|Home | About | Journals | Submit | Contact Us | Français|
Vaccinia virus (VACV) strain WR can enter cells by a low pH endosomal pathway or direct fusion with the plasma membrane at neutral pH. Here, we compared attachment and entry of five VACV strains in six cell lines and discovered two major patterns. Only WR exhibited pH 5-enhanced rate of entry following neutral pH adsorption to cells, which correlated with sensitivity to bafilomycin A1, an inhibitor of endosomal acidification. Entry of IHD-J, Copenhagen and Elstree strains were neither accelerated by pH 5 treatment nor prevented by bafilomycin A1. Entry of the Wyeth strain, although not augmented by pH 5, was inhibited by bafilomycin A1. WR and Wyeth were both relatively resistant to the negative effects of heparin on entry, whereas the other strains were extremely sensitive due to inhibition of cell binding. The relative sensitivities of individual vaccinia virus strains to heparin correlated inversely with their abilities to bind to and enter glycosaminoglycan-deficient sog9 cells but not other cell lines tested. These results suggested that that IHD-J, Copenhagen and Elstree have a more limited ability than WR and Wyeth to use the low pH endosomal pathway and are more dependent on binding to glycosaminoglycans for cell attachment.
Poxviruses comprise a family of large DNA viruses that replicate entirely in the cytoplasm of animal cells (Moss, 2007). Vaccinia virus (VACV), the prototype poxvirus has a 195 kbp double-stranded DNA genome that encodes nearly 200 proteins. There are two major infectious forms of VACV, the mature virion (MV) and enveloped virion (EV). MVs, which are comprised of a single membrane with more than 20 viral proteins surrounding the virus core, can be released by cell lysis. EVs are formed when a subset of MVs are wrapped by modified trans-Golgi or endosomal cisterna containing additional viral membrane proteins and are transported to the periphery of the cell where they exit through the plasma membrane (Smith et al., 2002). EVs differ from MVs primarily by the presence of an extra membrane that contains at least six unique viral proteins. This outer membrane is disrupted by cellular polyanionic molecules on the cell surface to allow for fusion of the inner MV membrane with the cell and entry of the core (Law et al., 2006).
Virus attachment, activation of fusion proteins, and membrane fusion comprise the typical steps involved in enveloped virus entry (Earp et al., 2005). The majority of studies with VACV have been carried out with the MV form. Attachment of MVs to glycosaminoglycans (GAGs) at the cell surface is mediated by three MV membrane proteins A27, D8 and H3 (Chung et al., 1998; Hsiao et al., 1999; Lin et al., 2000), but MVs can also enter cells in a GAG-independent manner (Carter et al., 2005). Two additional MV membrane proteins, A26 and L1, which bind to the extracellular matrix protein laminin and to a putative unknown protein, respectively, have also been identified (Chiu et al., 2007; Foo et al., 2009).
The entry fusion complex (EFC), comprising eight or more virus-encoded MV membrane proteins, is required for the entry of VACV (Izmailyan et al., 2006; Ojeda et al., 2006a; Ojeda et al., 2006b; Senkevich and Moss, 2005; Senkevich et al., 2005; Senkevich et al., 2004; Townsley et al., 2005a; Townsley et al., 2005b) and is also required for cell-cell fusion triggered by low pH or mutation of the A56 or K2 proteins (Senkevich et al., 2004; Wagenaar and Moss, 2007; Wagenaar et al., 2008). Additionally, the F9 and L1 proteins are essential for virion entry and associate with components of the EFC in non-stoichiometric ratios (Bisht et al., 2008; Brown et al., 2006). An eleventh protein, I2, has also been identified in VACV entry though its association with the EFC has not yet been determined (Nichols et al., 2008). Neither the mechanism of fusion nor the roles of the individual EFC and associated proteins have been elucidated.
Studies carried out with the Western Reserve (WR) strain of VACV have shown that cell entry of MVs can occur via a low-pH endosomal pathway (Townsley et al., 2006) in addition to direct fusion at the plasma membrane (Armstrong et al., 1973; Carter et al., 2005; Chang and Metz, 1976). However, earlier experiments with the IHD-J strain of VACV indicated that MV entry was not inhibited by weak bases suggesting independence of a low pH pathway (Vanderplasschen et al., 1998). Here, we systematically compared the attachment and entry of several VACV strains and found substantial differences with regard to enhancement by low pH, requirement for endosome acidification, inhibition by soluble heparin and binding to cell surface glycosaminoglycans.
Previously, we prepared a recombinant VACV strain WR (WRvFire) that expresses firefly luciferase (luc) regulated by an early-late promoter to measure virus entry (Townsley et al., 2006). Luc activity depends on virus attachment, fusion with the cell membrane, transcription and translation. However, due to the packaging of the transcription system within the virus core and the sensitivity of the assay, a robust response is detected by 1 h after infection. Two types of experiments suggested entry through a low pH endosomal route (Townsley et al., 2006). First, brief exposure of cell-bound virions to pH 5.0 buffer greatly increased luc expression during the first hour. By 24 h, however, luc activity was only slightly higher than in the neutral pH control. In addition, the low pH treatment had only a 1.4-fold enhancing effect on plaque formation. Therefore, the main effect of the brief low pH treatment was to accelerate rather than increase entry. Second, luc expression was inhibited at neutral pH by drugs that prevent endosomal acidification.
To determine whether other VACV strains also use a low pH endosomal route of entry, luc recombinants of the IHD-J (IHD-JvFire), Copenhagen (CopvFire), Wyeth (WyethvFire) and Elstree (ElstreevFire) strains were constructed. In each case, MVs were purified by sucrose gradient centrifugation and titers were determined by plaque assay. The protocol consisted of incubating virus with cells at 4°C for 1 h to allow adsorption at neutral pH, removing unattached virus, exposing the cells to buffer at pH 7.4 or 5.0 for 3 min at 37°C, and then continuing the incubation at 37°C at neutral pH. After 1 h, the cells were lysed and luc activity determined. (We emphasize that in all experiments in this paper, whether explicitly stated or not, the adsorption and 1 h incubation prior to measuring luc were at neutral pH). Without post-absorption low pH exposure, luc was highest after infection with IHD-J and Elstree (Fig. 1A). However, after low pH treatment, luc expression by WR was similar to that of IHD-J and Elstree (Fig. 1A). Although low pH greatly accelerated entry of WR, it had little effect on IHD-J, Copenhagen, Wyeth or Elstree (Fig. 1B). A longer experiment with IHD-J demonstrated superimposable increases in luc activity with or without low pH exposure at all time points reaching a maximum at about 10 h (data not shown). Thus, WR appeared to be exceptional, relative to the other VACV strains tested, with regard to low pH acceleration of entry.
Inhibition of endosomal acidification by a specific inhibitor of the vacuolar H+-ATPase, bafilomycin A1, reduces luc expression by WR (Townsley et al., 2006). Bafilomycin A1 reduced entry of WR (Fig. 2A) and Wyeth (Fig 2D) strains by about 50% at neutral pH in BS-C-1 cells, but had little or no effect on IHD-J (Fig 2B), Copenhagen (Fig 2C) and Elstree (Fig 2E). As previously shown (Townsley et al., 2006) and confirmed here, low pH exposure largely annulled the effects of bafilomycin A1 by inducing entry at the plasma membrane.
WR exhibits two low pH activation steps: pH 5.0 pretreatment before adsorption stably activates free virions for accelerated entry upon addition to cells at neutral pH without altering the titer determined by plaque assay, yet prevention of endosomal acidification by bafilomycin A1 is still inhibitory (Townsley and Moss, 2007). Compared to the other strains tested, only WR exhibited accelerated luc expression after exposure to low-pH prior to adsorption to cells at neutral pH (Fig. 3). Luc expression by the other strains was either unaffected or slightly reduced. The differences in low pH activation and the effects of bafilomycin A1 suggested strain variation in the mode of entry of different VACV strains.
The above experiments were carried out with BS-C-1 cells. Employing WR and IHD-J as representatives of strains that differ in their use of low pH mechanisms, we investigated entry in several additional cell lines. Monkey kidney cells (BS-C-1 and Vero), rabbit kidney cells (RK-13), marsupial potoroo kidney cells (PtK2), primary human epidermal keratinocytes (HEKn) and a mutant mouse L cell line defective in biosynthesis of GAGs (sog9) were incubated with virus at neutral pH to permit adsorption and entry was measured with and without subsequent exposure to low pH. Without pH 5 exposure, WR-induced higher luc expression in HEKn and PtK2 cells than the other cell lines (Fig. 4A). However, pH 5-treatment enhanced WR entry to the greatest extent in BS-C-1, RK-13 and Vero cells and to lesser extents in the other cell lines (Fig. 4A, C). Without low pH exposure, IHD-J induced high luc expression in all cell lines except sog9 (Fig. 4B). Again, there was no low pH enhancement of IHD-J in BS-C-1 cells; however, small enhancements were found in some of the other cells (Fig. 4B, D). These results suggested that the modes of VACV entry are affected by cell type as well as virus strain.
We considered that if cell surface GAGs are important for IHD-J entry, as suggested by the relative non-permissiveness of sog9 cells, then entry into other cells should be inhibited by soluble GAGs. To test this idea, IHD-J was incubated with heparin, heparan sulfate, chondroitin sulfate or their combinations prior to adsorption on BS-C-1 cells. At neutral pH conditions, 50 μg/ml of heparin and heparan sulfate inhibited entry by 90 and 60%, respectively, whereas chondroitin sulfate had no effect (Fig 5B). There was no discernible synergistic effect of combinations (Fig. 5B). Under the same conditions, WR was less inhibited by these GAGs (Fig. 5A). In particular, heparin reduced WR entry by only about 20% at neutral pH (Fig. 5A). The sensitivity of WR to heparin was increased when the cells were subjected to brief low pH treatment following neutral pH adsorption (Fig. 5C, E), whereas the sensitivity of IHD-J was unaltered (Fig. 5D, F). This result suggested that heparin inhibits WR entry through the plasma membrane, which occurs after low pH treatment.
The above experiments were carried out at 50 μg/ml concentrations of GAGs. To further compare the effects of heparin on IHD-J and WR, we tested a range of concentrations. A 90% reduction of IHD-J entry occurred at only 1 μg/ml, regardless of whether the cells were exposed to low pH following adsorption (Fig. 6B). In the absence of low pH exposure, heparin had no effect on WR entry (Fig. 6A). Brief low pH treatment enhanced the sensitivity of WR for heparin but a concentration of ~50 μg/ml was required to reduce entry of WR by 50% even under these conditions (Fig. 6A).
We also tested the heparin sensitivities of Copenhagen, Wyeth and Elstree strains of VACV. Entry of Wyeth, like WR, was not strongly inhibited by heparin; however, Copenhagen and Elstree like IHD-J were inhibited by ~90% compared to controls (Fig. 7). Thus, there seemed to be a correlation between strains resistant to bafilomycin A1 and sensitive to heparin.
Soluble laminin can reduce the ability of WR to bind and infect BS-C-40 cells (a derivative of BS-C-1 cells that is propagated at 40°C) (Chiu et al., 2007). To confirm and extend these results, we incubated laminin with WR and IHD-J MVs and determined luc expression at 1 h as a measure of entry. Luc expression by WR was reduced by almost 50% whether or not the cells were exposed to low pH following virus adsorption (Fig. 8A, C) whereas IHD-J expression was reduced about 25% (Fig. 8B, D). We also examined the effects of heparin as a control in these experiments and the results were similar to those in the previous section i.e. only the low pH enhanced entry of WR was reduced (Fig. 8A, C), whereas IHD-J entry was drastically reduced after neutral or low pH treatments (Fig. 8B, D). Heparin and laminin had additive effects on low pH enhanced WR entry (Fig. 8A, C) but not on IHD-J (Fig. 8B, D).
Previous reports indicated that heparin reduces binding of WR to cells (Chung et al., 1998), although this may be dependent on cell type (Carter et al., 2005). We first examined binding of WR and IHD-J virions that had been incubated with heparin by determining the recovery of infectious virus after adsorption to BS-C-1 cells. Following washing, the cells were frozen and thawed and sonicated to release bound virus. The amounts of infectious virus were then determined by plaque assay. Inhibition of WR and IHD-J binding after heparin treatment was 10 to 20% and 80 to 90%, respectively (data not shown). The results suggested that heparin affects entry of IHD-J by reducing binding to cells. However, there was a possibility that heparin selectively inactivated IHD-J, rather than prevented binding to cells.
To eliminate the above possibility, we designed a more direct analysis of binding. Purified WR and IHD-J MVs that contain green fluorescent protein (GFP) fused to the A4 core protein were incubated with or without heparin and then allowed to adsorb to BS-C-1 and sog9 cells. The amount of bound virus was measured by flow cytometry. Heparin reduced WR binding to BS-C-1 cells by about 25%, whereas binding of IHD-J was at the background level (Fig. 9A). WR also bound well to sog9 cells in the presence or absence of heparin (Fig. 9B). Interestingly, IHD-J bound very poorly to sog9 cells (Fig. 9B) providing an explanation for their low infectivity in these cells. We also determined the effect of laminin on binding of WR and IHD-J MVs to BS-C-1 cells. In each case, binding was reduced about 25% (Fig. 9C).
Enveloped viruses generally enter cells through the plasma membrane at neutral pH or through endosomal vesicles at low pH (Earp et al., 2005). VACV is unusual in that it can enter cells by either route, which may contribute to its wide host range (Townsley et al., 2006). Here we provide evidence that the relative utilization of these pathways varies with the VACV strain and to some extent with the host cell. Attachment and entry steps were experimentally separated by allowing adsorption to occur at 4°C for 1 h at neutral pH and then raising the temperature to 37°C for entry. Attachment was measured either by recovery of infectious virus after washing and lysing the cells or by flow cytometry of recombinant VACV with GFP-fused to a core protein. Because the early transcription system is packaged in virus particles and is activated when the core enters the cytoplasm, detection of luc after only 1 h was used as a measure of entry. Several parameters were measured: pH 5 activation before or after virus adsorption at neutral pH, sensitivity to bafilomycin A1, which prevents endosomal acidification, and inhibition by soluble GAGs and laminin. Analysis of the data revealed patterns in which WR and IHD-J were distinctively different (Table 1). Copenhagen and Elstree were similar to IHD-J, whereas Wyeth was intermediate. Thus, WR was the only strain that exhibited enhanced entry by pH 5 treatment of virions before adsorption or pH 5 treatment of cell-bound virions after adsorption, which correlated with sensitivity to bafilomycin A1. Entry of IHD-J, Copenhagen and Elstree were neither enhanced by pH 5 treatment nor inhibited by bafilomycin A1. These results suggested that low pH endosomal entry was more important for WR than for IHD-J, Copenhagen and Elstree strains and that the latter strains rely more on direct entry through the plasma membrane or an endosomal pathway that does not require low pH. Wyeth, although not enhanced by pH 5 treatment, was inhibited by bafilomycin A1. This apparent discrepancy, however, could be explained by previous evidence of two low pH steps in VACV entry(Townsley and Moss, 2007). Thus, if WR is activated with pH 5 buffer and then neutralized prior to adsorption, there is no further low pH enhancement after adsorption but the virus is still bafilomycin A1 sensitive (Townsley and Moss, 2007). Therefore, in this respect, Wyeth resembles activated WR.
Heparin was previously shown to inhibit the binding of VACV strain WR to cells, although the extent of this inhibition seemed to vary in different reports (Carter et al., 2005; Chung et al., 1998; Whitbeck et al., 2009). We found that entry of WR and Wyeth were both relatively resistant to heparin, whereas the other strains were more sensitive. Thus, 1 μg/ml of heparin was sufficient to inhibit IHD-J entry by 90%, whereas inhibition of WR was minimal even at 50 μg/ml. IHD-J was less sensitive to heparan sulfate than to heparin and both WR and IHD-J were insensitive to chondroitin sulfate. Flow cytometry demonstrated that heparin inhibited IHD-J entry at the step of binding to cells.
The greater sensitivity of IHD-J to heparin compared to WR was consistent with the differences in their abilities to infect sog9 cells, which are GAG-deficient (Banfield et al., 1995). Thus, sog9 cells were much more restrictive to IHD-J than to WR and this was shown to be due to lower virus binding. Taken together, these results suggest that binding to GAGs is more critical for viruses that predominantly enter by direct fusion with the plasma membrane rather than through endocytosis. The enhanced sensitivity of WR to heparin when the adsorbed virus was briefly treated with pH 5 buffer to accelerate entry through the plasma membrane was consistent with this idea.
In addition to effects of VACV strain on entry, there were also effects of cell type. We compared several cell lines (BS-C-1, RK-13, HEKn, Vero, pTK2 and Sog9) with regard to low pH enhanced entry. With WR, each of the cell lines except HEKn showed between 2.6- to 4.6-fold enhancement. The sensitivity of WR to bafilomycin A1 could be demonstrated in BS-C-1, RK-13 and HuTK- cells (Townsley et al., 2006). Recently, Whitbeck and coworkers (Whitbeck et al., 2009)reported that bafilomycin A1 inhibition was stronger in BS-C-1 and B78H1 cells than Vero and HeLa cells, though whether this was due to differences in the ability of the drug to lower the pH or to differences in entry were not determined. In addition, they found that WR entry into HeLa, B78H1 and L cells was more strongly inhibited by heparin than entry into Vero and BS-C-1 cells (Whitbeck et al., 2009).
The differences in entry of WR and other VACV strains is remarkable and could be due to their continuous propagation on diverse cell types as well as other factors. It is known that WR and IHD-J differ in the release of extracellular virions due to a point mutation in the A34R ORF (Blasco et al., 1993; Payne, 1979), but this is unrelated to the difference in entry of MVs (A.C.T., unpublished). We are currently in the process of swapping genes between WR and IHD-J to determine the molecular basis for the differences in entry. Once this is accomplished, we may be able to evaluate the effect of mode of entry on virus distribution and pathogenicity in animal models. It will also be interesting to determine the entry pathways of isolates of other orthopoxviruses that have not been passaged extensively in cell culture.
African green monkey kidney BS-C-1, rabbit kidney epithelial RK-13, and potoroo kidney (PtK2) cells were maintained in minimum essential medium with Earle’s salts (EMEM, Quality Biological, Gaithersburg, MD). Mouse sog9 cells were maintained in Dulbecco’s modified Eagle’s medium. Media were supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin. Primary human epidermal keratinocytes (HEKn) cells (Cascade Biologics, Portland, OR) were maintained in the recommended medium. The following strains of VACV were used: WR (ATCC VR-1354; GenBank accession number NC_006998), IHD-J from S. Dales, Copenhagen from E. Paoletti, Elstree (Lister; ATCC VR-1549) and Wyeth New York City Board of Health from a Wyeth Laboratory seed stock. The recombinant VACV WR expressing firefly luc via a synthetic early-late promoter (WRvFire) was previously described (Townsley et al., 2006). Similar vFire recombinants were made using the IHD-J, Copenhagen, Lister and Wyeth strains. WR-A4GFP and IHD-J-A4GFP core-fusion recombinant viruses were generated by the amplification of GFP (accession number AAG27429) and overlap extension PCR for fusion of the GFP open reading frame to the N-terminal codon of A4L with WR and IHD-J VACV genomic DNA as the template. To ensure efficient homologous recombination, flanking sequences of A4L of at least 500 bp were appended to the termini of the PCR product. Cells were infected with 0.1 plaque forming unit of WR or IHD-J per cell and at 1 h post infection transfected with 0.3 μg of gel-purified PCR product. The infected and transfected cells were lysed during three freeze-thaw cycles in a dry-ice bath and clonally purified 5 times by picking GFP positive plaques on BS-C-1 cells. The inserted DNA of the purified recombinant viruses was verified by sequencing.
HeLa cells were infected with VACV WRvFire, IHD-JvFire, CopvFire, ElstreevFire, WyethvFire, WR-A4GFP and IHD-J-A4GFP. MVs were isolated at 48 h after infection as previously described (Earl et al., 2001). Briefly, infected cells were lysed by Dounce homogenization and MVs were purified by sedimentation through two 36% (w/v) sucrose cushions and banding once on a 25 to 40% (w/v) sucrose gradient. Purified stocks were stored at −80°C and sonicated on ice for 1 min prior to infection.
Cells were seeded in 12-well plates and incubated at 37°C overnight. The cells were chilled for 10 min at 4°C and virus allowed to adsorb for 1 h at 4°C at neutral pH. Unattached virus was removed by washing. For pH activation, the cells were then incubated for 3 min at 37°C with Dulbecco’s phosphate buffered saline with Ca2+ and Mg2+ at pH 7.4 or adjusted with HCl and 1 mM 2-morpholinoethane-sulfonice acid to pH 5. The pH was then neutralized and washed with EMEM containing 2.5% FBS, 2 mM L-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin (EMEM-2.5) and incubated with 1 ml EMEM-2.5 at 37°C for 1 h. Cells were then harvested by washing with phosphate buffered saline (pH 7.4) and incubation with 300 μl of Cell Culture Lysis Reagent (Promega, Madison, WI) for 30 min at room temperature on an orbital shaker. Luc assay was performed by adding 20 μl of cell lysate to 100 μl of luc activity assay substrate (Promega), mixed, and chemiluminescence was measured using a luminometer (Berthold Sirius, Bad Wilbad, Germany).
Cells were seeded in 12-well plates and treated with bafilomycin A1 (Sigma, St. Louis, MO) at the indicated concentrations for 1 h at 37°C. Cells were then cooled and infected as above with the exception that bafilomycin A1 was present throughout the adsorption and subsequent incubations.
Purified MVs were incubated in phosphate buffered saline at pH 7.4 or 5 adjusted as described above in a final volume of 50 μl for 3 min at 37°C. The pH was neutralized with excess EMEM-2.5 and added to cells for adsorption at neutral pH as described above.
Virus was treated with the indicated concentrations of heparin, chondroitin sulfate, heparan sulfate, and laminin (Sigma) in EMEM-2.5 for 30 min on ice. Without removing the GAGs, virus was then added to pre-chilled cells for adsorption at neutral pH as described above.
Treated virus was added to pre-chilled cells at neutral pH and incubated at 4°C for 1 h. Cells were then washed immediately and harvested by scraping. Cells were lysed by rapid freeze-thaw three times and the lysates were diluted for plaque assay on BS-C-1 cells. The number of plaques on untreated virus was used as control.
Recombinant strains of VACV with GFP fused to the A4 core protein were incubated with cells at neutral pH for 1 h at 4°C at a multiplicity of infection of 5. Unbound virus was washed with cold medium and cells were harvested and fixed in 2% paraformaldehyde and analyzed with a FACSCalibur flow cytometer using CellQuest (BD Biosciences) and FlowJo Software (Tree Star, Inc, Ashland, OR).
We thank Norman Cooper and Catherine Cotter for providing cells and virus stocks, and Timothy R. Wagenaar, Subbian Sathesh Kumar Panayampalli, Amanda Howard, and Jason Laliberte for helpful discussions. This work was supported by the Division of Intramural Research, National Institute of Allergy and Infectious Diseases, National Institutes of Health.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.