| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
First published online July 29, 2005; 10.1104/pp.105.066019 Plant Physiology 138:1877-1895 (2005) © 2005 American Society of Plant Biologists The Potato Virus X TGBp2 Movement Protein Associates with Endoplasmic Reticulum-Derived Vesicles during Virus Infection1Department of Entomology and Plant Pathology (H.-J.J., T.D.S., R.M., K.K., J.V.-L.), and Department of Statistics (M.P.), Oklahoma State University, Stillwater, Oklahoma 74078; and Plant Biology Division, The Samuel Roberts Noble Foundation, Ardmore, Oklahoma 73401 (Y.-S.W., E.B., R.S.N.)
The green fluorescent protein (GFP) gene was fused to the potato virus X (PVX) TGBp2 gene, inserted into either the PVX infectious clone or pRTL2 plasmids, and used to study protein subcellular targeting. In protoplasts and plants inoculated with PVX-GFP:TGBp2 or transfected with pRTL2-GFP:TGBp2, fluorescence was mainly in vesicles and the endoplasmic reticulum (ER). During late stages of virus infection, fluorescence became increasingly cytosolic and nuclear. Protoplasts transfected with PVX-GFP:TGBp2 or pRTL2-GFP:TGBp2 were treated with cycloheximide and the decline of GFP fluorescence was greater in virus-infected protoplasts than in pRTL2-GFP:TGBp2-transfected protoplasts. Thus, protein instability is enhanced in virus-infected protoplasts, which may account for the cytosolic and nuclear fluorescence during late stages of infection. Immunogold labeling and electron microscopy were used to further characterize the GFP:TGBp2-induced vesicles. Label was associated with the ER and vesicles, but not the Golgi apparatus. The TGBp2-induced vesicles appeared to be ER derived. For comparison, plasmids expressing GFP fused to TGBp3 were transfected to protoplasts, bombarded to tobacco leaves, and studied in transgenic leaves. The GFP:TGBp3 proteins were associated mainly with the ER and did not cause obvious changes in the endomembrane architecture, suggesting that the vesicles reported in GFP:TGBp2 studies were induced by the PVX TGBp2 protein. In double-labeling studies using confocal microscopy, fluorescence was associated with actin filaments, but not with Golgi vesicles. We propose a model in which reorganization of the ER and increased protein degradation is linked to plasmodesmata gating.
Plant viruses use the cellular endomembrane system to support virus replication (Schaad et al., 1997  ; Reichel and Beachy, 1998; Carette et al., 2002a, 2002b; Schwartz et al., 2002). Among positive-strand RNA viruses, the viral replicase is typically anchored to membranes associated with the endoplasmic reticulum (ER), the endocytic pathway, or cellular organelles (Rubino and Russo, 1998; Mas and Beachy, 1999; Rubino et al., 2001; Schwartz et al., 2002, 2004). Many viral replication-associated proteins cause reorganization of cellular membranes with the purpose of creating centers that protect the viral replication complexes (VRCs) from cellular nucleases and host defenses (Schwartz et al., 2002, 2004; Ding et al., 2004). Specifically, the replication complexes of plant viruses belonging to the genera Potyvirus, Comovirus, Dianthovirus, Pecluvirus, and Bromovirus cause proliferation and vesiculation of either the perinuclear or the cortical ER (Schaad et al., 1997; Dunoyer et al., 2002; Lee and Ahlquist, 2003). The replicases of cowpea mosaic virus and red clover necrotic mosaic virus induce ER proliferation (Carette et al., 2002a, 2002b; Turner et al., 2004). The best-studied example is the brome mosaic virus (BMV) 1a replicase, which localizes to the ER and induces membrane invaginations to form spherules (Lee and Ahlquist, 2003). These invaginations, or spherules, are centers for virus replication and contain the BMV 2a protein and viral RNA (Schwartz et al., 2004). It was suggested that the BMV-induced spherules resemble budding retrovirus cores and that these membrane invaginations containing VRCs have a place in virus evolution (Schwartz et al., 2002, 2004).
Recent information supports the involvement of the endomembrane system in virus intracellular and intercellular movement (for review, see Reichel and Beachy, 1998
Potato virus X (PVX) encodes three movement proteins from a genetic module of three overlapping open reading frames (ORFs), termed the triple gene block. These triple gene block proteins, named TGBp1, TGBp2, and TGBp3, have molecular masses of roughly 25, 12, and 8 kD (Huisman et al., 1988 In this study, the green fluorescent protein (GFP) gene was fused to the PVX TGBp2 gene to study protein subcellular targeting during virus infection. Time-course experiments were conducted in BY-2 protoplasts and in tobacco leaves to examine the association of the GFP:TGBp2 protein with the ER network during the course of virus infection. We present evidence that the PVX TGBp2 protein, similar to many viral membrane-associated proteins, induces formation of vesicles from the ER. In addition, the PVX TGBp3 protein also associates with the ER. Last, we present evidence that TGBp2-induced vesicles associate with the actin network and are unrelated to the Golgi apparatus.
GFP:TGBp2 Associates with Cellular Membranes during Virus Infection in Protoplasts
The PVX-GFP infectious clone (Fig. 1) expresses the GFP gene from a duplicated coat protein subgenomic promoter. GFP is used as a visual marker to follow virus infection over time. The GFP:TGBp2 fused genes were inserted into the PVX genome in place of the GFP gene to study TGBp2 protein subcellular targeting during virus infection (Fig. 1). Most of the endogenous TGBp2 gene was deleted from the PVX-GFP:TGBp2 infectious clone (Verchot et al., 1998
Infectious PVX-GFP and PVX-GFP:TGBp2 transcripts were inoculated to BY-2 protoplasts, and the subcellular distribution of fluorescence was studied at 18, 24, 36, 48, and 72 h postinoculation (hpi) using confocal microscopy. In PVX-GFP-inoculated protoplasts, fluorescence was mainly cytosolic and nuclear, with no obvious change in the pattern of GFP expression over time (Fig. 2, F and K). These data support related studies in which the GFP protein is typically cytosolic and nuclear when expressed in tobacco leaves and protoplasts (Oparka et al., 1999 ; Crawford and Zambryski, 2000, 2001; Yang et al., 2000).
There were three patterns of GFP:TGBp2 protein fluorescence noted at each time point: (1) in vesicles only, (2) in vesicles and cytosol, and (3) in cytosol and nucleus (Fig. 2, A–D). The vesicles containing GFP:TGBp2 proteins were perinuclear as well as lining the plasma membrane (Fig. 2, A, C, and D). Vesicles sometimes aggregated (Fig. 2B). GFP:TGBp2 fluorescence associated with the nucleus, but not with the nucleolus, as visualized after 4',6-diamino-phenylindole (DAPI) staining of protoplasts (Fig. 2, E, G, H, and I). We quantified the patterns of GFP:TGBp2 protein fluorescence to determine whether the subcellular distribution of GFP:TGBp2 fluorescence changes over time. Thirty protoplasts were scored at each time point for fluorescence associated with the vesicles, nucleus, and/or cytosol (Fig. 2J). The proportion of protoplasts containing GFP:TGBp2 fluorescence in vesicles was 80% at 18 hpi and declined to 27% by 48 hpi. This level was maintained through 72 hpi (Fig. 2J). The proportions of protoplasts containing GFP:TGBp2 fluorescence in the cytosol and nucleus were 40% at 18 hpi and increased to 80% and 90% at 48 and 72 hpi, respectively (Fig. 2J). Thus, GFP:TGBp2 fluorescence associates mainly with vesicles early in virus infection, but increasingly with the cytosol and nucleus as infection progresses (Fig. 2J). Statistical analyses were performed using the data represented in Figure 2, J and K, to determine whether the subcellular distribution of fluorescence in PVX-GFP:TGBp2- and PVX-GFP-infected protoplasts is regulated over time. Time is a likely variable in studying protein subcellular trafficking. If the TGBp2 proteins traffic from one location to another in the cell, then changes in the distribution of GFP:TGBp2 fluorescence within the cell over time would not be random. For example, movement of the GFP:TGBp2 proteins from vesicles into the cytosol may be regulated by the host degradation machinery. Since GFP:TGBp2 is an ER-associated protein, it is possible that ER stress stimulates export of the GFP:TGBp2 proteins to the cytosol for degradation by the 26S proteasome. It is also possible that movement of the GFP:TGBp2 proteins would be regulated and not random if it is controlled by interactions with other PVX proteins. Therefore, a logistic regression was used to test whether time is a variable in the subcellular accumulation patterns of GFP:TGBp2 and GFP fluorescence. The presence/absence of fluorescence in each subcellular address (vesicle, nucleus, and cytoplasm; dependent variable) were compared over time (independent variable). We hypothesized that the observed changes in the distribution of GFP:TGBp2 proteins over time would indicate subcellular trafficking of the fusion protein. If time is not a variable controlling the distribution of fluorescence within the cell, then statistical analyses would indicate that the distribution of protein in each subcellular address is independent of time (i.e. random) and no correlation would exist. Slope estimates and P values were determined. A nonsignificant slope (i.e. large P value) indicates no relationship to time. Conversely, a significant slope (i.e. small P value) indicates the response is affected by time. A negative slope indicates a decreased presence of fluorescence in each address over time, and a positive slope indicates an increase of fluorescence in each address over time. For PVX-GFP:TGBp2, the slope values for vesicles, nucleus, and cytoplasm were –0.0367, 0.0313, and 0.0463, respectively. The P values were 0.0001, 0.0014, and 0.0001, respectively. The negative slope for PVX-GFP:TGBp2 in vesicles indicates that fluorescence decreased in these locations over time. The positive slope for PVX-GFP:TGBp2 in the cytoplasm and nucleus indicates fluorescence increased in these locations over time. The extremely low P values indicate that changes in subcellular targeting of GFP:TGBp2 proteins over time were significant. For PVX-GFP, the slope values for vesicles, nucleus, and cytoplasm were 0.0145, –0.0155, and –0.0155, respectively. The P values were 0.3631, 0.4009, and 0.4009, respectively. The nonsignificant slopes and high P values for PVX-GFP indicate that the change in fluorescence at each subcellular address had no relationship to time and was random. Thus, statistical analysis supports the hypothesis that GFP:TGBp2 proteins first resided in vesicles and over time became increasingly cytosolic and nuclear.
PVX-GFP and PVX-GFP:TGBp2 transcripts were inoculated to tobacco leaves and fluorescence was examined between 2 and 5 d postinoculation (dpi) to determine whether the pattern of fluorescence observed in tobacco leaves resembles the pattern reported in protoplasts. Within a single infection focus, cells were viewed early and late during virus infection. The intracellular fluorescence patterns located at the leading edge and at the center of infection foci were compared to determine whether the subcellular distribution of fluorescence changed as infections developed. Since tobacco leaf cells are not synchronously infected, we could not use statistical methods to assess changes in the distribution of GFP or GFP:TGBp2 proteins over time.
In PVX-GFP-inoculated leaves, the patterns of fluorescence at the leading edge and at the center of the infection foci were similar. Fluorescence was observed in the cytosol, nucleus, and perinuclear inclusion bodies (Fig. 3, A and B). Sometimes it was difficult to discern between GFP fluorescence inside the nucleus and perinuclear inclusion bodies. These inclusion bodies are likely X-bodies described in early literature as masses containing virus particles, ER, and ribosomes (Kikumoto and Matsui, 1961
PVX-GFP:TGBp2 infection foci were observed from 2 to 5 dpi. Fluorescence was associated with the ER network and vesicles in all infected cells. Perinuclear inclusion bodies were observed in all infected cells (Fig. 3, C–E). Vesicles appeared to move along the ER strands in tobacco epidermal cells (data not shown). A thick band of cytosolic fluorescence was evident in cells located at the center of infection, but was less obvious at the leading edge of infection. Although the intensity of fluorescence was weaker at the leading edge of infection, the patterns of fluorescence in the ER network and vesicles at this location were similar to those observed in cells at the center of the infection foci (Fig. 3, C–F). Single optical cross-sections taken through the middle of infected cells revealed vesicles lining the cell wall (Fig. 3, E and F). These vesicles were sometimes twin structures along the walls of adjacent cells and resemble the peripheral bodies reported to occur in beet necrotic yellow vein virus (BNYVV; a benyvirus), potato mop top (PMTV; a pomovirus) or the poa semilatent virus (PSLV; a hordeivirus; Erhardt et al., 2000 ; Solovyev et al., 2000; Gorshkova et al., 2003; Zamyatnin et al., 2004; Haupt et al., 2005) infected cells. Data obtained in inoculated plants confirmed observations in protoplasts. In early infected cells and protoplasts, the GFP:TGBp2 proteins accumulated mainly in vesicles. In cells and protoplasts infected for a long period of time, we observed an increase in cytosolic fluorescence. In plants, GFP:TGBp2 proteins were found simultaneously in the ER as well as in vesicles. We never obtained evidence that the GFP:TGBp2 proteins move from the ER into vesicles, or from vesicles into the ER. Rather, the data indicate that GFP:TGBp2 proteins are either simultaneously targeted to the ER and cellular vesicles or that the GFP:TGBp2 proteins induce reorganization of the ER soon after translation.
In PVX-GFP:TGBp2-infected plants, we observed GFP:TGBp2 fluorescence in the ER as well as in vesicles, suggesting that the vesicles may be invaginations of the ER network. In protoplasts inoculated with PVX-GFP:TGBp2, fluorescence was redistributed from the vesicles into the cytosol. It is therefore reasonable to consider that ER retention of viral proteins and/or ER reorganization by the TGBp2 protein triggers ER-associated protein degradation as a result of ER stress (Brandizzi et al., 2003 Fluorometric assays were used to quantify GFP expression over time as a measure of both protein accumulation and degradation in BY-2 protoplasts. We then compared the rates of protein turnover in protoplasts inoculated with the PVX viruses with the rates of protein turnover in protoplasts transfected with the pRTL2 plasmids. Inoculated and transfected protoplasts were chased at 24 h with cycloheximide to halt protein synthesis. Fluorometric values were obtained at 0, 4, 8, and 12 h following treatment with cycloheximide and the protein half-lives were calculated. The fluorometric values were normalized to a time 0 measurement. The data were plotted and a linear regression was used to calculate protein half-lives. We predicted that if protein turnover was stimulated as the result of virus infection, then fluorescence would decline more rapidly in virus-infected protoplasts than in pRTL2-transfected protoplasts. On the other hand, if TGBp2 contained a degradation signal, then we would expect greater turnover of GFP:TGBp2 than GFP proteins in both PVX-inoculated and pRTL2-transfected protoplasts. In protoplasts transfected with PVX-GFP or PVX-GFP:TGBp2 transcripts, there was a linear increase in fluorescence between 12 and 48 hpi (Fig. 4A). The fluorescence values obtained at 48 hpi were between 8- and 12-fold higher than the values obtained at 12 hpi. The plotted fluorometric values for PVX-GFP:TGBp2- and PVX-GFP-infected protoplasts were almost identical to each other during most of the time course (Fig. 4A). Since the decrease in GFP and GFP:TGBp2 fluorescence had similar rates, the stability of GFP is not altered by fusion to the TGBp2 protein (Fig. 4B). The half-lives of GFP and GFP:TGBp2 proteins were similar, 8.9 and 8.6 h, respectively.
In protoplasts transfected with pRTL2-GFP or -GFP:TGBp2 plasmids, fluorescence did not show the same linear increase between 12 and 48 h post-transfection (hpt) as reported for the PVX-GFP- and PVX-GFP:TGBp2-infected protoplasts (compare Fig. 4, A and C). Instead, the fluorometric values obtained for these pRTL2 constructs reached a plateau at approximately 24 hpt (Fig. 4C). The values representing GFP expression were 4-fold greater than the values representing GFP:TGBp2 fluorescence between 24 and 48 hpt (Fig. 4C). This was surprising since the fluorometric values obtained in PVX-GFP- and PVX-GFP:TGBp2-infected protoplasts were not significantly different. Perhaps protein accumulation in these transient assays is different when GFP:TGBp2 is expressed from a DNA or RNA promoter (cauliflower mosaic virus [CaMV] 35S promoter or PVX coat protein subgenomic promoter). Since protein expression reaches a plateau at 24 hpt in protoplasts transfected with pRTL2 constructs but continues to increase in a linear fashion in PVX-inoculated protoplasts, it is possible that nuclear gene expression from a CaMV 35S promoter is more tightly regulated than cytosolic expression of the same gene(s) from a PVX viral promoter. The rate of decrease in GFP fluorescence following cycloheximide treatment was greater than the rate of decrease in GFP:TGBp2 fluorescence. The half-lives calculated for the GFP and GFP:TGBp2 proteins were 12.4 and 27.1 h, respectively. The GFP:TGBp2 fusion protein was more stable than the nonfused GFP protein (Fig. 4D). Evidence that the GFP and GFP:TGBp2 proteins have shorter half-lives when expressed from the PVX genome than from pRTL2 constructs indicates that protein degradation is stimulated during virus infection. The half-life of the GFP:TGBp2 protein is 8.6 h when expressed from the PVX genome, and this may account for increasing cytosolic fluorescence during the course of infection, representing the last cellular location during degradation. Since the GFP:TGBp2 protein has a 2- to 3-fold longer half-life when expressed from the pRTL2 plasmid, cytosolic accumulation of fluorescence would be minimal in studies involving these plasmids.
To determine whether the PVX TGBp2 protein itself associates with vesicles in the absence of virus infection, pRTL2-GFP:TGBp2 plasmids were transfected into protoplasts or bombarded to tobacco leaves. For comparison, plasmids expressing GFP, GFP:TGBp3, or mGFP5-ER, which encodes a version of GFP that has an N-terminal basic chitinase signal peptide and a C-terminal HDEL sequence for ER targeting and retention (Siemering et al., 1996 Ninety-three percent (28/30) of the protoplasts expressed GFP:TGBp2 in vesicles only, and 13% (4/30) of the protoplasts showed fluorescence in the ER as well as in vesicles associated with the ER (Table I). All cells contained some aggregates of vesicles (Table I; Fig. 5, A–C). As in the PVX-GFP:TGBp2-infected protoplasts, the vesicles or aggregates were either perinuclear or along the plasma membrane (compare Fig. 2, A and B, with Fig. 5A).
Both GFP:TGBp3 and mGFP5-ER proteins were observed in the ER network in protoplasts and tobacco leaves (Fig. 5, E–H). The tubules and branches of the cortical ER network were shorter in protoplasts (Fig. 5, D–H) than in leaf epidermal cells, making the network in BY-2 protoplasts more difficult to resolve microscopically than in intact leaves. The small volume of the protoplasts cause the ER network to appear more compact than it appears in tobacco leaf epidermal cells. We did not observe fluorescent vesicles in protoplasts or tobacco leaf epidermal cells expressing either the GFP:TGBp3 or mGFP5-ER constructs (Table I; data not shown). In protoplasts or bombarded leaves expressing only GFP, fluorescence was cytosolic and nuclear (Table I; Fig. 5, I–J). In comparing the data obtained using GFP:TGBp2, GFP:TGBp3, and mGFP5-ER constructs, it appears that the GFP:TGBp2 proteins have the unique ability to associate with vesicles.
It is common that viral proteins associating with the endomembrane system cause changes in membrane architecture. The PVX TGBp2 protein may cause invaginations of the ER network similar to those caused by the membrane-binding proteins encoded by TMV, BMV, tobacco etch virus, and peanut clump virus (PCV; Schaad et al., 1997
To determine the origin of the GFP:TGBp2-associated vesicles, we conducted immunogold labeling and electron microscopic analysis of freeze-substituted leaf tissue taken from transgenic tobacco expressing mGFP5-ER, GFP:TGBp2, or GFP:TGBp3 proteins (Figs. 6 and 7; Table II). The mGFP5-ER or GFP:TGBp3 transgenic tobacco was used for comparison because these fusion proteins are known to associate with the ER (Krishnamurthy et al., 2003
Immunolabeling was carried out using antiserum against GFP (since antibodies to the PVX TGBp2 and TGBp3 proteins were unavailable) and/or binding protein (BiP; Table II). BiP antisera were chosen because the BiP protein is an ER resident chaperone that is not in other areas of the endomembrane system (Fontes et al., 1991 ). Ten-nanometer gold particles were associated with GFP antisera, and 25-nm gold particles were associated with BiP antisera in all experiments. The presence of immunogold labeling in the ER, cell wall, and plasma membrane was quantified and compared statistically (Table II). The average numbers of gold particles (of each type) associated with the cell wall, plasma membrane, ER, vesicles, and Golgi apparatus were calculated in 1-µm2 fields. Gold particles embedded in the cell wall were scored as cell wall particles. Gold particles deposited on the cytosolic side of the cell wall near the plasma membrane were considered to be associated with the plasma membrane. Immunogold labeling with the GFP and BiP antisera was specific (Table II). In mGFP5-ER or GFP:TGBp3 transgenic samples treated with GFP and/or BiP antisera, the gold particles were mainly along the ER (Fig. 6, A, C, and D, and Fig. 7, D–F; Table II) and were rarely in the plasma membrane, cell wall, or Golgi network. There was minimal label in nontransgenic samples treated with GFP antisera (data not shown; Table II). Only the BiP antisera labeled the ER in nontransgenic cells (Fig. 6G; Table II). Statistical analyses confirm the mGFP5-ER and GFP:TGBp3 fusion proteins localize mainly to the ER network (Table II). Additional controls included treating samples with buffer, TMV antisera, or secondary antisera. There was little or no labeling of the mGFP5-ER, GFP:TGBp2, and GFP:TGBp3 transgenic or nontransgenic samples following any of these treatments (Fig. 6, B–F; nontransgenic data not shown; Table II). Membrane vesicles were plentiful in GFP:TGBp2-expressing tissues (Fig. 7, A–C), but were absent from mGFP5-ER transgenic, GFP:TGBp3 transgenic, or nontransgenic leaf samples (Table II; Figs. 6 and 7). At high magnification, we saw granules along the outline of many vesicles that we regarded as ribosomes (Fig. 7, A–C). Vesicles were mainly observed in the cortical region of the cell and were often abutting the cell wall (Fig. 7, B and C). Vesicles ranged from 150 to 500 nm in diameter, and the number of vesicles per 1-µm2 field ranged from 1 to 3. Roughly 70% of the fields analyzed in GFP:TGBp2 samples contained vesicles (Table II). In GFP:TGBp2 transgenic samples, immunolabeling with GFP was mainly associated with the ER and TGBp2-induced vesicles (Fig. 7, A–C; Table II). In GFP:TGBp3 transgenic samples, GFP and BiP immunolabeling was mainly along the ER (Fig. 7, D–F; Table II). In both GFP:TGBp2 and GFP:TGBp3 samples, gold particles were rarely present in the cell wall, plasma membrane, or Golgi apparatus (Fig. 7, B and C; Table II). These data indicate that GFP:TGBp2 and GFP:TGBp3 are ER-associated proteins when expressed alone. The vesicles seen in GFP:TGBp2 transgenic cells were never present in GFP:TGBp3, mGFP5-ER, or nontransgenic leaves, suggesting that they are TGBp2-induced structures. Immunogold labeling indicated that GFP:TGBp2 proteins associate with these vesicles. Since these vesicles were specific to GFP:TGBp2-expressing samples and were studded with ribosomes (Fig. 7, A and C), it is likely the GFP:TGBp2 protein induced reorganization of the ER to produce these structures.
Brefeldin A (BFA) causes Golgi proteins to redistribute into the ER and impedes transport vesicles from cycling between the two compartments (Brandizzi et al., 2002
GFP:TGBp2-, GFP:TGBp3-, and mGFP5-ER-expressing leaf segments were treated with 200 µg mL–1 BFA for 3 to 4 h (Ritzenthaler et al., 2002
Previous studies have shown that movement is facilitated by interactions with the actin microfilaments (Heinlein et al., 1995 ; McLean et al., 1995; Liu et al., 2005). While the immunogold experiments indicate that the GFP:TGBp2-induced vesicles are ER derived, we were uncertain whether these vesicles associate with the actin network. To further characterize the behavior of these GFP:TGBp2-induced vesicles, leaf segments expressing GFP:TGBp2 were treated with latrunculin B, which is a potent disrupter of actin filaments (Nebenfuhr et al., 1999; Ritzenthaler et al., 2002). Latrunculin B also has minor effects on the shape of the ER, causing an increase in cisternal ER and blind-ending tubules (Knebel, et al., 1990; Nebenfuhr et al., 1999; Brandizzi et al., 2003). Leaf segments expressing GFP:TGBp2, GFP:TGBp3, mGFP5-ER, or GFP:TGBp3 were treated for 1 to 2 h with latrunculin B. The reticulate pattern typical of the ER network was absent from GFP:TGBp2-expressing cells treated with latrunculin B (Fig. 8I). Instead, we saw GFP:TGBp2-containing vesicles in aggregates, which appeared as many pinwheels within a cell (Fig. 8I). The vesicles seemed to coalesce as a result of latrunculin B disruption of the actin network. The pattern of fluorescence in GFP:TGBp2 differed from GFP:TGBp3 and mGFP5-ER cells treated with latrunculin B. In the mGFP5-ER and GFP:TGBp3 samples treated with latrunculin B, the cisternal ER was more prevalent than the tubular ER (Fig. 8, G and H). In both the mGFP5-ER and GFP:TGBp3 samples treated with latrunculin B, the tubules of the ER have fewer branches (Fig. 8, G and H).
To further examine the association of GFP:TGBp2-induced vesicles and actin filaments, tobacco leaves were bombarded with plasmids expressing GFP:TGBp2 and either DsRed:Talin or GFP:Talin. DsRed:Talin- and GFP:Talin-expressing cells displayed filamentous arrays typical of the actin network (Fig. 9, A–E). Higher magnification images clearly showed GFP:TGBp2-containing vesicles along the length of the red- or green-labeled actin filaments (Fig. 9, B–E). When these cells were treated with latrunculin B, the actin network dissolved and GFP:TGBp2 vesicles were dispersed in the cytosol (Fig. 9F).
Although immunogold labeling indicates that the TGBp2-induced vesicles are primarily associated with the ER, confocal images suggest that the GFP:TGBp2 vesicles resemble the pattern of Golgi vesicles reported in previous studies (Boevink et al., 1998 ; Saint-Jore et al., 2002). To validate the immunogold-labeling experiments, we cobombarded plasmids expressing GFP:TGBp2 and DsRed-sialyl transferase (ST) to tobacco leaves. DsRed-ST has the signal anchor of a rat ST fused to DsRed-2 and is targeted to the Golgi apparatus (Dixit and Cyr, 2002). The red and green fluorescent vesicles observed in bombarded cells were due to DsRed-ST and GFP:TGBp2 expression, respectively, in cobombarded cells (Fig. 9, G–J). However, the DsRed and GFP signals did not overlap, demonstrating that GFP:TGBp2 did not localize to the Golgi network (Fig. 9, G–J), consistent with the results from our immunogold-labeling studies. While the cycloheximide chase experiments support the hypothesis that the PVX TGBp2 protein accumulates in the cytosol and nucleus as a result of protein degradation, we could not rule out the possibility that the TGBp2 protein was redirected to the cytosol by other PVX proteins. This hypothesis is also supported by the finding that GFP:TGBp2 is cytosolic when expressed from the PVX genome, but is not cytosolic when expressed from pRTL2 plasmids. For example, in protoplasts expressing only GFP:TGBp2, we saw vesicles and aggregates of vesicles (Fig. 9K). In protoplasts expressing only GFP:TGBp1, fluorescence was nuclear and cytosolic (Fig. 9L). It is possible that, if TGBp1 and TGBp2 interact during virus infection, some populations of GFP:TGBp2 could appear to be nuclear and cytosolic as well. Further investigation is needed to explore the effects of viral protein-protein interactions on their subcellular accumulation patterns.
There are several reports that the TGBp2 and TGBp3 proteins of PSLV, BNYVV, and PMTV associate with the ER and with peripheral bodies (sometimes called punctate bodies; Erhardt et al., 2000 ; Solovyev et al., 2000; Cowan et al., 2002; Gorshkova et al., 2003; Morozov and Solovyev, 2003; Zamyatnin et al., 2004; Haupt et al., 2005). Since studies were conducted using confocal microscopy, the nature of the peripheral bodies was a source of speculation. Some have speculated that these peripheral bodies are vesicles similar to Golgi vesicles (Morozov and Solovyev, 2003). The peripheral bodies observed in studies involving BNYVV, PMTV, and PSLV were sometimes described as twin structures forming on opposite walls of adjacent cells, as seen in this study in PVX-infected cells (Fig. 3, D and E; Erhardt et al., 2000; Solovyev et al., 2000; Cowan et al., 2002; Gorshkova et al., 2003; Morozov and Solovyev, 2003; Zamyatnin et al., 2004; Haupt et al., 2005). This observation has led to the suggestion that these viruses use the secretory pathway to move within the cell toward the plasmodesmata (Morozov and Solovyev, 2003). Here, we demonstrated that the PVX TGBp2 protein induces similar peripheral bodies that could be seen under the confocal and electron microscope (Figs. 2, 3, 5, and 7). GFP:TGBp2 was observed in the ER and in vesicles in PVX-GFP:TGBp2-infected protoplasts and tobacco plants, in pRTL2-GFP:TGBp2-transfected protoplasts and bombarded tobacco leaves, and in GFP:TGBp2 transgenic tobacco leaves. Vesicles were not observed in mGFP5-ER- or GFP:TGBp3-expressing samples, suggesting that these vesicles were uniquely induced by the GFP:TGBp2 protein. Using three different experimental systems, new evidence was obtained indicating that the peripheral bodies induced by the PVX TGBp2 protein are vesicles resulting from invagination of the ER network and are unrelated to the Golgi apparatus. By electron microscopy, it was determined that the vesicles were sometimes studded with ribosomes, suggesting that they were ER derived (Fig. 7, A–C). BiP and GFP antisera colabeled vesicles in GFP:TGBp2 transgenic tissues. Immunolabeling using GFP antiserum did not detect GFP:TGBp2 along the Golgi apparatus, indicating that these vesicles were not likely to be Golgi related (Table II). Further evidence that these vesicles were not related to the Golgi apparatus was obtained in transient expression studies using the Golgi marker DsRed-ST or treatment with BFA (Figs. 8 and 9). These combined data provide further evidence that the vesicles were more likely to be derived from the ER network rather than the Golgi apparatus.
The peripheral bodies reported for PSLV were induced by the TGBp3 protein (Solovyev et al., 2000
The association of PVX TGBp2 with the actin cytoskeleton and ER, but not with the Golgi, is consistent with recent results with the PMTV TGBp2/3 complex (Haupt et al., 2005
Viral proteins that are known to associate with cellular membranes often cause changes in the endomembrane architecture (Schwartz et al., 2002 We observed GFP:TGBp2 in the cytosol and nucleus during PVX infection but not when GFP:TGBp2 was expressed from pRTL2 plasmids. Statistical analysis of the data presented in Figure 2 support the hypothesis that the GFP:TGBp2 protein was redirected to the cytosol by either viral or cellular factors. The pattern of fluorescence observed in protoplasts expressing GFP:TGBp1 was somewhat similar to the pattern of fluorescence observed in PVX-GFP:TGBp2-infected cells (compare Fig. 9L with Fig. 2, C and D). Presently, very little is known regarding interactions among the triple gene block proteins. Further research is needed to learn whether protein subcellular targeting is affected by viral protein-protein interactions.
Another explanation for cytosolic and nuclear accumulation of GFP:TGBp2 relates to the increased levels of protein turnover observed in protoplasts infected with PVX (Brandizzi et al., 2003
The TMV movement protein and the turnip yellow mosaic virus movement proteins are degraded by the 26S proteasome (Reichel and Beachy, 2000
A model was proposed in which the regulation of calcium stores in the ER by calreticulin controls unconventional myosin VIII activities within the plasmodesmata and thereby controls plasmodesmata gating (Baluska et al., 1999
To test this model, further research is needed to determine whether PVX infection triggers ER stress responses and whether there is a link between ER stress response and plasmodesmata gating. The ability of virus infection to cause fluctuations in calcium levels across the ER is also worthy of further investigation.
Bacterial Strains and Plasmids
All plasmids were constructed using Escherichia coli strain JM109 (Sambrook et al., 1989
The plasmids pRTL2-GFP, -GFP:TGBp1, -GFP:TGBp2, and -GFP:TGBp3 were prepared previously (Fig. 1; Yang et al., 2000
Suspension cells of tobacco BY-2 (Nagata et al., 1992
Protoplasts were isolated from the suspension cells as previously described with slight modifications (Gaire et al., 1999
Protoplasts were inoculated with infectious viral RNA transcripts. Transcripts were prepared from pPVX-GFP, pPVX12D-GFP, and pPVX-GFP:TGBp2 plasmids described previously (Baulcombe et al., 1995
Protoplasts were also transfected with pRTL2 plasmids by electroporation as described by Gaire et al. (1999) RNA- or DNA-transfected protoplasts were collected, washed with phosphate-buffered saline (PBS), and then fixed with PBS containing 3% paraformaldehyde (v/v) and 5 mM EGTA for 1 h at room temperature, followed by washing twice with PBS. Fixed samples were analyzed by confocal microscopy. Some protoplasts were stained with DAPI. Transfected protoplasts were incubated in 10 µg mL–1 DAPI in growth media. Protoplasts were then analyzed by confocal microscopy.
In some experiments, protoplasts were inoculated with PVX transcripts or transfected with pRTL2 plasmids and fluorescence was monitored fluorometrically at 12, 18, 24, 30, 36, and 48 h as described previously (Howard et al., 2004
Tobacco (Nicotiana benthamiana and Nicotiana tabacum) plants were used for studying the subcellular targeting of most proteins. Transgenic N. tabacum plants expressing mGFP5-ER, GFP:TGBp2, or GFP:TGBp3 were prepared previously using Agrobacterium transformation, as previously described (Krishnamurthy et al., 2003 |
TOP
ABSTRACT
RESULTS
, and 125 µF using 0.4-cm gap cuvettes (Bio-Rad Laboratories, Hercules, CA) and a gene pulser (Bio-Rad Laboratories). Following electroporation, protoplasts were immediately transferred into a new tube containing 1 mL of solution 2 and incubated on ice for 30 min, and then at 30°C for 5 min. Protoplasts were collected by centrifugation at 39g for 5 min, resuspended in 1 mL of solution 3 (BY-2 culture media plus 0.45 M mannitol), and transferred to 6-well cell culture plates (Corning, Corning, NY) coated with solution 3 plus 1.0% agarose (w/v; pH 5.7). Protoplasts were cultured at 27°C and then collected at 18, 24, 36, and 48 hpi by centrifugation at 39g for 5 min.