*Short slender actin filaments were found instead of thick actin tails.

†By intranasal route (G. L. Smith, unpublished data).

Electron microscopy indicated that most EEV particles originate from IEV (Payne & Kristenson, 1979). Consistent with this proposal, levels of EEV were reduced when the formation of IEV was inhibited by drugs (Payne & Kristenson, 1979; Ulaeto et al., 1995) or the deletion of gene B5R (Engelstad & Smith, 1993; Wolffe et al., 1993) or F13L (Blasco & Moss, 1991). However, mutants lacking gene A33R or A34R produced higher levels of EEV despite the wrapping of IMV being less efficient (Duncan & Smith, 1992; Wolffe et al., 1997) or incomplete (Roper et al., 1998). This observation questions whether EEV made by these mutants is formed by an alternative pathway such as budding.

Here, we have investigated the roles of IMV, IEV, CEV and EEV in VV spread in vitro, by using a panel of virus mutants lacking individual IEV- or EEV-specific genes, and Abs that neutralize IMV or EEV. We demonstrate that comet-shaped plaques are probably made by convection currents, that VV spreads from cell to cell by Ab-sensitive and Ab-resistant pathways, and that the A33R protein has a role in Ab-resistant spread.

Cells and viruses. RK₁₃ and BS-C-1 cells were grown in Dulbecco's modified Eagle's medium (DMEM) (Gibco BRL) containing 5 % heat-inactivated foetal bovine serum (HFBS). For infection of cells, virus was diluted in DMEM supplemented with 2 % HFBS (DMEM/2 %). VV strains IHD-J and WR (Alcamí & Smith, 1992), and deletion mutants vA33R (Roper et al., 1998), vA34R (McIntosh & Smith, 1996), vA36R (Parkinson & Smith, 1994), vA56R (Sanderson et al., 1998a), vB5R (Engelstad & Smith, 1993), vF12L (Zhang et al., 2000) and vF13L (Blasco & Moss, 1992) originating from strain WR, were described previously. vRevA34R is the revertant virus of vA34R in which the A34R gene was inserted back into the virus genome (McIntosh & Smith, 1996). IMV was purified as described elsewhere (Law & Smith, 2001) and was used for all infections.

Antibodies. Mouse monoclonal Ab (mAb) 2D5 against the IMV L1R protein (Ichihashi & Oie, 1996), mAb AB1.1 against the IMV D8L protein (Parkinson & Smith, 1994), rabbit antiserum against the B5R protein (-B5R) (Galmiche et al., 1999) and VV-immune rabbit antiserum Rb-WR2 (Law & Smith, 2001) were described previously. Antisera were heat-inactivated at 56 °C for 30 min before use.

Plaque assays. (i) Liquid overlay for comet formation. IMV was diluted in DMEM/2 % and adsorbed onto cells for 2 h at 37 °C. Unbound virus was washed away with PBS and the cells were overlaid with liquid medium (DMEM/2 %) and stained 2 days later (unless specified otherwise) with 0.05 % crystal violet in 15 % ethanol. Antibodies and 10 µg/ml IMCBH (N₁-isonicotinoyl-N₂-3-methyl-4-chlorobenzoylhydrazine) (Payne & Kristenson, 1979) were included in the overlays where indicated. (ii) Semi-solid overlay. As for (i) except that virus was adsorbed for 1 h and cells were overlaid with DMEM/2 % containing 1.5 % carboxymethylcellulose (CMC).

Titration of EEV. The infectivity of EEV was quantified as described (Law & Smith, 2001). Briefly, fresh virus supernatants were collected at the indicated times, diluted and incubated with mAb 2D5 (diluted 1/2000) for 1 h at 37 °C to neutralize contaminating IMV. When specified, Rb-WR2 Ab was included to neutralize EEV. The virus was adsorbed onto cells for 1 h, washed, and overlaid with 1.5 % CMC in DMEM/2 %. After incubation, the plaques were stained as above.

Microscopy. Methods for indirect immunofluorescent staining have been described elsewhere recently (Law & Smith, 2001). VV-infected cells were detected using mAb AB1.1 (5 µg/ml) and fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG (diluted 1/100, Jackson Laboratories). The methods for electron microscopy have been described elsewhere recently (Hollinshead et al., 2001).

VV long-range spread

The ability of VV to form a comet-shaped plaque under liquid overlay has been noted for years, but there is no explanation why the virus spreads unidirectionally to produce the comet tail. VV strains, such as IHD-J, that make high levels of EEV form comet-shaped plaques under liquid overlay (Fig. 1b), but round plaques under semi-solid overlay (Fig. 1c). In contrast, VV strains, such as WR, yielding less EEV form round, defined plaques under liquid overlay (Fig. 1a). The formation of comet-shaped plaques can be blocked by Ab raised against VV infection (Fig. 1d) or EEV proteins (Engelstad et al., 1992; Galmiche et al., 1999), resulting in a plaque phenotype similar to that produced under semi-solid overlay (Appleyard et al., 1971; Payne, 1980). Immunofluorescent microscopy showed that comet-shaped plaques are formed by the distant spread of EEV (Fig. 1e) rather than by the migration of infected cells because infected cells moved 0.2 mm/24 h (Sanderson et al., 1998b; Sanderson & Smith, 1999) far less than the 4.3 mm from the centre of the plaque to the end of the comet tail shown in Fig. 1(e). To examine why EEV spreads unidirectionally, the possible role of gravity was investigated. Cells infected with VV were incubated with the tissue culture plate tilted 10°. Surprisingly, instead of having comets pointing in different directions as when incubated flat (Fig. 1f), all the comet tails became parallel and moved uphill (Fig. 1g). This suggested that EEV was dispersed by an upward convection current instead of gravity. Such currents might be generated by a temperature gradient resulting from evaporation or fluctuation of temperature inside the incubator.

Fig. 1. VV plaque phenotypes. Plaques formed on RK₁₃ cells infected with VV strains, WR (a) and IHD-J (b) under liquid overlay, IHD-J under CMC overlay (c) and liquid overlay with anti-VV Ab Rb-WR2 diluted at 1/100 (d). Plaques were stained 44 h p.i. (e) Indirect immunofluorescent staining of a VV IHD-J plaque on RK₁₃ cells under liquid overlay 24 h p.i. Bar, 500 µm. (f) VV strain IHD-J plaques on RK₁₃ cells under liquid overlay (44 h p.i.) with the plate incubated flat or (g) tilted 10°. The arrow points uphill, the direction of virus spread.

Previously, anti-EEV Ab was reported to inhibit EEV release from cells (Ichihashi, 1996; Vanderplasschen et al., 1997) and to neutralize EEV in solution (Boulter & Appleyard, 1973; Galmiche et al., 1999; Law & Smith, 2001). To investigate if a single IEV or EEV protein was essential for this inhibition, we studied whether VV mutants lacking individual IEV- and EEV-specific genes would escape inhibition of virus dissemination (formation of secondary plaques) in the presence of Rb-WR2, a rabbit Ab raised by repeated immunization with live VV WR (Law & Smith, 2001). This Ab reacted strongly with IMV proteins and also moderately with EEV proteins in immunoblot (see supplementary data). In assays using Rb-WR2 (diluted 1/100) and BS-C-1 cells, which gave clear plaques for all mutants, none of the mutants escaped inhibition of secondary plaque formation (data not shown). This is consistent with a recent study showing that EEV made by wild-type and each mutant virus were neutralized by Ab in solution (Law & Smith, 2001).

VV cell-to-cell spread

Ab Rb-WR2 not only inhibited the long-range spread of all the mutants but also caused significant reduction in the plaque size of some of the mutants. These observations suggest there are at least two mechanisms of virus cell-to-cell spread that are either sensitive or resistant to Ab. This was studied further by comparing the plaque size in the presence of Ab against different forms of VV and under a semi-solid overlay so as to measure virus cell-to-cell spread only. Fig. 2(a) shows the plaques formed by the mutants 10 days p.i. in the presence of mAb 2D5 (diluted 1/500), Rb-WR2 or -B5R antisera (diluted 1/100), and their sizes are quantified in Fig. 2(b–f). MAb 2D5 is an IMV-neutralizing Ab (Ichihashi, 1996), Rb-WR2 inhibits both IMV and EEV (Law & Smith, 2001) whereas -B5R neutralizes only EEV (Galmiche et al., 1999; Law & Smith, 2001).

Fig. 2. (a) Cell-to-cell spread of VV in the presence of Ab. BS-C-1 cells were infected by WR or vA56R at 20 p.f.u. per dish and other mutants at 75 p.f.u. per well without Ab (control), or with mAb 2D5 diluted 1/500, Rb-WR2 or -B5R antiserum diluted at 1/100. Cells were overlaid with CMC and stained with crystal violet at 10 days p.i. (b–f) Quantification of plaque sizes in the presence or absence of Ab. (b) The diameters of plaques made by the different viruses in (a) were measured. The data represent the mean ąSEM of the diameter of six plaques selected randomly. An asterisk indicates that vF13L plaques in the presence of Rb-WR2 were too faint to measure accurately. (c) The relative plaque size (diameter) of each mutant compared to WR is plotted. (d–f) The relative size of plaques formed by the mutants in the presence of mAb 2D5 (d), Rb-WR2 (e) or -B5R (f), compared to control are shown.

In the same conditions, plaques formed by all the mutants in the absence of Ab had at least a threefold reduced diameter compared with WR, except for those formed by vA56R that were equivalent (Fig. 2c). For the other mutants, the next largest were made by vA36R and vB5R, followed by vA34R and vA33R, and the smallest plaques were made by vF12L and vF13L.

Addition of mAb 2D5 had little effect on the plaque size of the mutants, although vF13L was reduced slightly more than the others (Fig. 2d). Possibly, IMV was released late during infection with vF13L and therefore its plaques were affected by mAb 2D5. These data indicate that IMV is not sufficient for VV cell-to-cell spread.

In contrast to mAb 2D5, the antiserum Rb-WR2 raised against a live infection had varying effects on the mutants (Fig. 2e). Most strikingly, plaques were not made by vA33R in the presence of Rb-WR2. Rb-WR2 also reduced the plaque size of WR, vA34R, vA56R and vB5R, but not vA36R or vF12L. Very faint plaques were seen with vF13L, but these were difficult to quantify. These results allow the viruses to be divided into three groups: complete blockage of virus spread (vA33R), partial blockage (WR, vA34R, vA56R, vB5R and vF13L) and no blockage (vA36R and vF12L). Since anti-IMV Ab did not affect cell-to-cell spread, the inhibition should be mediated by the anti-EEV Ab present in Rb-WR2. The blockage of plaque formation by vA33R using Rb-WR2 is not due to an increase in Ab sensitivity of the mutant because EEV made by vA33R showed similar sensitivity to neutralization by antibody (Law & Smith, 2001) and further analyses using a range of dilutions of Ab confirmed that the sensitivity was the same as wild-type (Fig. 3). On the other hand, EEV of vB5R is less sensitive to Rb-WR2 because of the absence of major virus-neutralizing epitopes on the B5R protein (Law & Smith, 2001). Interestingly, antibody enhancement was observed in both vA33R and WR at similar Ab dilutions.

Fig. 3. Neutralization of EEV by Rb-WR2 Ab. EEV of the indicated viruses was grown in RK₁₃ cells and incubated with Rb-WR2 IgG, which had been diluted serially in threefold steps, for 1 h at 37 °C in the presence of mAb 2D5 (diluted 1/2000). After washing, the cells were overlaid with 1 % low-melting agarose in DMEM/2 % and the plaques were counted at 40 h p.i. for WR, 4 days p.i. for vB5R and 5 days p.i. for vA33R. Data point '0' represents a concentration of 100 µg/ml of Rb-WR2 IgG. Two independent experiments gave similar results and each data point represents the mean ąSEM of duplicate measurements in one experiment.

The effect of -B5R also varied with different mutants (Fig. 3f). As expected, it had no effect on vB5R. Similar to Rb-WR2, it did not affect vA36R and vF12L, but inhibited WR, vA56R and vA34R slightly (~20 %). Surprisingly, it did not affect vF13L at all, while tiny faint plaques were formed by vA33R.

Does abortive infection by vA33R occur in the presence of Rb-WR2?

The complete blockage of plaque formation by vA33R using Rb-WR2 Ab was unexpected. To investigate whether this blockage was a result of an Ab-sensitive spreading mechanism or a post-binding neutralization, we compared vA33R- and vA34R-infected cells by immunofluorescent microscopy (Fig. 4). Without Ab (Fig. 4a, d) or with mAb 2D5 (Fig. 4b, e), extensive cytopathic effect, such as cell rounding, cell flattening and projection formation (Sanderson et al., 1998b), was found in cells infected by either mutant. At 72 h p.i., the infected foci were the same size with or without mAb 2D5 and were about 0.24 mm and 0.38 mm for vA33R and vA34R, respectively. In the presence of Rb-WR2, cells infected with vA34R developed similar cytopathic effect, although virus spread was restricted by the Ab (Fig. 4f). With vA33R in the presence of Rb-WR2 (Fig. 4c), very few infected cells were found in all the infected foci examined. In addition, the infected cells did not develop extensive cytopathic effect and were not clustered together. The individual infected cells within these atypical foci were unlikely to have derived from several separate infections because only a few plaque-forming units were used to inoculate the cell monolayer. Possibly, virus-induced cell movement might have caused the infected cells to disconnect from each other after the spread of the virus.

Fig. 4. Immunofluorescent microscopy of vA33R- and vA34R-infected cells. BS-C-1 cells growing on 22 mm glass coverslips were infected with vA33R or vA34R at 20 p.f.u. per coverslip for 2 h at 37 °C. The cells were covered with semi-solid overlay with or without Ab and were incubated for 72 h. Infected cells were detected using mAb AB1.1 followed by FITC-conjugated goat -mouse IgG secondary Ab. (a, d) Control; (b, e) mAb 2D5; (c, f) Rb-WR2. Bar, 100 µm.

These results demonstrated that, in the presence of Rb-WR2, productive virus infection occurred in vA33R-infected cells but virus spread was greatly diminished.

Is IMV wrapping important for the Ab-resistant spreading mechanism?

vA33R and vA34R make higher levels of EEV (Table 1) despite the wrapping of IMV being either incomplete (Roper et al., 1998) or inefficient (Duncan & Smith, 1992; Wolffe et al., 1997). Therefore, we investigated whether an alternative mechanism for CEV and EEV formation, such as budding of IMV from the cell surface, occurred with vA33R and resulted in the increased susceptibility to Rb-WR2.

The wrapping of IMV to form IEV can be inhibited pharmacologically by IMCBH (Kato et al., 1969; Payne & Kristenson, 1979; Hiller et al., 1981), which targets the F13L protein (Hiller et al., 1981; Schmutz et al., 1991). Fig. 5(a) shows the plaques formed in the presence of IMCBH at 5 days p.i. As expected, IMCBH had no effect on vF13L due to the absence of the target protein, while WR and vA56R produced plaques similar in size to those of vF13L. All other mutants were inhibited severely and only tiny infected foci were seen.

Fig. 5. (a) Cell-to-cell spread of VV mutants in the presence of IMCBH. BS-C-1 cells were infected by the indicated viruses at 75 p.f.u. per well for 2 h at 37 °C and the cells were covered with semi-solid overlay with (+) or without (–) IMCBH (10 µg/ml). The plaques were stained 5 days p.i. (b) EEV production in the presence of IMCBH. EEV of the indicated viruses was grown in RK₁₃ cells with or without IMCBH (10 µg/ml) for 18 h and was titrated on BS-C-1 cells using mAb 2D5 (final dilution 1/2000). Two independent experiments gave very similar results. Each data point represents the mean ąSEM of triplicate measurements in one experiment.

The levels of EEV produced by these mutants in the presence of IMCBH were also studied (Fig. 5b). To measure EEV infectivity, fresh EEV was mixed with mAb 2D5 (diluted 1/2000) to neutralize contaminating IMV (Law & Smith, 2001). Without IMCBH (open bars), vA33R and vA34R produced 2.6 and 4.1 times more EEV than WR while vA36R, vB5R, vF12L and vF13L produced 2.5, 5.5, 2.9 and 5.3 times less EEV than WR, respectively, in broad agreement with published data (Table 1). The level of EEV made by vA56R is similar to WR (G. L. Smith, unpublished data). Here, we found that vA56R produced 1.8-fold less EEV than WR. In most reports, except vB5R and vF12L (Mathew et al., 1998; Zhang et al., 2000), EEV infectivity was measured as the total present in the supernatant. Here, only the infectivity of intact EEV was measured and this might account for the small variations observed.

In the presence of IMCBH (Fig. 5b, black bars), the levels of EEV made by vA33R and vA34R were reduced drastically, suggesting EEV were formed by the normal IMCBH-sensitive pathway. Similar results were noted for vB5R and vF12L. Some EEV were still made by WR, vA56R and vF13L, consistent with the observations on virus cell-to-cell spread above. Surprisingly, a significant level of EEV was made by vA36R in the presence of IMCBH, despite plaque formation being inhibited by this drug (Fig. 5a). These data may suggest that the inhibition of EEV production by IMCBH is also partially dependent on the presence of A36R in addition to F13L.

The formation of IEV by vA33R and vA34R was investigated further by electron microscopy. In RK₁₃ cells infected with VV strain IHD-J most IMV wrapping occurred before 20 h p.i. (Payne & Kristenson, 1979). Since vA33R and vA34R make enhanced levels of EEV at 18–24 h p.i. (Table 1), the deletion of A33R or A34R might have accelerated the kinetics for IMV wrapping and EEV release. Therefore, RK₁₃ cells infected with vA33R, vA34R, WR and vRevA34R (the revertant virus of vA34R) were analysed at 6 h and 8 h p.i. (Fig. 6) and intracellular virions (IMV and IEV) were quantified (Table 2). Fully wrapped IEV particles were found in cells infected with both mutants (Fig. 6) but, compared to the other viruses, lower levels of IEV were found in vA34R-infected cells at both 6 and 8 h p.i. This suggested that either IEV formation is less efficient or IEV dissemination is quicker. In all samples, we found no evidence that EEV were generated by budding and conclude that wrapping of IMV is necessary for the formation of CEV/EEV by vA33R and vA34R. Therefore, the sensitivity to Ab of vA33R cell-to-cell spread is unlikely to be caused by an alternative pathway of CEV/EEV formation.

Fig. 6. Electron microscopy of vA33R- and vA34R-infected cells. RK₁₃ cells were infected with WR, vA33R, vA34R or vRevA34R at 10 p.f.u. per cell. The cells were harvested at 6 h and were processed for thin section transmission electron microscopy. Wrapped particles of vA33R (left panels) and vA34R (right panels) are shown. Bars, 200 nm.

Table 2. Intracellular virions (IMV and IEV) in VV-infected cells

This paper describes an investigation of the mechanisms of cell-to-cell and long-range spread of VV in cell culture and the effects of antibody against IMV or EEV on these processes.

Long-range virus dissemination was analysed in cell culture by the formation of comet-shaped plaques and was shown to be mediated by EEV that is probably dispersed by convection currents to infect distant cells. By tilting the culture dish 10° all comets became parallel and went uphill. In vivo, EEV is released early in infection and is important for systemic spread (Payne, 1980).

The mechanism of cell-to-cell spread was investigated by a direct comparison of the plaques formed by all mutants lacking individual IEV or EEV proteins and the effects of anti-IMV and -EEV Ab. The plaque sizes of all mutants, except vA56R, was reduced at least threefold in diameter compared to WR. vA56R formed syncytia (Ichihashi & Dales, 1971; Sanderson et al., 1998a) but had similar sensitivity to Ab compared with WR, indicating that syncytium formation did not contribute to virus spread. Syncytium formation by vA56R occurs late during infection and might occur between infected cells rather than between infected and uninfected cells.

After vA56R, the next largest plaque is made by vA36R followed by vB5R, vA34R, vA33R, v12L and vF13L. All mutants make near normal levels of IMV (Table 1), demonstrating that IMV is insufficient for plaque formation. Plaque size is also not determined by whether or not IMV are wrapped to form IEV because mutants that can (vA36R) or cannot (vF13L) make IEV each form a small plaque. The levels of CEV and EEV also are not important for plaque size since viruses IHD-J and WR produce very different levels of CEV but make similar plaque sizes, and viruses with either enhanced (vA34R) or reduced (vB5R) EEV levels each make small plaques (Table 1). The major factor determining plaque size is the ability to make actin tails that mediate efficient cell-to-cell spread (Table 1).

The mechanism(s) of VV spread were dissected further using neutralizing Ab specific to IMV or EEV. A reduction in plaque size suggests that the virus spreads from cell to cell in a pathway that is exposed to Ab, whereas, if the plaque size is unchanged the virus spreads in a pathway that is protected from Ab. Plaques made by all viruses were largely unaffected by anti-IMV Ab (Fig. 2d). Therefore, cell-to-cell spread must involve the enveloped virions CEV/EEV, and consistent with this, anti-EEV Ab had varying effects on the different mutants (Fig. 2e, f). In the presence of Rb-WR2 or -B5R Ab, WR and vA56R plaque sizes were inhibited to a similar extent (Fig. 2e), but the plaques were still bigger than those made by other mutants, even in the absence of the Ab (Fig. 2b). This reiterated the importance of actin tails for efficient cell-to-cell spread and showed that this mechanism of spread is resistant to Ab. Resistance to Ab was also observed in other mutants unable to make actin tails, e.g. vA36R and vF12L, and therefore an actin tail-independent and Ab-resistant pathway exists. Both A36R and F12L proteins are IEV-specific and not exposed to Ab, and so these results are unlikely to be due to the removal of neutralizing targets on CEV/EEV. Rb-WR2 reduced the size of plaques formed by vA34R and vB5R and only very faint plaques were formed by vF13L, implying that these viruses used both Ab-sensitive and Ab-resistant pathways for their spread. Most dramatic was the abrogation of vA33R plaque formation by Rb-WR2, showing this mutant spreads entirely via an Ab-sensitive pathway.

Anti-B5R Ab affected vA33R the most among all the mutants, showing B5R is a target in the Ab-sensitive pathway. However, other virus antigen(s) must be involved because -B5R did not block vA33R completely, affected vA34R only marginally and had no effect on vF13L (Fig. 2f). IMV wrapping by vF13L is inhibited severely and therefore this mutant makes very low levels of CEV and EEV (Table 1). vF13L was also slightly more sensitive to mAb 2D5 than other mutants (Fig. 2c), suggesting that IMV might contribute to some levels of virus spread in this mutant. This might account for the resistance to -B5R. Cell-to-cell spread of vA34R was fairly resistant to -B5R despite vA34R making 19- to 25-fold higher levels of EEV. This suggested A34R is required for the Ab inhibition via B5R. A34R may have a role in the conformation, recruitment or function of the B5R protein.

In summary, these data show that VV uses a combination of mechanisms to spread between cells. The mechanisms can be divided into (i) actin tail-dependent Ab-resistant pathway (WR and vA56R); (ii) actin tail-independent Ab-resistant pathway (vA36R and vF12L); and (iii) Ab-sensitive pathway (vA33R). Actin tails have a major role in the Ab-resistant pathway. This is probably because actin tails push the virions into the neighbouring cells directly without exposing the virions to Ab. WR and vA56R spread by a combination of these three pathways whereas vA34R, vB5R and vF13L use pathways (ii) and (iii).

The Ab-susceptibility of vA33R is analogous to HSV-1 gE and gI deletion mutants. HSV-1 entry kinetics and replication were not affected by deletion of either gene; however, cell-to-cell spread of these mutants was impaired (Dingwell et al., 1994), and neutralizing Ab reduced the yields of the mutants but not wild-type virus. HSV gE and gI form a complex (Johnson et al., 1988) that accumulates at sites of cell–cell contact, possibly by interacting with junctional components (Dingwell & Johnson, 1998), and which may mediate HSV transfer across the cell junctions. Similarly, varicellar-zoster virus gE expression in polarized epithelial cells altered the F-actin organization and accelerated the formation of tight junctions between cells (Mo et al., 2000). The VV A33R protein expressed by Semliki forest virus or VV accumulated on microvillus-like cell surface projections (Lorenzo et al., 2000). A33R might aid virus spread through cell junctions by interacting with junctional proteins in a similar fashion to HSV gE–gI. A33R might also interact with surface molecules of neighbouring cells to facilitate cell–cell contacts for virus spread. Deletion of A33R could disrupt these cell–cell interactions and permit vA33R to spread only in an Ab-sensitive pathway. Consistent with this proposal, the passive transfer of an anti-A33R mAb, and immunization with A33R DNA (Hooper et al., 2000) or recombinant A33R protein (Galmiche et al., 1999), protected mice from VV challenge. However, neither the anti-A33R mAb nor Ab raised against recombinant A33R proteins neutralized EEV in vitro.

The production of several infectious forms of VV is explained by these virions having different roles in the virus life-cycle. IMV is highly immunogenic (see supplementary data) (Law & Smith, 2001) and susceptible to neutralization by complement (Vanderplasschen et al., 1998b), and therefore is poorly suited to virus spread within a host. It would be advantageous for the majority of VV infectivity (IMV) to be retained and protected within cells from Ab and complement. However, IMV particles are physically robust and are well suited for dissemination between hosts. To aid virus dissemination within a host, VV has exploited several features of the cell biology. It uses cellular membranes to wrap IMV particles and by the acquisition of host complement factors protects EEV particles from destruction by complement (Vanderplasschen et al., 1998b), and uses virus-encoded proteins in the EEV outer envelope to bind to different receptors from IMV (Vanderplasschen & Smith, 1997) and to enhance the range of cell types that may be infected.

The role of IEV is less characterized and might appear unnecessary, for to make CEV or EEV an IMV might bud through the plasma membrane. Indeed, limited budding occurs late during infection in VV strain IHD-W (Tsutsui, 1983) and in fowlpox virus (Boulanger et al., 2000). The apparent inefficiency of IMV wrapping but enhanced EEV production by mutants vA33R and vA34R (Duncan & Smith, 1992; McIntosh & Smith, 1996; Roper et al., 1998) further questioned the significance of IEV in the VV life-cycle. However, by pharmacological and microscopic approaches, we demonstrated here that wrapping not only occurs with these mutants but is an essential step for both mutants (Figs 5 and 6). Two functions for IEV are proposed. First, asymmetric distribution of the A36R protein enables the unidirectional polymerization of actin after the outmost IEV membrane has fused with the plasma membrane (van Eijl et al., 2000). This could not easily occur via budding. Second, both IMV and IEV utilize microtubules for intracellular movement (Sanderson et al., 2000; Hollinshead et al., 2001; Ward & Moss, 2001). The wrapping of IMV by additional membranes places different proteins on the IMV and IEV surface, and consequently these virions may have different interactions with microtubule components enabling movement towards (IMV) or away from (IEV) the site of wrapping.

Finally, we propose an explanation for the existence of CEV. CEV is structurally indistinguishable from EEV, but with most VV strains many enveloped virions are retained on the cell surface rather than being released. This appears curious since many viruses enhance virus release by down-regulating cell receptors. However, the efficient cell-to-cell dissemination of virus requires actin tail formation from the cell surface, and so enveloped virions need to be retained long enough at the surface to promote this activity. A34R and B5R are involved in the retention of CEV on the cell surface. The deletion of the entire B5R gene or fusion with the VV A56R extracellular domain reduced wrapping and EEV release (Engelstad & Smith, 1993; Wolffe et al., 1993; Mathew et al., 2001) while deletion of any of the short consensus repeat domains of B5R enabled wrapping and increased EEV release (Herrera et al., 1998; Mathew et al., 1998). The deletion or mutation of A34R enhanced the release of EEV (Blasco et al., 1993; McIntosh & Smith, 1996) and reduced wrapping compared to WR and vRevA34R (Table 2). B5R interacts with A34R (Röttger et al., 1999) and, interestingly, -B5R Ab did not inhibit the cell-to-cell spread of vA34R (Fig. 2f). A34R and B5R may work in concert to affect EEV release.

In summary, VV benefits from having four different virus forms, IMV, IEV, CEV and EEV, for efficient cell-to-cell spread (CEV and actin tails) and long-range spread (EEV) within the host, and reserving the majority of infectivity (IMV) for transmission between hosts.

We thank Michael Hollinshead for help with microscopy, Alain Vanderplasschen for discussion, and Riccardo Wittek, Yasuo Ichihashi and Bernard Moss for Ab, viruses and IMCBH. M.L. was funded by the Croucher Foundation Scholarship, Hong Kong, China. This work was supported by Programme Grant P8901790 from the UK Medical Research Council and an equipment grant from The Wellcome Trust. G.L.S. is a Wellcome Trust Principal Research Fellow.

† Present address: Wright–Fleming Institute, Imperial College School of Medicine, St Mary's Campus, Norfolk Place, London W2 1PG, UK.

Alcamí, A. & Smith, G. L. (1992). A soluble receptor for interleukin-1 beta encoded by vaccinia virus: a novel mechanism of virus modulation of the host response to infection. Cell 71, 153–167.

Appleyard, G. & Andrews, C. (1974). Neutralizing activities of antisera to poxvirus soluble antigens. Journal of General Virology 23, 197–200.

Appleyard, G., Hapel, A. J. & Boulter, E. A. (1971). An antigenic difference between intracellular and extracellular rabbitpox virus. Journal of General Virology 13, 9–17.

Blasco, R. & Moss, B. (1991). Extracellular vaccinia virus formation and cell-to-cell virus transmission are prevented by deletion of the gene encoding the 37,000-Dalton outer envelope protein. Journal of Virology 65, 5910–5920.

Blasco, R. & Moss, B. (1992). Role of cell-associated enveloped vaccinia virus in cell-to-cell spread. Journal of Virology 66, 4170–4179; erratum 5703–5704.

Blasco, R., Sisler, J. R. & Moss, B. (1993). Dissociation of progeny vaccinia virus from the cell membrane is regulated by a viral envelope glycoprotein: effect of a point mutation in the lectin homology domain of the A34R gene. Journal of Virology 67, 3319–3325.

Boulanger, D., Smith, T. & Skinner, M. A. (2000). Morphogenesis and release of fowlpox virus. Journal of General Virology 81, 675–687.

Boulter, E. A. & Appleyard, G. (1973). Differences between extracellular and intracellular forms of poxvirus and their implications. Progress in Medical Virology 16, 86–108.

Boulter, E. A., Zwartouw, H. T., Titmuss, D. H. & Maber, H. B. (1971). The nature of the immune state produced by inactivated vaccinia virus in rabbits. American Journal of Epidemiology 94, 612–620.

Dingwell, K. S. & Johnson, D. C. (1998). The herpes simplex virus gE–gI complex facilitates cell-to-cell spread and binds to components of cell junctions. Journal of Virology 72, 8933–8942.

Dingwell, K. S., Brunetti, C. R., Hendricks, R. L., Tang, Q., Tang, M., Rainbow, A. J. & Johnson, D. C. (1994). Herpes simplex virus glycoproteins E and I facilitate cell-to-cell spread in vivo and across junctions of cultured cells. Journal of Virology 68, 834–845.

Duncan, S. A. & Smith, G. L. (1992). Identification and characterization of an extracellular envelope glycoprotein affecting vaccinia virus egress. Journal of Virology 66, 1610–1621.

Engelstad, M. & Smith, G. L. (1993). The vaccinia virus 42-kDa envelope protein is required for the envelopment and egress of extracellular virus and for virus virulence. Virology 194, 627–637.

Engelstad, M., Howard, S. T. & Smith, G. L. (1992). A constitutively expressed vaccinia gene encodes a 42-kDa glycoprotein related to complement control factors that forms part of the extracellular virus envelope. Virology 188, 801–810.

Firsching, R., Buchholz, C. J., Schneider, U., Cattaneo, R., ter Meulen, V. & Schneider-Schaulies, J. (1999). Measles virus spread by cell–cell contacts: uncoupling of contact-mediated receptor (CD46) downregulation from virus uptake. Journal of Virology 73, 5265–5273; erratum 10556.

Flexner, C., Hugin, A. & Moss, B. (1987). Prevention of vaccinia virus infection in immunodeficient mice by vector-directed IL-2 expression. Nature 330, 259–262.

Galmiche, M. C., Goenaga, J., Wittek, R. & Rindisbacher, L. (1999). Neutralizing and protective antibodies directed against vaccinia virus envelope antigens. Virology 254, 71–80.

Herrera, E., Lorenzo, M. M., Blasco, R. & Isaacs, S. N. (1998). Functional analysis of vaccinia virus B5R protein: essential role in virus envelopment is independent of a large portion of the extracellular domain. Journal of Virology 72, 294–302.

Hiller, G. & Weber, K. (1985). Golgi-derived membranes that contain an acylated viral polypeptide are used for vaccinia virus envelopment. Journal of Virology 55, 651–659.

Hiller, G., Eibl, H. & Weber, K. (1981). Characterization of intracellular and extracellular vaccinia virus variants: N₁-isonicotinoyl-N₂-3-methyl-4-chlorobenzoylhydrazine interferes with cytoplasmic virus dissemination and release. Journal of Virology 39, 903–913.

Hirt, P., Hiller, G. & Wittek, R. (1986). Localization and fine structure of a vaccinia virus gene encoding an envelope antigen. Journal of Virology 58, 757–764.

Hollinshead, M., Rodger, G., van Eijl, H., Law, M., Hollinshead, R., Vaux, D. J. T. & Smith, G. L. (2001). Vaccinia virus utilizes microtubules for movement to the cell surface. Journal of Cell Biology 154, 389–402.

Hooper, J. W., Custer, D. M., Schmaljohn, C. S. & Schmaljohn, A. L. (2000). DNA vaccination with vaccinia virus L1R and A33R genes protects mice against a lethal poxvirus challenge. Virology 266, 329–339.

Ichihashi, Y. (1996). Extracellular enveloped vaccinia virus escapes neutralization. Virology 217, 478–485.

Ichihashi, Y. & Dales, S. (1971). Biogenesis of poxviruses: interrelationship between hemagglutinin production and polykaryocytosis. Virology 46, 533–543.

Ichihashi, Y. & Oie, M. (1980). Adsorption and penetration of the trypsinized vaccinia virion. Virology 101, 50–60.

Ichihashi, Y. & Oie, M. (1996). Neutralizing epitope on penetration protein of vaccinia virus. Virology 220, 491–494.

Isaacs, S. N., Wolffe, E. J., Payne, L. G. & Moss, B. (1992). Characterization of a vaccinia virus-encoded 42-kilodalton class I membrane glycoprotein component of the extracellular virus envelope. Journal of Virology 66, 7217–7224.

Johnson, D. C., Frame, M. C., Ligas, M. W., Cross, A. M. & Stow, N. D. (1988). Herpes simplex virus immunoglobulin G Fc receptor activity depends on a complex of two viral glycoproteins, gE and gI. Journal of Virology 62, 1347–1354.

Kato, N., Eggers, H. J. & Rolly, H. (1969). Inhibition of release of vaccinia virus by N₁-isonicotinoly-N₂-3-methyl-4-chlorobenzoylhydrazine. Journal of Experimental Medicine 129, 795–808.

Krijnse-Locker, J., Kuehn, A., Schleich, S., Rutter, G., Hohenberg, H., Wepf, R. & Griffiths, G. (2000). Entry of the two infectious forms of vaccinia virus at the plasma membrane is signaling-dependent for the IMV but not the EEV. Molecular Biology of the Cell 11, 2497–2511.

Law, M. & Smith, G. L. (2001). Antibody neutralization of the extracellular enveloped form of vaccinia virus. Virology 280, 132–142.

Lorenzo, M. M., Galindo, I., Griffiths, G. & Blasco, R. (2000). Intracellular localization of vaccinia virus extracellular enveloped virus envelope proteins individually expressed using a Semliki Forest virus replicon. Journal of Virology 74, 10535–10550.

McIntosh, A. A. & Smith, G. L. (1996). Vaccinia virus glycoprotein A34R is required for infectivity of extracellular enveloped virus. Journal of Virology 70, 272–281.

Madeley, C. R. (1968). The immunogenicity of heat-inactivated vaccinia virus in rabbits. Journal of Hygiene 66, 89–107.

Mathew, E., Sanderson, C. M., Hollinshead, M. & Smith, G. L. (1998). The extracellular domain of vaccinia virus protein B5R affects plaque phenotype, extracellular enveloped virus release, and intracellular actin tail formation. Journal of Virology 72, 2429–2438.

Mathew, E. C., Sanderson, C. M., Hollinshead, R. & Smith, G. L. (2001). A mutational analysis of the vaccinia virus B5R protein. Journal of General Virology 82, 1199–1213.

Mo, C., Schneeberger, E. E. & Arvin, A. M. (2000). Glycoprotein E of varicella-zoster virus enhances cell–cell contact in polarized epithelial cells. Journal of Virology 74, 11377–11387.

Moss, B. (1996). Poxviridae: the viruses and their replication. In Fields Virology, 3rd edn, pp. 2637–2671. Edited by B. N. Fields, D. M. Knipe & P. M. Howley. Philadelphia: Lippincott–Raven.

Palese, P., Tobita, K., Ueda, M. & Compans, R. W. (1974). Characterization of temperature sensitive influenza virus mutants defective in neuraminidase. Virology 61, 397–410.

Parkinson, J. E. & Smith, G. L. (1994). Vaccinia virus gene A36R encodes a M_r 43–50 K protein on the surface of extracellular enveloped virus. Virology 204, 376–390.

Payne, L. G. (1978). Polypeptide composition of extracellular enveloped vaccinia virus. Journal of Virology 27, 28–37.

Payne, L. G. (1980). Significance of extracellular enveloped virus in the in vitro and in vivo dissemination of vaccinia. Journal of General Virology 50, 89–100.

Payne, L. G. (1992). Characterization of vaccinia virus glycoproteins by monoclonal antibody precipitation. Virology 187, 251–260.

Payne, L. G. & Norrby, E. (1978). Adsorption and penetration of enveloped and naked vaccinia virus particles. Journal of Virology 27, 19–27.

Payne, L. G. & Kristenson, K. (1979). Mechanism of vaccinia virus release and its specific inhibition by N₁-isonicotinoyl-N₂-3-methyl-4-chlorobenzoylhydrazine. Journal of Virology 32, 614–622.

Piguet, V., Schwartz, O., Le Gall, S. & Trono, D. (1999). The downregulation of CD4 and MHC-I by primate lentiviruses: a paradigm for the modulation of cell surface receptors. Immunological Reviews 168, 51–63.

Roper, R. L., Payne, L. G. & Moss, B. (1996). Extracellular vaccinia virus envelope glycoprotein encoded by the A33R gene. Journal of Virology 70, 3753–3762.

Roper, R. L., Wolffe, E. J., Weisberg, A. & Moss, B. (1998). The envelope protein encoded by the A33R gene is required for formation of actin-containing microvilli and efficient cell-to-cell spread of vaccinia virus. Journal of Virology 72, 4192–4204.

Röttger, S., Frischknecht, F., Reckmann, I., Smith, G. L. & Way, M. (1999). Interactions between vaccinia virus IEV membrane proteins and their roles in IEV assembly and actin tail formation. Journal of Virology 73, 2863–2875.

Sanderson, C. S. & Smith, G. L. (1999). Cell motility and morphology: viruses in control. Expert Reviews in Molecular Medicine http://www-ermm.cbcu.cam.ac.uk/99000629h.htm

Sanderson, C. M., Frischknecht, F., Way, M., Hollinshead, M. & Smith, G. L. (1998a). Roles of vaccinia virus EEV-specific proteins in intracellular actin tail formation and low pH-induced cell–cell fusion. Journal of General Virology 79, 1415–1425.

Sanderson, C. M., Way, M. & Smith, G. L. (1998b). Virus-induced cell motility. Journal of Virology 72, 1235–1243.

Sanderson, C. M., Hollinshead, M. & Smith, G. L. (2000). The vaccinia virus A27L protein is needed for the microtubule-dependent transport of intracellular mature virus particles. Journal of General Virology 81, 47–58.

Schmelz, M., Sodeik, B., Ericsson, M., Wolffe, E. J., Shida, H., Hiller, G. & Griffiths, G. (1994). Assembly of vaccinia virus: the second wrapping cisterna is derived from the trans-Golgi network. Journal of Virology 68, 130–147.

Schmutz, C., Payne, L. G., Gubser, J. & Wittek, R. (1991). A mutation in the gene encoding the vaccinia virus 37,000-M_r protein confers resistance to an inhibitor of virus envelopment and release. Journal of Virology 65, 3435–3442.

Shida, H. (1986). Nucleotide sequence of the vaccinia virus hemagglutinin gene. Virology 150, 451–462.

Shida, H. & Dales, S. (1981). Biogenesis of vaccinia: carbohydrate of the hemagglutinin molecules. Virology 111, 56–72.

Shinkai, K. (1975). Plaque morphology of herpes simplex virus in various cells under liquid overlay as a marker for its type differentiation. Japanese Journal of Microbiology 19, 459–462.

Tooze, J., Hollinshead, M., Reis, B., Radsak, K. & Kern, H. (1993). Progeny vaccinia and human cytomegalovirus particles utilize early endosomal cisternae for their envelopes. European Journal of Cell Biology 60, 163–178.

Tsutsui, K. (1983). Release of vaccinia virus from FL cells infected with the IHD-W strain. Journal of Electron Microscopy 32, 125–140.

Turner, G. S. & Squires, E. J. (1971). Inactivated smallpox vaccine: immunogenicity of inactivated intracellular and extracellular vaccinia virus. Journal of General Virology 13, 19–25.

Ulaeto, D., Grosenbach, D. & Hruby, D. E. (1995). Brefeldin A inhibits vaccinia virus envelopment but does not prevent normal processing and localization of the putative envelopment receptor P37. Journal of General Virology 76, 103–111.

Vanderplasschen, A. & Smith, G. L. (1997). A novel virus binding assay using confocal microscopy: demonstration that the intracellular and extracellular vaccinia virions bind to different cellular receptors. Journal of Virology 71, 4032–4041.

Vanderplasschen, A., Hollinshead, M. & Smith, G. L. (1997). Antibodies against vaccinia virus do not neutralize extracellular enveloped virus but prevent virus release from infected cells and comet formation. Journal of General Virology 78, 2041–2048.

Vanderplasschen, A., Hollinshead, M. & Smith, G. L. (1998a). Intracellular and extracellular vaccinia virions enter cells by different mechanisms. Journal of General Virology 79, 877–887.

Vanderplasschen, A., Mathew, E., Hollinshead, M., Sim, R. B. & Smith, G. L. (1998b). Extracellular enveloped vaccinia virus is resistant to complement because of incorporation of host complement control proteins into its envelope. Proceedings of the National Academy of Sciences, USA 95, 7544–7549.

van Eijl, H., Hollinshead, M. & Smith, G. L. (2000). The vaccinia virus A36R protein is a type Ib membrane protein present on intracellular but not extracellular enveloped virus particles. Virology 271, 26–36.

van Eijl, H., Hollinshead, M., Zhang, W.-H. & Smith, G. L. (2002). The vaccinia virus F12L protein is associated with intracellular enveloped virus particles and is required for their egress to the cell surface. Journal of General Virology 83, 195–207.

Ward, B. M. & Moss, B. (2001). Visualization of intracellular movement of vaccinia virus virions containing a green fluorescent protein–B5R membrane protein chimera. Journal of Virology 75, 4802–4813.

Wolffe, E. J., Isaacs, S. N. & Moss, B. (1993). Deletion of the vaccinia virus B5R gene encoding a 42-kilodalton membrane glycoprotein inhibits extracellular virus envelope formation and dissemination. Journal of Virology 67, 4732–4741; erratum 5709–5711.

Wolffe, E. J., Katz, E., Weisberg, A. & Moss, B. (1997). The A34R glycoprotein gene is required for induction of specialized actin-containing microvilli and efficient cell-to-cell transmission of vaccinia virus. Journal of Virology 71, 3904–3915.

Wolffe, E. J., Weisberg, A. S. & Moss, B. (1998). Role for the vaccinia virus A36R outer envelope protein in the formation of virus-tipped actin-containing microvilli and cell-to-cell virus spread. Virology 244, 20–26.

Zhang, W. H., Wilcock, D. & Smith, G. L. (2000). Vaccinia virus F12L protein is required for actin tail formation, normal plaque size, and virulence. Journal of Virology 74, 11654–11662.

Supplementary data

Immunoblot of purified IMV and EEV using Rb-WR2 Ab. Purified IMV and EEV (10 µg each) were resolved by SDS–PAGE (12 % gel) and transferred to a nitrocellulose membrane. Virus-specific proteins were detected with Rb-WR2 IgG (0.25 µg/ml) and horseradish peroxidase-conjugated goat anti-rabbit IgG (diluted 1/2000, Sigma) using an ECL Western Blotting detection system (Amersham Pharmacia Biotech). Common bands represent the immunodominant IMV and core proteins shared by IMV and EEV. Potential EEV-specific proteins are indicated by arrows.

JGV Direct table of contents

© 2002 SGM

This article is now available in the January 2001 print issue of JGV (vol. 83, 209–222). The complete issue of the journal may be seen in electronic form on JGV Online.