large cytosolic domain (33), and protein F13L is a peripheral membrane protein which associates with the membrane by a palmitic acid moiety (15, 16, 37). All of the proteins except A56R (the virus hemagglutinin) have roles in virus wrapping or in the induction of actin tails (4, 5, 12, 32-34, 46, 47).

Besides their function in contributing to IMV wrapping, IEV transport, and CEV or EEV infectivity, we hypothesized that at least some of the EEV envelope proteins have to interact with cellular structures to determine the cellular compartment for wrapping and to perform functions related to the transport and egress of enveloped virions. In an attempt to unravel the relevant cell biological features of these proteins, we have carried out their individual expression in cells and studied their intracellular fates in the absence of vaccinia virus infection.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. BHK-21 (ATCC CCL10) cells were grown in BHK medium G-MEM containing 5% fetal bovine serum, 3 g of tryptose phosphate broth/ml, and 0.01 M HEPES in a 5% CO₂ atmosphere at 37°C. Anti-A33R and anti-A34R rabbit polyclonal antisera were provided by M. Way (European Molecular Biology Laboratory). Anti-A36R mouse monoclonal antibody was made available by G. L. Smith (Oxford University). Monoclonal antibody B2D10, anti-A56R, was kindly provided by Y. Ichihashi. Rat monoclonal antibodies 19C2 (anti-B5R) and 15B6 (anti-F13L) were kindly made available by G. Hiller. Vaccinia viruses WB5R, W-B5R_SCR1-4, and W-B5Rc have been described previously (17, 23) and were kindly provided by S. Isaacs (University of Pennsylvania). The plasmids for the Semliki Forest virus (SFV) expression system were provided by H. Garoff (Karolinska Institute).

Plasmid construction. The coding sequence of the A33R, A36R, and A56R genes were amplified by PCR using vaccinia virus genomic DNA as the template and were inserted into plasmid pSFV1 (20). The A33R gene was amplified using the oligonucleotide primers A335'Bam2 (5'-GACATAAATAGGATCCATTACCATGATG-3') (the BamHI site is underlined) and A33R3'Sma (5'-CATTTATTAATGTACCCGGGTAAATATTAG-3') (the SmaI site is underlined). The PCR product was cut with BamHI and SmaI and inserted into plasmid pSFV-1 to generate pSVF-A33R. The A36R gene was similarly amplified using the oligonucleotide primers A36R5'Bam (5'-CGTATATTGAGGATCCAGAAATGATGC-3') (the BamHI site is underlined) and A36R3'Sma (5'-CTTCAATTTTATAACCCGGGAACTAATC-3') (the SmaI site is underlined). The PCR product was cut with BamHI and SmaI and inserted into plasmid pSFV-1 to generate pSVF-A36R. The A56R gene was amplified using the oligonucleotide primers HA5'Bam (5'-AAATCACTTTGGATCCTAATATGACACG-3') (the BamHI site is underlined) and HA3'Sma (5'-TTTTACTATCCCGGGATTTATGTAAG-3') (the SmaI site is underlined). The PCR product was cut with BamHI and SmaI and inserted into plasmid pSFV-1 to generate pSVF-A56R. The A34R gene was amplified using the oligonucleotide primers A34R5' (5'-TTGTAGGATCCTCAATGAAATCGCT-3') and A34R3' (5'-CGTACGGATCCGACTTATTATT-3') (the BamHI sites are underlined). The PCR product was cut with BamHI, inserted into plasmid pGAT to generate pGAT-A34R, and subsequently subcloned into the BamHI site of pSFV1 to generate pSVF-A34R. The B5R gene was amplified using the oligonucleotide primers B5R5' (5'-CTATTTCTAGACCCGGGAATAAAAA-3') (the XbaI and SmaI sites are underlined) and B5R3' (5'-TTAATTATGGTACCGGATTTAT-3') (the KpnI site is underlined). The PCR product was cut with XbaI and KpnI, inserted into plasmid pGEM7 to generate pGEM-B5R, and subsequently subcloned into the SmaI site of pSFV1 to generate pSVF-B5R. The F13L gene was excised from pUC19-F13L (7) and subcloned into the BamHI site of pSFV1(NruI) to generate pSVF-F13L.

SFV expression. For each construct, 3 µg of linearized recombinant pSFV plasmid was used as a template for in vitro transcription with SP6 RNA polymerase (Pharmacia Biotech). For in vivo packaging of recombinant RNA into SFV particles, in vitro-transcribed RNA was electroporated into BHK-21 cells together with SFV helper RNA (3, 21). After 24 h, SFV particles in the culture medium were collected and frozen rapidly to be stored as a virus stock. The titers of stocks were determined by infecting cells in coverslips with serial dilutions of the stocks followed by indirect immunofluorescence assay for the expressed proteins.

Western blotting. To obtain cell extracts, monolayers were washed with phosphate-buffered saline (PBS) and incubated for 20 min on ice with lysis buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 2 mM EDTA, 1% NP-40, 1 µg of leupeptin/ml, 1 mM phenylmethylsulfonyl fluoride). The cells were scraped and recovered. Proteins were resolved by electrophoresis in sodium dodecyl sulfate-12% polyacrylamide gels and transferred onto nitrocellulose membranes by electroblotting (17 V for 1 h and 30 min at room temperature). After transfer, the membranes were blocked in PBS buffer containing 5% nonfat dry milk and then incubated overnight at 4°C with primary antibody diluted in PBS containing 1% nonfat dry milk, 0.05% Tween 20 (1:500 for anti-A33R, 1:100 for anti-A34R, 1:100 for anti-A36R, 1:100 for anti-A56R, 1:25 for anti-B5R, and 1:50 for anti-F13L). After being washed extensively with PBS-0.05% Tween 20, the membranes were incubated for 1 h at room temperature with horseradish peroxidase-conjugated secondary antibody diluted 1:3,000 in PBS-0.05% Tween 20 (sheep anti-rat immunoglobulin [Ig], sheep anti-mouse Ig, or sheep anti-rabbit Ig [Amersham]). After being washed, the membranes were incubated for 1 min with a 1:1 mix of solution A (2.5 mM luminol [Sigma]-0.4 mM p-coumaric acid [Sigma]-100 mM Tris HCl [pH 8.5]) and solution B (0.018% H₂O₂-100 mM Tris HCl [pH 8.5]) and exposed to an autoradiographic film.

Immunofluorescence microscopy. Cells grown on round coverslips were infected with vaccinia virus at a multiplicity of infection of 10 PFU per cell or with 200 µl of supernatants containing SFV particles. After 6.5 h of infection, the cells were incubated with 10 µg of cycloheximide/ml for 30 min. The cells were washed twice with PBS, fixed for 15 min at room temperature with cold 4% paraformaldehyde, and permeabilized by incubation for 15 min with PBS-0.1% Triton X-100. After incubation with PBS-glycine (0.1 M), the coverslips were incubated with primary antibody diluted in PBS-20% horse serum (1:100 for anti-A33R, 1:50 for anti-A34R, 1:100 for anti-A36R, 1:100 for anti-A56R, 1:100 for anti-B5R, and 1:50 for anti-F13L), washed, and incubated with secondary antibody diluted in PBS-20% horse serum (1:200 rabbit anti-mouse Ig-tetramethyl rhodamine isocyanate [TRITC] or 1:200 swine anti-rabbit Ig-TRITC [Dako, Glostrup, Denmark]). Some preparations were also incubated with 0.02 mg of TRITC-conjugated Triticum vulgaris lectin/ml together with the secondary antibody (Sigma). Both incubations were carried out at room temperature for 30 min. Finally, the coverslips were washed, mounted with FluorSave reagent (Calbiochem), and observed by fluorescence microscopy.

Electron microscopy. For electron microscopy, electroporation of RNA was used in order to obtain a high percentage of transfected cells (20). BHK-21 cells grown in 83-cm² flasks to 80% confluence were transfected by electroporation with RNA transcribed in vitro from the pSFV constructs. Parallel transfections with pSFV-GFP were carried out to monitor the transfection efficiency, which was close to 100%. After 8 h, the cells were fixed with 4% paraformaldehyde and 0.1% glutaraldehyde in PBS for 15 min at room temperature. After being washed twice with PBS, the cells were maintained in 4% paraformaldehyde until they were sectioned. For this, the cell pellets were immersed in 2 M sucrose for 30 min and then cryosectioned by the Tokuyasu procedure. The sections were incubated with the primary antibodies followed by protein A-gold and then dried using a mixture of methyl cellulose and uranyl acetate (see reference 14 for details).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Localization of EEV envelope proteins during vaccinia virus infection. Models for EEV formation and release imply that EEV envelope proteins must be present in the wrapping membranes, derived from the TGN. We first compared the distribution of the six known EEV envelope proteins in vaccinia virus-infected cells at late infection times (Fig. 1). The presence of the proteins at the cell surface was revealed in nonpermeabilized cells, which were compared with parallel immunofluorescences on Triton X-100-permeabilized cells. In permeabilized cells, proteins A36R, B5R, and F13L produced the typical EEV envelope pattern, with a strong juxtanuclear signal and peripheral staining of virions. However, proteins A33R, A34R, and A56R significantly deviated from this pattern. Proteins A33R and A34R were concentrated in a central area that was clearly more extended than the Golgi area and also in numerous small membrane structures. These structures were smaller and more abundant in the case of anti-A33R staining. Protein A56R was present both in the juxtanuclear area and in the plasma membrane but was not apparent in viruslike structures. Therefore, the EEV envelope proteins A36R, B5R, and F13L have the typical EEV envelope pattern, whereas proteins A33R, A34R, and A56R have a different distribution, which is partially coincident with the pattern of the former proteins.

View larger version (63K): [in this window] [in a new window]   FIG. 1.   Distribution of vaccinia virus EEV envelope proteins. Immunofluorescence staining was performed on nonpermeabilized cells (NP) and Triton X-100-permeabilized cells (P). Note the bright central staining in permeabilized cells, which corresponds to the Golgi complex area.

Expression of EEV envelope proteins in the SFV system. We wished to study the cell biological features of the six EEV envelope proteins, in particular with reference to the intracellular targeting signals that could direct these proteins to the TGN. Thus, the genes for the six known EEV envelope proteins were cloned and expressed by using an SFV replicon. Figure 2 shows the detection of the protein products by immunoblotting with specific antibodies. In all cases, SFV-expressed proteins were similar in size to the native proteins. In the case of protein A36R, two smaller proteins were detected which were absent in vaccinia virus-infected cells. These probably represent proteolytic products of the full-size protein.

View larger version (37K): [in this window] [in a new window]   FIG. 2.   Western blots of proteins expressed using SFV replicons. For each protein, Western blotting was carried out on cell extracts from mock-infected cells (lanes A), cells infected with vaccinia virus WR (lanes B and F) or vaccinia virus deficient in the corresponding protein (lanes C and G), control SFV expressing GFP (lanes D and H) or SFV expressing the corresponding protein (lanes E and I). Lanes A to E correspond to extracts prepared at 7 h postinfection, and lanes F to I correspond to extracts prepared at 24 h postinfection. Protein molecular weight markers are shown at the left of each panel. The arrow at the right of each panel indicates the position of the full-size protein.

A33R protein. As noted above, the distribution of A33R in vaccinia virus-infected cells was different from that of the other EEV envelope proteins. In permeabilized cells, a strong juxtanuclear staining, significantly larger than the Golgi area (as defined by the remaining EEV proteins or by wheat germ agglutinin [WGA] staining) was prominent (Fig. 3). In addition, a fine punctate pattern was visible in nonpermeabilized cells. When expressed in the absence of vaccinia virus infection (using the SFV-A33R construct), the protein produced surface staining similar to that seen in vaccinia virus-infected cells. In contrast, the level of intracellular labeling was relatively low and clearly distinguishable from that of vaccinia virus-infected cells.

View larger version (93K): [in this window] [in a new window]   FIG. 3.   Localization of A33R. Immunofluorescence staining was carried out on BHK-21 cells fixed at 7 h postinfection. Both nonpermeabilized (NP) and Triton X-100-permeabilized (P) cells are shown. The cells were either mock infected (M), infected with vaccinia virus (V), or infected with SFV-A33R particles (S). Note the fine punctate surface staining in cells infected with SFV-A33R.

View larger version (117K): [in this window] [in a new window]   FIG. 4.   Localization of A33R expressed from SFV-A33R by immunogold staining. Note the high concentration of label on the plasma membrane (arrowheads). Details of membrane labeling of a microvilluslike projection are shown in the inset. Bars, 100 nm.

A34R protein. In vaccinia virus-infected cells, labeling with anti-A34R antibody revealed the characteristic pattern in which Golgi area staining, as well as a peripheral punctate staining corresponding to virions and smaller structures, was apparent (Fig. 5A). Expression of the protein in the absence of vaccinia virus infection produced a completely different immunofluorescence pattern. The protein was retained intracellularly, was enriched in the perinuclear area, and was not transported in significant amounts to the cell surface. To ascertain whether the perinuclear staining corresponded to the Golgi area, double labeling with anti-A34R antibody and WGA was carried out (Fig. 5B). The immunofluorescence pattern showed that, while in vaccinia virus-infected cells the strong juxtanuclear staining overlapped well with the position of the Golgi complex, the perinuclear labeling pattern seen in SFV-A34R-infected cells did not coincide with the position of the Golgi complex, although they were in roughly the same area (Fig. 5B).

View larger version (65K): [in this window] [in a new window]   FIG. 5.   Localization of A34R. (A) Immunofluorescence staining was carried out on BHK-21 cells fixed at 7 h postinfection. Both nonpermeabilized (NP) and Triton X-100-permeabilized (P) cells are shown. The cells were either mock infected (M), infected with vaccinia virus (V), or infected with SFV-A34R particles (S). (B) Costaining with rhodamine-labeled WGA. Cells infected with vaccinia virus (V) or with SFV-A34R (S) were fixed and permeabilized and subjected to WGA and anti-A34R labeling. The same cells are shown on the left and right. Note the localization of A34R in the Golgi region in vaccinia virus-infected cells and the more extended localization in SFV-A34R-infected cells.

A36R protein. The A36R protein is a type Ib transmembrane protein, with most of the protein facing the cytosol. Since the epitope recognized by the antibody is on the cytoplasmic domain, immunofluorescence labeling with anti-A36R antibody was seen only after permeabilization of the cells (Fig. 6A). In vaccinia virus-infected cells, A36R staining was similar to that of other EEV envelope proteins. In contrast, immunofluorescence in cells infected with SFV-A36R showed a complex staining pattern. Staining was noted in an area in proximity to the nucleus. Also, plasma membrane staining was evident and was more pronounced at the tips of cell projections. In addition, fibrillar structures, typically running from side to side in the cell, were labeled. These structures were actin filaments, as demonstrated by double labeling with fluorescently labeled phalloidin (Fig. 6B).

View larger version (68K): [in this window] [in a new window]   FIG. 6.   Localization of A36R protein. (A) Immunofluorescence staining was carried out on BHK-21 cells fixed at 7 h postinfection. Both nonpermeabilized (NP) and Triton X-100-permeabilized (P) cells are shown. The cells were either mock infected (M), infected with vaccinia virus (V), or infected with SFV-A36R particles (S). (B) Costaining with rhodamine-labeled phalloidin to reveal actin fibers. Cells infected with SFV-A36R were fixed and permeabilized and subjected to phalloidin and anti-A36R labeling. The same cells are shown on the left and right. Note the localization of A36R in stress fibers and the strong labeling at the tips of cell projections.

View larger version (243K): [in this window] [in a new window]   FIG. 7.   Immunoelectron microscopy of cells expressing A36R protein. (A) Uniform high labeling of the plasma membrane. There is also some label on the membrane of a putative endosome (E), whereas the mitochondrion (M) is unlabeled. (B) Details of plasma membrane labeling (small arrowheads), including a prominent surface projection (large arrowhead). (C) Heavy labeling for A36R in membrane structures that are in continuity (arrowhead) with the nuclear envelope (N, nucleus). Note the label in vesicles close to the nucleus and in the nuclear envelope (arrows). (D) Significant labeling (arrows) on one side of the Golgi stacks (G). The arrowheads indicate putative clathrin-coated vesicles, whose presence is an indication that the labeled aspect of the Golgi is the TGN. Bars, 100 nm.

A56R protein. Protein A56R, the viral hemagglutinin, is a type I glycoprotein which is heavily glycosylated. In vaccinia virus-infected cells it was present in the Golgi region as well as at the cell surface. We did not detect A56R enrichment in any defined structures that could represent virions, although the protein has been described as a component of the EEV envelope. When expressed separately, the protein was efficiently transported to the cell surface, producing immunofluorescence images similar to those of vaccinia virus-infected cells (Fig. 8).

View larger version (93K): [in this window] [in a new window]   FIG. 8.   Localization of A56R. Immunofluorescence staining was carried out on BHK-21 cells fixed at 7 h postinfection. Both nonpermeabilized (NP) and Triton X-100-permeabilized (P) cells are shown. The cells were either mock infected (M), infected with vaccinia virus (V), or infected with SFV-A56R particles (S). Note the strong labeling of the plasma membrane in both vaccinia virus- and SFV-A56R-infected cells.

B5R protein. The B5R protein is a type I transmembrane glycoprotein with homology to complement control proteins. As mentioned above, immunofluorescence on vaccinia virus-infected cells showed B5R in a juxtanuclear area, as well as in enveloped virions. When expressed from the SFV construct, the protein was efficiently retained intracellularly in a region close to the nucleus (Fig. 9A). The area labeled with anti-B5R antibody corresponded to the Golgi region, as suggested by double-labeling experiments with WGA (Fig. 9B). We also carried out immunoelectron microscopy of cells expressing B5R (Fig. 10). In agreement with the immunofluorescence results, the protein was associated with vesicles, which were often found in the proximity of the nucleus. Thus, except for the absence of enveloped virions, the distribution of the protein expressed by itself was indistinguishable from that seen in vaccinia virus-infected cells.

View larger version (202K): [in this window] [in a new window]   FIG. 9.   Localization of B5R protein. (A) Immunofluorescence staining was carried out on BHK-21 cells fixed at 7 h postinfection. Both nonpermeabilized (NP) and Triton X-100-permeabilized (P) cells are shown. The cells were either mock infected (M), infected with vaccinia virus (V), or infected with SFV-B5R particles (S). (B) Costaining with rhodamine-labeled WGA. Cells infected with vaccinia virus (V) or SFV-B5R (S) were fixed and permeabilized and subjected to WGA and anti-B5R labeling. The same cells are shown on the right and left. Note that localization of B5R is coincident with the Golgi region in both vaccinia virus-infected cells and SFV-B5R-infected cells.

View larger version (126K): [in this window] [in a new window]   FIG. 10.   Immunoelectron microscopy of cells expressing B5R protein. (A) Overview of labeling, with prominently labeled intracellular vesicles (arrows). There is little labeling on the plasma membrane (P), and one gold particle is seen over the nuclear envelope (N, nucleus). (B and C) Details of the vesicle labeling. Bars, 100 nm.

F13L protein. The F13L protein is the most abundant protein in the EEV envelope. It is palmitylated and associates with the cytoplasmic side of the membrane. In permeabilized vaccinia virus-infected cells, F13L showed a pattern similar to that of the B5R protein, except for some diffuse labeling. Expression from the SFV construct resulted in a different localization pattern (Fig. 11). The protein was not visible in nonpermeabilized cells, as would be expected from the protein topology. In permeabilized cells, the protein was distributed throughout the cytoplasm, producing a diffuse staining pattern.

View larger version (84K): [in this window] [in a new window]   FIG. 11.   Localization of F13L protein. Immunofluorescence staining was carried out on BHK-21 cells fixed at 7 h postinfection. Both nonpermeabilized (NP) and Triton X-100-permeabilized (P) cells are shown. The cells were either mock infected (M), infected with vaccinia virus (V), or infected with SFV-F13L particles (S).

B5R chimeric constructs. The localization of B5R in the absence of vaccinia virus infection indicates that the B5R protein contains Golgi localization signals. In order to determine the portion of the protein responsible for Golgi targeting, we expressed fusion proteins in which the GFP was fused to the 42-amino-acid C-terminal portion of B5R, encompassing the transmembrane and cytosolic portions of the protein. These chimeric genes were introduced for expression in vaccinia virus, as well as in SFV constructs. Expression in the context of vaccinia virus infection resulted in a pattern that was similar to that of normal B5R (Fig. 12), indicating that the fusion protein was normally targeted to the Golgi complex and was incorporated efficiently into wrapped virions. Interestingly, when GFP-B5R was expressed independently (Fig. 12) of the SFV construct, the fusion protein also accumulated in the juxtanuclear region, demonstrating that the C-terminal 42 amino acids are sufficient for retention of the protein.

View larger version (105K): [in this window] [in a new window]   FIG. 12.   Localization of B5R protein chimeras. Cells infected with the indicated vaccinia viruses (WR-GFP or WR-GFPmB5R) or with the corresponding SFV replicons (SFV-GFP or SFV-GFPmB5R) are shown.

Immunofluorescence in B5R-deficient virus. The above-mentioned results point to B5R as a candidate to determine the Golgi complex as the retention site for the remaining EEV envelope proteins. Therefore, we tested whether the wrapping process was altered when B5R was absent or modified. We used F13L as a marker for the wrapping compartment. Figure 13 shows the distribution of F13L in cells infected with mutant viruses lacking either the complete B5R gene (W-B5R), the cytoplasmic tail (W-B5R_c), or most of the luminal domain (W-B5R_SCR1-4). A decrease in the number of individual virions was visible in cells infected with W-B5R, consistent with the role of B5R in virus envelopment (12, 46). In these cells, the F13L protein was efficiently localized to the Golgi area. Interestingly, deletion of the B5R cytoplasmic tail, where Golgi localization signals have been described, had no effect on F13L distribution (Fig. 13, W-B5R_c). This observation is consistent with the normal plaque phenotype of this mutant virus (23). These results did not reveal any major change in the distribution of F13L as a result of B5R mutation, suggesting that the absence of B5R protein by itself does not affect the determination of the wrapping compartment.

View larger version (78K): [in this window] [in a new window]   FIG. 13.   Localization of EEV envelope proteins in vaccinia virus B5R mutants. Cells infected with the indicated vaccinia viruses were fixed and stained with antibody to F13L protein. Wild-type virus (WR) or mutant viruses lacking most of the B5R gene (W-B5R), most of the extracellular domain (W-B5RSCR1-4), or the cytoplasmic tail (W-B5Rc) were used. Note the juxtanuclear localization of F13L in all cases. Differences in the number of enveloped virions reflect differences caused by B5R mutation.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

At the beginning of this study, we anticipated that at least some of the proteins of the EEV envelope contained signals responsible for their retention in the Golgi complex. If at least one was retained, other EEV-specific proteins could be indirectly retained in the wrapping vesicles by protein-protein interactions. Indeed, such interactions between some of these proteins have been postulated (33). From this point of view, the protein(s) responsible for TGN targeting should in principle be identifiable by individual expression of the different EEV envelope proteins. The results presented here show a more complex scenario. On one hand, B5R is the only EEV envelope protein that, when expressed by itself, showed a clear-cut Golgi complex localization similar to that seen in vaccinia virus-infected cells, indicating that it contains Golgi localization signals. This conclusion is also supported by other recent studies. Thus, a number of lines of evidence point to protein B5R as the major determinant of the Golgi complex as the wrapping compartment. However, the subcellular compartment for the wrapping process does not seem to be altered in B5R-deficient virus, suggesting that other proteins which show a partial Golgi localization, such as A33R, A34R (this report), and F13L (7), may also contribute to this process. In agreement with this idea, a recent report showed that retention of a significant fraction of the B5R protein in the endoplasmic reticulum does not affect the wrapping process but rather leads to a reduction in the amount of B5R incorporated into EEV (26).

Several of the EEV-specific proteins have been expressed previously in transfected cells (7, 19, 38, 40). In general, our results using the SFV expression system are in good agreement with these studies. It is to be noted that protein F13L showed a somewhat different pattern when expressed transiently (19) in a stable cell line (7) or by SFV-directed expression (this report). Since membrane association of F13L depends on palmitylation, it is likely that differences in the amount and/or the period of expression may account for these different patterns.

One important observation is that different EEV envelope proteins showed different distributions in vaccinia virus-infected cells (Fig. 1). From the immunofluorescence patterns on vaccinia virus-infected cells, proteins B5R, F13L, and A36R appear to be the most specific markers for the EEV envelope. In contrast, proteins A33R, A34R, and A56R showed significant deviation from the typical pattern, suggesting that they are targeted to a number of additional membrane locations in the cell. Thus, it is likely that multiple protein-protein interactions of different efficiencies, as well as the relative concentrations of the various proteins, determine both the intracellular distributions of individual proteins and their degree of incorporation into the wrapping membranes.

A56R, the viral hemagglutinin, is a heavily glycosylated protein that has been shown to be present in purified EEV preparations (28, 30). However, our results and previous reports (38, 40) show that it is efficiently transported to the plasma membrane. It is not clear whether A56R is enriched in the EEV envelope, since viruslike structures are not visible in infected cells by anti-A56R staining. Therefore, as an alternative to specific incorporation in the wrapping membranes, it is possible that, in the absence of specific retention and due to its transit through the Golgi complex, some of the protein may be passively incorporated in the EEV envelope. In fact, nonspecific incorporation of many membrane proteins, including host proteins, in the EEV envelope has recently been demonstrated (43), reinforcing the idea that enrichment, rather than the presence of a particular protein in the EEV envelope, is required to demonstrate specific retention of that protein and selective incorporation into the TGN.

The A36R protein also requires special consideration. On one hand, the localization of this protein during vaccinia virus infection indicates that it is specifically targeted to the Golgi complex and incorporated in the wrapping membranes. Despite this, and in contrast to other EEV envelope proteins, we consider it likely that A36R is an IEV but not an EEV protein. We have consistently observed that a proportion of paraformaldehyde-fixed CEVs have envelopes that have been disrupted, as demonstrated by labeling with anti-F13L antibodies (Fig. 11) or antibodies to IMV surface proteins that are inaccessible to intact CEV or EEV (44). In contrast, anti-A36R antibody did not label CEV under the same conditions, and labeling of intracellular A36R required permeabilization of the cells with detergent. These observations may suggest that A36R is exclusively in the outer membrane of IEV and is excluded from EEV.

SFV expression of A36R also shows localization of the protein at cell projections and association with actin fibers. The latter observation suggests specific interactions of B5R with cytoskeletal elements, whose significance goes beyond the scope of this report. Interestingly, A36R has been identified as a key protein in the induction of actin polymerization by IEV (13, 33).

Our results also complement studies of the characterization of functional domains within B5R protein. The extracellular domain, which shows similarity to complement control proteins, has been shown to be involved in retention of viruses at the plasma membrane (17, 25). Other reports indicate that the transmembrane domain is important for the function of the protein in the wrapping process (17, 23, 25). The complete B5R can be targeted to the Golgi complex in the absence of other viral proteins (reference 19 and this report). We have extended these studies to show that the C-terminal 42 amino acids of the B5R protein, spanning the transmembrane and cytosolic domains, are sufficient to confer Golgi complex targeting on a GFP fusion protein. When the chimeric GFP fusion protein was expressed from a vaccinia virus recombinant, it was also incorporated into EEV (data not shown). Significantly, the C terminus of B5R has also been shown to mediate incorporation of chimeric human immunodeficiency virus Env-B5R proteins into the virus particle (19). A recent report has revealed a contribution of certain residues within the cytoplasmic tail to the process of intracellular transport of the protein (45). However, we present evidence here that deletion of the cytoplasmic tail of B5R does not significantly affect either the virus wrapping process or the Golgi complex targeting of other EEV envelope proteins.

Taken together, these results imply that multiple proteins in the EEV envelope contribute to determine the wrapping compartment and that several of these proteins interact specifically with different cellular membrane structures.