|Home | About | Journals | Submit | Contact Us | Français|
Poxviruses are the only DNA viruses known to replicate and assemble in the cytoplasm of infected cells. Poxvirus morphogenesis is a complicated process in which four distinct infectious forms of the virus are produced: intracellular mature virus, intracellular enveloped virus, cell-associated enveloped virus, and extracellular enveloped virus. The source of primary membrane wrapping the intracellular mature virus, the first infectious form, is still unknown. Although the membrane was suggested to originate from the endoplasmic reticulum-Golgi intermediate compartment, none of the marker proteins from this or any other cell compartments has been found in the intracellular mature virus. Thus, it was hypothesized that the membrane is either extensively modified by the virus or synthesized de novo. In the work described here, we demonstrate that a host cell protein residing in the trans-Golgi network membrane, golgin-97, is transported to the sites of virus replication and assembly and becomes incorporated into the virions during poxvirus infection. Inside the virion, golgin-97 is associated with the insoluble core protein fraction. Being able to adopt a long rod-like structure, the protein apparently extends through the virion envelope and protrudes from its surface. Here we discuss the potential role and functions of golgin-97 in poxvirus replication and propose two working models.
Poxviruses are large enveloped double-stranded DNA viruses with a complex morphology. Vaccinia virus (VV), the most extensively studied member of the Poxviridae, possesses a genome of approximately 192 kbp in size that encodes over 200 proteins whose expression is temporally regulated (22). Unlike other DNA viruses, VV replicates and assembles in the perinuclear compartments within the cytoplasm, termed viral factories or virosomes (22). The process of virus morphogenesis is very complicated and occurs in several stages. It begins with the assembly of membrane crescents around electron-dense material containing viral DNA and core proteins (11), leading to formation of spherical immature virus (IV). This step is followed by a series of proteolytic cleavages and condensation of the virus core, transforming IV into oval or brick-shaped intracellular mature virus (IMV), the first infectious form of VV (21). IMV represents the majority of produced virions and remains trapped inside the cell until its lysis. A subfraction of IMV moves on microtubules from the viral factories to the trans-Golgi network (TGN) (30) to acquire two additional membranes and become intracellular enveloped virus (IEV). Transported to the cell surface along microtubule networks (5, 9, 27, 38), IEV particles bud through the plasma membrane and lose the outermost envelope. This process generates cell-associated enveloped virus (CEV). CEV induces formation of actin tails that propel the particles, now referred to as extracellular enveloped virions (EEV), outside the cell and facilitates cell-to-cell and long-range dissemination of the virus.
The origin of the virion membranes is not well understood. At the present time, there are two hypotheses explaining formation of the primary (crescent) membrane surrounding IV and IMV particles. One of them suggests that the membrane is synthesized de novo, induced by virus infection (3). Other studies have proposed that the primary membrane is derived from the intermediate endoplasmic reticulum (ER)-Golgi compartment (ERGIC). It has been shown that the membranes contiguous with crescents are labeled with intermediate ERGIC markers (33) and that IMV envelope proteins A17L, A14L, and A13L accumulate at the ERGIC during virus infection (14, 29). So far, all attempts to confirm this model and detect host cell proteins from ERGIC or any other cellular compartment incorporated into IMV particles have been unsuccessful (13). It is also not clear what intracellular membranes are used to generate the secondary membrane that wraps IEV. Schmelz and coworkers (31) suggested that IEV is surrounded by modified TGN membranes. Others provided evidence suggesting a hypothesis that the secondary membrane originates from endosomal cisternae (35, 37). Supporting these models, both TGN and endosomal markers β-1,4-galactosyltransferase and transferrin receptor CD71, respectively, were shown to be associated with purified EEV particles (13).
In the work reported here, we demonstrate that a TGN membrane protein, golgin-97, not only copurifies with VV particles but is also found to be a virion component. In VV-infected cells, golgin-97 is shown to translocate to the viral factories. In the purified virions the protein is detected in a proteolytically processed form. Immunoelectron microscopy, proteinase protection assays, and virion fractionation results suggest that golgin-97 is associated with insoluble virus core proteins, but consistent with its rod-like structure the protein seems to penetrate the virion envelope and protrude through the virion surface.
HeLa cells, BSC-40 monkey kidney cells, and McCoy mouse fibroblast cells were grown in minimal essential medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS; Invitrogen). Human embryonic kidney 293 cells were grown in Dulbecco's modified Eagle's medium (Invitrogen) with 10% FBS. Vaccinia virus strains Western Reserve (WR), Copenhagen (COP), IHD-J, and cowpox virus (CPX) were propagated and purified by two cycles of sucrose gradient centrifugation as described previously (10). Briefly, the viruses were grown in monolayers of BSC-40 cells maintained in minimal essential medium supplemented with 5% FBS for 40 h. The infected cells were harvested, resuspended in 10 mM Tris-HCl (pH 8.0), and homogenized in a Dounce homogenizer. Nuclei and unbroken cells were removed by centrifugation at 1,500 × g for 10 min, 4°C. The supernatant was layered onto the 36% sucrose cushion and centrifuged at 40,000 × g for 80 min, 4°C. The resultant pellet, resuspended in 1 mM Tris-HCl (pH 8.0) and homogenized in a Duall homogenizer, was layered onto a 25% to 40% continuous sucrose gradient and centrifuged at 22,500 × g for 40 min. The opalescent band containing virus was withdrawn with a syringe and spun down at 22,000 × g for 40 min.
golgin-97 (GOLGA1) cDNA clone (ID IOH27115) was purchased from Invitrogen (Ultimate ORF Clones collection). In order to insert FLAG sequence upstream or downstream of the golgin-97 open reading frame, its cDNA sequence was amplified by PCR with the following sets of primers: N-terminal FLAG-golgin-97 fusion, 5′-CACCATGGATTATAAAGACGATGACGACAAGTTTGCAAAACTGA-3′ and 5′-CTAGGACCATGGTATCC-3′; C-terminal FLAG-golgin-97 fusion, 5′-CACCATGTTTGCAAAACTGA-3′ and 5′-CTACTTGTCGTCATCGTCTTTATAATCGGACCATGGTATCC-3′. The PCR product was subcloned into the pcDNA6.2/V5/GW/D-TOPO vector (Invitrogen), yielding pN-FLAG:G97 and pC-FLAG:G97 constructs expressing N-terminal and C-terminal FLAG-golgin-97 fusion proteins, respectively.
Anti-human golgin-97 (GOLGA1) mouse monoclonal antibodies and CDF4 and CDFX clones were purchased from Invitrogen and GeneTex, Inc. (San Antonio, TX), respectively. Anti-FLAG mouse monoclonal antibodies were purchased from Sigma-Aldrich, Inc. (St. Louis, MO). Rabbit polyclonal anti-I3L antibodies (40) were kindly provided by J. Krijnse-Locker (European Molecular Biology Laboratory, Heidelberg, Germany). Rabbit polyclonal anti-L1R and anti-p25 antibodies were produced by our lab and have been described elsewhere (15, 25).
HeLa cells were grown in six-well plates to 80% confluency and transfected with 1μg of plasmid DNA using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. At 24 h posttransfection, the cells were infected with VV WR at a multiplicity of infection (MOI) of 5.0 PFU per cell. After 24 h of incubation the cells were harvested and analyzed by immunoblot assay. For virus purification, the cells were grown in eight 100-mm plates and transfected with 3 μg of plasmid DNA using the same protocol. At 3 h posttransfection the cells were infected with VV WR at an MOI of 5.0 PFU/cell and incubated for 40 h. Virus purification was carried out according to the protocol described above.
The cells were plated on glass coverslips in 24-well plates at 30% to 40% confluency. The next day, the cells were infected with vaccinia virus strain WR at an MOI of 1.0 PFU/cell. At the indicated times postinfection, HeLa and McCoy cells were fixed with 100% methanol for 10 min. 293 and BSC-40 cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS, pH 7.4) for 20 min and then permeabilized in 0.2% Triton X-100 in PBS for 10 min. The coverslips were washed three times with PBS, blocked in 2% bovine serum albumin (BSA)-PBS (pH 7.4) (P-BSA) for 30 min, and incubated with primary rabbit anti-I3L (1:2,000) and mouse anti-golgin-97 (1 μg/ml) CDF4 antibodies diluted in P-BSA for 1 h. The cells were washed three times with PBS and once with P-BSA and then incubated with anti-rabbit (Southern Biotechnology Associates, Inc., Birmingham, AL) and anti-mouse (Invitrogen) antibodies conjugated to tetramethylrhodamine isothiocyanate (TRITC) and Alexa Fluor 488, respectively. The coverslips were mounted in ProLong Gold antifade reagent with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen) and analyzed with a Zeiss LSM 510 Meta confocal laser scanning microscope. The channels were collected in multitrack mode.
HeLa and 293 cells were infected with VV WR at an MOI of 5 PFU/cell. At 24 h postinfection (hpi), the cells were scraped from the plates, resuspended in culture medium, and pelleted by centrifugation at 700 × g, 4°C. The cells were fixed in 1% paraformaldehyde, 0.5% glutaraldehyde, 200 mM cacodylate buffer (pH 7.4) for 4 h. Embedding in LR White resin and sectioning were performed by personnel of the Oregon State University Electron Microscopy laboratory. The thin sections were placed on uncoated gold 300-mesh grids. The grids were blocked in P-BSA for 1 h at room temperature and incubated with primary mouse anti-golgin-97 CDF4 antibodies (4 μg/ml) in P-BSA for 1 h at room temperature and then overnight at 4°C. The sections were washed in six droplets of PBS and one droplet of P-BSA (5 min/each) and then incubated with secondary anti-mouse antibodies conjugated to 18-nm gold (Jackson Immunoresearch Laboratories, Inc., West Grove, PA) diluted 1:20 in P-BSA for 3 h at room temperature. The grids were washed in six droplets of PBS (5 min/each) and deionized water and stained with lead citrate and uranyl acetate. The samples were examined on a Philips EM 300 electron microscope.
Approximately 108 particles were applied to carbon-coated gold 300-mesh grids and air dried. The samples were processed as described previously (1) with minor modifications. The virions were fixed in 4% paraformaldehyde in PBS for 1 h and permeabilized with 0.2% Triton X-100 in PBS or left unpermeabilized. The grids were washed three times in PBS, blocked in 5% BSA, 2% normal goat serum in PBS (pH 7.4) for 30 min, and incubated with primary golgin-97 antibodies CDF4 (4 μg/ml) diluted in blocking solution as described above. For virion permeabilization experiments, the grids were washed and incubated with secondary 5-nm gold-conjugated anti-mouse antibodies (Sigma-Aldrich, Inc.) diluted 1:20 in blocking solution. Alternatively, the virions were labeled with anti-golgin-97 CDF4 antibodies and anti-mouse antibodies conjugated to 30-nm gold particles (Ted Pella, Inc., Redding, CA) using the same protocol. The samples were stained with ammonium molybdate and examined on a Philips EM 300 electron microscope.
HeLa, BSC-40, and McCoy cells were mock infected or infected with the indicated viruses at an MOI of 10 PFU/cell and incubated for 17 hpi. The infected cells were harvested, pelleted by centrifugation at 700 × g, 4°C, and resuspended in 10 mM Tris (pH 8.0). The cells were lysed by freeze-thawing and cleared by centrifugation at 1,500 × g, 4°C. The samples were equalized based on the measured optical density at 280 nm. VV and CPX virions were purified as described above. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene fluoride membrane (Immobilon-PSQ; Millipore, Billerica, MA). The membrane was blocked in 3% gelatin in TTBS (0.05% Tween 20, 150 mM NaCl, 50 mM Tris-HCl, pH 7.4) and incubated with primary rabbit anti-L1R (1:4,000), rabbit anti-p25 (1:2,000), or mouse anti-golgin-97 CFD4 (1 μg/ml) and CDFX (1:100) antibodies in 1% gelatin-TTBS overnight. The membrane was washed three times in TTBS and incubated with secondary anti-rabbit (Promega, Madison, WI) or anti-mouse (Invitrogen) antibodies conjugated to horseradish peroxidase. The proteins were detected with a chemiluminescence kit (West-Pico; Pierce Biotechnology, Inc., Rockford, IL).
The virions were fractionated by detergent treatment according to a previously published protocol (1) with some modifications. Briefly, the virions (1010 particles) were incubated in 1% NP-40, 50 mM dithiothreitol (DTT), 5 mM MgCl2, Tris-HCl (pH 8.0) buffer for 30 min at room temperature. Membrane and core proteins were separated by centrifugation at 14,000 × g, 4°C. The cores were divided into soluble and insoluble fractions by incubation with 1% NP-40, 0.1% SDS, 0.5% sodium deoxycholate for 30 min at room temperature, followed by centrifugation at 14,000 × g, 4°C. The fractions were analyzed by immunoblotting.
Purified virions (4 × 108 particles) were incubated in 1 mM CaCl2, 50 mM Tris-HCl (pH 8.0) buffer in either the presence or absence of trypsin (50 μg/ml; Promega) for 1 h at room temperature. The digestion reaction was stopped by the addition of SDS-PAGE sample buffer. The samples were analyzed by immunoblotting.
Screening the distribution of a variety of ER and Golgi apparatus resident proteins in VV-infected cells (data not shown), we were surprised to note that one of the TGN membrane proteins, golgin-97, accumulates in the viral factories at 24 hpi. To determine when golgin-97 is transported into the virosome during infection, we performed a time course experiment using immunofluorescence microscopy. HeLa cells infected with VV (WR strain) were fixed at 3, 6, 9, 12, and 24 hpi and processed for immunostaining with antibodies to golgin-97 and the VV single-stranded DNA-binding protein, I3L, which is expressed early in infection (28, 36). Because I3L was previously shown to localize to the viral DNA replication sites (28, 40), antibodies to this protein were used as a marker for the viral factories, in particular early in infection when viral DNA cannot be visualized with DAPI staining. The golgin-97 distribution in uninfected and VV-infected cells is shown in Fig. 1A to F (http://web.science.oregonstate.edu/~hruby/supp.htm). Viral DNA replication factories appear to organize initially in close proximity to golgin-97 accumulation sites, the TGN membranes (Fig. (Fig.1B).1B). The golgin-97 translocation to the viral factories was detected at 6 hpi (Fig. (Fig.1C).1C). By 24 hpi, most of golgin-97 accumulated in the factories as discrete punctate bodies (Fig. (Fig.1F).1F). Although an overlap in golgin-97 and I3L accumulation sites could be noticed in some cells, the proteins seemed to distribute to distinct subdomains within the virosome. Similar relocation of golgin-97 into the virosome was also detected in another human cell line, 293 cells infected with VV WR (data not shown). All of the 126 examined VV-infected HeLa cells contained virosomes with golgin-97 inclusions, although the size and number of inclusions varied for each individual cell.
We also investigated localization of golgin-97 orthologs in uninfected and VV WR-infected BSC-40 monkey kidney cells and McCoy mouse fibroblasts using the anti-human golgin-97 antibodies. Due to significant homology between golgin-97 and its orthologs, the antibodies recognized a TGN resident protein in most vertebrate species. In our immunoblotting experiments, the anti-human golgin-97 antibodies detected a protein of the correct size (~97 kDa) in both the BSC-40 and McCoy cell lines (data described below). The immunofluorescence assay showed that by 24 hpi, BSC-40 and McCoy golgin-97 orthologs also assembled in punctate structures inside viral factories analogous to those found in human cell lines (data not shown). This suggested that similar mechanisms of golgin-97 redistribution exist in all other host species infected with VV WR, implying that this phenomenon is induced by virus infection.
To study golgin-97 accumulation at viral DNA replication sites in more detail, ultrathin sections of VV-infected cells were examined by immunoelectron microscopy. The infected cells were fixed at 24 hpi and processed for immunolabeling with primary anti-golgin-97 antibodies and secondary antibodies conjugated to 18-nm gold particles. In agreement with the earlier study, gold particles indicating golgin-97 accumulation were found in the virosomal area. Heavy labeling was detected around electron-dense structures that appeared to be the previously described DNA crystalloids (Fig. (Fig.2A)2A) and inclusion bodies (Fig. (Fig.22 A, B, and C) (8, 34). Moreover, some of the gold grains were observed in close proximity to IV and IMV particles and seemed to be associated with both of the virion forms (Fig. 2B and C), leading to a hypothesis that golgin-97 might be a virion component.
To investigate the potential association of golgin-97 with VV virions, we purified the virus from three different cell lines (human HeLa cells, monkey kidney BSC-40 cells, and mouse fibroblast McCoy cells) and analyzed the preparations by immunoblotting with anti-human golgin-97 CDF4 antibodies. As demonstrated above, the anti-human golgin-97 antibodies cross-react with the protein orthologs from other vertebrate species. Immunoblotting of the uninfected and infected cell lysates demonstrated that human golgin-97 antibodies detect a protein of approximately 97 kDa in size from all three cell lines, HeLa, BSC-40, and McCoy (Fig. (Fig.3).3). The VV virions purified from these cells also contained a protein recognized by antibodies to golgin-97 (Fig. (Fig.3).3). Interestingly, the protein band from purified virus was smaller in size (~60 kDa). To confirm that the 60-kDa band is a golgin-97 fragment, we tested VV virions with the other antibodies against golgin-97, clone CDFX. These antibodies also recognized a 60-kDa protein fragment in the purified virus, suggesting that golgin-97 is processed in the infected cell either before or after its incorporation into the virus particle. To exclude the possibility that the protein band resulted from immunological cross-reaction between golgin-97 antibodies and abundant virus core proteins p4a and p4b, we tested the antibodies by immunofluorescence and immunoblot assays (data not shown). Both assays revealed no immunological cross-reactivity between the core proteins and golgin-97 antibodies and vice versa.
To investigate golgin-97 incorporation in virions of other orthopoxviruses, we purified VV strains COP and IHD-J and also CPX, a close relative of VV, using the protocol mentioned above. All three virus preparations contained proteolytically processed golgin-97 of the same size (~60 kDa) (Fig. (Fig.3),3), implying that digestion of golgin-97 and its incorporation into the virion could be common for other orthopoxviruses as well.
Because the antibodies used in the experiment described above are monoclonal, the other products of proteolytic cleavage of golgin-97 remained undetected by immunoblot analysis. In order to identify other fragments of the protein potentially incorporated into poxvirus virions and to confirm golgin-97 targeting to the virions, we have propagated and purified VV from cells transiently expressing a FLAG-golgin-97 fusion protein.
The FLAG sequence was inserted either upstream or downstream of the golgin-97 open reading frame and subcloned under control of the cytomegalovirus promoter. The resulting plasmids expressing N-terminal and C-terminal FLAG-golgin-97 fusion proteins were named pN-FLAG:G97 and pC-FLAG:G97, respectively.
Transient expression of the fusion proteins in uninfected HeLa cells demonstrated that the protein fusions are expressed at the full length (~97 kDa) and detected by both anti-golgin-97 and anti-FLAG antibodies (Fig. 4A and B). In the infected cell lysates expressing either N-terminal or C-terminal fusion protein, anti-golgin-97 antibodies detected an additional band of approximately 80 kDa (Fig. (Fig.4A),4A), thus confirming that the protein is processed in the presence of virus. The failure to detect those bands in untransfected cells is likely due to insufficient amounts of the processed endogenous golgin-97 protein. Also, we cannot exclude that golgin-97 overexpression or presence of the FLAG sequence may result in a different proteolytic digestion pattern. Immunoblot analysis with anti-FLAG antibodies of the infected cell lysates expressing the C-terminal FLAG-golgin-97 fusion showed two other protein bands (~70 kDa and ~25 kDa) that were not recognized by golgin-97 CDF4 antibodies (Fig. (Fig.4B).4B). Similarly, these protein bands were not detected by another golgin-97 antibody, clone CDFX (data not shown). Because the 70-kDa band does not contain anti-golgin-97 antibody epitopes, it is likely that the previously found golgin-97 60-kDa fragment is produced by an alternative processing of the protein.
VV WR virions were propagated in HeLa cells untransfected or transfected with either pN-FLAG:G97 or pC-FLAG:G97 and purified by the standard protocol described above. Immunoblot analysis of the purified viruses with anti-FLAG antibodies revealed that one of the C-terminal FLAG-golgin-97 fusion protein fragments (~70 kDa) is incorporated into the virions (Fig. (Fig.4B).4B). These results confirm targeting of golgin-97 fragments into poxvirions and strongly suggest that the previously detected 60-kDa protein band is in fact a golgin-97 fragment. None of the N-terminal FLAG-golgin-97 fusion protein fragments was detected in the purified virions, indicating that the FLAG sequence may be cleaved prior to or interfere with incorporation of the protein into virions. Hence, it is possible that there are additional golgin-97 fragments incorporated in the virion which have yet to be identified.
To determine the subviral localization of golgin-97, we fixed purified VV virions. The samples were labeled with anti-golgin-97 CDF4 antibodies and 5-nm gold-conjugated secondary antibodies either before or after permeabilization with Triton X-100 and examined by immunoelectron microscopy. Both permeabilized (Fig. (Fig.5B)5B) and nonpermeabilized (Fig. (Fig.5A)5A) virions were labeled in a similar manner. The gold grains remained outside the virion, suggesting that the protein is exposed at the virion surface and is likely to be incorporated in the virion envelope. This interpretation would be also consistent with the fact that golgin-97 is a membrane protein.
To confirm golgin-97 localization in the virion envelope, purified VV WR particles were fractionated by detergent treatment (1% NP-40, 50 mM DTT) into membrane and core proteins. Surprisingly, golgin-97 was absent in the membrane fraction (Fig. (Fig.6A).6A). Further separation of cores into soluble and insoluble fractions in the presence of 0.1% SDS and 0.5% sodium deoxycholate revealed that golgin-97 fractionates strictly with insoluble core proteins (Fig. (Fig.6A).6A). As controls for this experiment, we used viral proteins L1R and p25. L1R is a VV myristoylprotein shown to specifically associate with the virion envelope and fractionate with membrane proteins (4, 25, 26). P25 is a structural VV protein previously localized to the core fraction (25, 39). Taken together, these results suggest that golgin-97 is held in the virion by its strong association with insoluble VV core proteins. At the same time, the protein seems to penetrate the virion envelope and protrude from its surface.
To test our hypothesis we carried out a proteinase protection assay. Purified VV WR virions were incubated in the presence or absence of trypsin. In such an assay, all proteins exposed at the virion surface would be degraded by a proteinase while virion proteins protected by the virion envelope would remain intact. Consistent with our immunoelectron microscopy results, most of the golgin-97 CDF4 epitope was destroyed in the presence of trypsin (Fig. (Fig.6B).6B). As expected, when exposed at the virion surface viral membrane protein, L1R was also degraded (25), but viral core protein p25 was protected from digestion by the membrane (25) (Fig. (Fig.6B6B).
Purified virions were also examined with golgin-97 CDFX antibodies by immunogold electron microscopy and a proteinase protection assay. The results were similar to those obtained with CDF4 antibodies (data not shown), suggesting that the epitopes for both golgin-97 antibodies are exposed at the virion surface and are likely located in the same protein region close to each other. With both permeabilized and nonpermeabilized virions, we observed some variation in the number of associated gold grains. The proteinase protection assay demonstrated high sensitivity of the protein to proteinases, suggesting that a fraction of the exposed golgin-97 epitopes could be degraded during the virus purification process. Hence, it is possible that some of the incorporated protein would remain undetectable by the monoclonal antibodies used in our experiments.
This study began with screening the distribution of the ER and Golgi marker proteins in VV-infected cells in order to identify potential host factors recruited by the virus. As a result of these experiments, we found that one of the TGN membrane proteins, golgin-97, redistributes to the viral factories in VV-infected cells. The protein assembles into discrete punctate structures inside the factories by 24 hpi. Analogous translocation and accumulation patterns were found for monkey and mouse orthologs of golgin-97. Further investigation demonstrated that golgin-97 and its orthologs are incorporated into the virions of all three tested VV strains and of Cowpox virus, another orthopoxvirus. It is interesting that the detected protein was smaller in size (~60 kDa) and likely to be a product of virus-induced proteolytic processing of golgin-97. Transient expression of N-terminal and C-terminal FLAG-golgin-97 fusion proteins in the infected cells revealed other potential products of the protein proteolytic digestion. In addition to this, one of those fragments, ~70-kDa C-terminal FLAG:golgin-97, was found inside the virions propagated and purified from the cells transiently expressing the fusion protein, thus confirming that golgin-97 becomes processed and incorporated into poxvirus particles during virus infection. The results of virion fractionation and immunogold electron microscopy showed that the golgin-97 60-kDa fragment fractionates with insoluble core proteins and at the same time is exposed at the virion surface.
The golgin-97 protein (97 kDa), an autoantigen associated with Sjogren's syndrome, was cloned by screening a HeLa cDNA library with serum from Sjogren's syndrome patients (7). golgin-97 belongs to the group of GRIP domain proteins. It has a rod-like structure with extensive coiled-coil regions and a conserved GRIP domain at its extreme C terminus (2, 12, 23). The presence of the GRIP domain was shown to be necessary and sufficient for protein targeting to the cytoplasmic surface of the TGN membrane (2, 12, 23). Immunoprecipitation experiments, cross-linking, and yeast two-hybrid analyses suggest that all GRIP domain proteins, including golgin-97, form parallel coiled-coil homodimers (20). Recruitment of golgin-97 to the TGN membrane is facilitated by the Arl1 protein, a member of the ARF/Arl family of small GTPases via its interactions with the GRIP domain (17-19, 32). The functions of golgin-97 in the cell are still unclear. Due to shared similarities in structure and localization with other golgins, the protein was proposed to play the role of a tethering molecule in vesicle trafficking and/or to be a TGN matrix protein (6). Recent studies have implicated golgin-97 in regulation of both endocytic and exocytic vesicular trafficking (16, 19).
Our results on golgin-97 subviral localization are fully supported by the presence of large coiled-coil regions in the protein sequence, implying that it can adopt a long rod-like structure. Taking into consideration the ability of the protein to form dimers, we propose that golgin-97 in monomeric form or as a homodimer becomes incorporated into the virion via its interactions with one or several core proteins comprising the insoluble fraction. The protein extends through the virion envelope and becomes exposed at the surface of the virion (Fig. (Fig.7).7). Although golgin-97 does not possess a transmembrane domain, its transmembrane localization inside the virion could be explained either by its potential lipid modification, possibly induced by virus, or by its interactions with viral membrane proteins. Immunoblotting analysis of transiently expressed golgin-97 and its truncated versions tagged with FLAG (data not shown) suggests that the epitopes for the monoclonal antibodies used in our experiments lie close to the N terminus of the protein. Therefore, it is likely that the golgin-97 N-terminal region stays outside the virion. It has yet to be determined whether golgin-97 proteolysis occurs before its incorporation into the virion or after, but we cannot exclude the possibility that there are additional golgin-97 fragments trapped inside the virion.
Data showing translocation of golgin-97 and its orthologs to the sites of virus replication and following incorporation of these proteins into the virions, demonstrated for two orthopoxviruses and in different host cell lines, suggest that the protein could be important for virus replication. As a cellular membrane protein, golgin-97 might participate in the process of virion envelopment either by facilitating delivery of the TGN membranes to the virion assembly sites or, being incorporated into the virion, by transporting virus particles to the TGN. According to the first model, golgin-97 attached to TGN membrane fragments might interact with core proteins and aid IV maturation. This hypothesis argues with the idea of an ERGIC origin of the IMV wrapping membrane. Alternatively, being exposed at the virion surface, golgin-97 might traffic IMV particles to the TGN, where they acquire double-walled membranes and transform into IEV. The second hypothesis does not conflict with the existing models of the formation of the primary membrane and supports the hypothesis about a TGN-derived secondary membrane, implicating golgin-97 in IEV morphogenesis.
This study was supported by NIH grant number AI21335 and in part by NIH grant number 1S10RR107903-01.
We acknowledge the Confocal Microscopy Facility of the Center for Gene Research and Biotechnology and the Environmental and Health Sciences Center at Oregon State University. We thank Jacomine Krijnse-Locker (EMBL, Heidelberg, Germany) for providing anti-I3L-antibodies, Michael Nesson (Electron Microscopy Laboratory, Oregon State University) for assistance with electron microscopy, and Cliff Gagnier for help with virus purification.
Published ahead of print on 20 September 2006.