The Poliovirus Empty Capsid Specifically Recognizes the Poliovirus Receptor and Undergoes Some, but Not All, of the Transitions Associated with Cell Entry

Abstract

Experimental results presented here demonstrate that the poliovirus empty capsid binds with saturable character to poliovirus-susceptible cells, binds preferentially to susceptible cells, and competes with mature virus for binding sites on cells. Hence, induced changes in the structure and/or stability of the particle by RNA encapsidation and virus maturation are not necessary for recognition by receptor. In mature virus, heat-induced rearrangements mimic those induced by receptor at physiological temperatures in several important respects, namely, expulsion of VP4 and externalization of the VP1 N-terminal arm. It is shown here that in the empty capsid the VP1 N-terminal arm is externalized but the VP4 portion of VP0 is not. Thus, these two hallmark rearrangements associated with cell entry can be uncoupled.

That the first step in cell entry by poliovirus is attachment to specific receptors on the cell surface has been well established (16, 17, 31). However, the structural features in the poliovirus particle necessary for receptor recognition and the ensuing structural changes that allow the viral RNA to enter the host cell cytoplasm are not well understood. According to a widely accepted model, binding to receptor at physiological temperatures induces a specific set of structural changes that give rise to the cell entry intermediate termed the 135S particle. The changes that characterize the conversion to the 135S particle, in addition to the shift in sedimentation coefficient from 160S to 135S, include a transition from the native (N) antigenic state to the heated (H) antigenic state, an increase in hydrophobicity, and an enhancement in protease sensitivity (6, 8, 11, 14, 25, 26, 28). Two dramatic specific changes are also known to occur, namely, externalization of the N-terminal arm of the VP1 polypeptide and expulsion of the entire VP4 polypeptide (14, 36), both of which are internal in the native poliovirus (15). Later in infection the second type of particle, the 80S particle, accumulates, with a coordinate loss of internalized 135S particles (14). The 80S particle does not contain genomic RNA. The 80S particle may be the empty protein shell which remains after the RNA is released from the 135S particle.

The 135S particle has been considered a required cell entry intermediate since it is the major type of internalized virus-derived particle observed soon after the start of infection (10, 24). However, cold-adapted mutants of poliovirus that do not accumulate 135S particles have been reported recently (9). It is not clear whether these mutants bypass the 135S stage entirely or whether the kinetics of cell entry have changed such that the 135S stage is no longer rate limiting. In the former case, the genuine cell entry intermediate may arise from subtle transitions akin to the breathing motions previously described (22), and the 135S particle may represent an exaggerated form of these changes (9). In either case, understanding the mechanism that leads to the 135S particle should provide clues about the transitions necessary for cell entry.

Examination of the crystal structure of the poliovirus native empty capsid (3) indicates that the native empty capsid can be used as a unique probe in investigating receptor recognition as well as the mechanism of conversion to the 135S particle. The native empty capsid is a putative assembly intermediate that contains the full complement of capsid proteins but not the genomic RNA (20, 29). The native empty capsid is in an immature form, in that the polypeptide VP0 has not been cleaved to form the VP4 and VP2 polypeptides present in the mature virus (18, 19). This particle is considered to be in the native state because it has the same N antigenic surface as mature virus.

Comparison of the crystal structures of the native empty capsid (henceforth referred to simply as the empty capsid) and the mature virus reveals that their outer surfaces and the bulk of their shells are very similar (3). The primary difference is the presence of three amino acid residues at the C terminus of VP3 in the empty capsid (3). These are not observed in the mature virus structure and may be cleaved off during virus maturation. In contrast, the inner surfaces of the protein shells are radically different. One set of major differences in the inner surface is due to the very different disposition of the 10 residues on either side of the VP0 scissile bond in the empty capsid and the corresponding residues in the mature virus. These segments must undergo large-scale rearrangements in the transition from empty capsid to mature virus. The other set of major differences arises from the disorder in the N-terminal arm of VP1 in the empty capsid. This arm is ordered in the mature virus and makes numerous intraprotomer, intrapentamer, and interpentamer contacts. The absence of these contacts in the empty capsid may explain, at least partially, the empty capsid’s greater lability. Here, we exploited the similarities and differences between the empty capsid and the mature virus to arrive at a better understanding of the structural requirements for receptor recognition and the subsequent conformational changes.

Preparation of virions and empty capsids.

HeLa cells in suspension culture were maintained in Joklik’s modified minimal essential medium supplemented with 10% bovine calf serum, 0.3 g of glutamine per liter, 0.5 g of pluronic acid per liter, 3.5 g of glucose per liter, and nonessential amino acids (Gibco). L cells in monolayers were maintained in Eagle’s minimal essential medium containing Earle’s salt solutions (BME) supplemented with 10% fetal bovine serum. Both mature virus and empty capsid particles were prepared from HeLa cells infected with the Mahoney strain of type 1 poliovirus.

Unlabeled mature virus particles were propagated in suspension culture as described elsewhere (22) except that after attachment the cells were resuspended in the supplemented medium indicated above. Labeled empty capsids were grown by the infection procedure described previously (3). Labeled mature virus particles were grown by a procedure similar to that for the empty capsids except that guanidine-HCl was not added. Labeled mature virus and labeled empty capsids for the cell binding studies were prepared from a single radiolabeled infected culture which was divided into equal volumes 3.5 h after the start of infection. Guanidine-HCl was added to a concentration of 0.20 mM to the volume to be used for empty capsid preparation. Guanidine-HCl inhibits poliovirus RNA polymerase and thus favors accumulation of empty capsids. Mature virus particles were purified by isopycnic centrifugation in a CsCl gradient, as described previously (1). Empty capsid particles were purified by isopycnic centrifugation in a Nycodenz gradient followed by rate zonal centrifugation in a 15 to 30% sucrose gradient (3). Particle protein concentration was quantitated by bicinchoninic acid protein assay (Pierce) and Bradford protein assay (Bio-Rad). Mature virus concentration was also determined by absorbance at 260 nm and an extinction coefficient of 7.7 ml · mg⁻¹ · cm⁻¹.

Empty capsid attachment to cells in a receptor-dependent fashion.

The similar exterior surfaces and dissimilar interior surfaces of the empty capsids and the mature virus prompted us to test whether the empty capsid possesses the structural features necessary for specific attachment to cells. These tests consisted of (i) saturation in binding to cells, (ii) differential binding to cells with and without receptor, and (iii) competition with mature poliovirus for receptor binding sites. In all of these experiments, the behavior of the empty capsid was compared to that of mature virus.

Cell attachment was assayed by adding varying concentrations of radiolabeled particles (empty capsid or virion) to 1 ml of washed cells at a density of 1.3 × 10⁷ cells in phosphate-buffered saline (PBS) and incubating for 30 min at room temperature with constant gentle mixing. After incubation, the cells were pelleted by centrifugation at 16,000 × g for 5 min in a microcentrifuge. The supernatant and a 200-μl wash of the cells with PBS were combined and counted for radiolabel. The washed cells were resuspended in 600 μl of PBS and transferred to scintillation vials, along with 600-μl washes of the tubes, and counted for radiolabel.

The binding curves from these experiments (Fig. 1A) indicate that the empty capsid is able to bind to HeLa cells and does so in a manner very similar to that of the mature virus. In both cases, at low concentrations of input particles the binding curve possesses a saturable component. However, at higher concentrations of input particles, the binding does not plateau. The presence of only a partial saturability in the binding of poliovirus to susceptible cells has been reported previously with poliovirus (23). The lower-affinity mode may represent nonspecific, adventitious binding or weak binding to an as yet unidentified class of cell surface molecules. The level of binding observed with the empty capsid sample cannot be due to contamination of the sample with mature poliovirus. The purification protocol used for empty capsid preparation yields samples that, when analyzed by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis, have undetectable levels of VP2 and therefore negligible amounts of mature virus.

FIG. 1.

Binding of empty capsids and mature virions to HeLa cells. Various amounts of labeled particles (empty capsids or mature virus) were added to a constant number of HeLa cells in suspension. After incubation, the cells were pelleted and the radioactivity copelleting with the cells was measured. (A) Saturation binding curve. (B) Scatchard plot of the data in panel A. As a rough approximation, for each of the binding curves two lines have been fitted by the linear least-squares method. The solid and broken lines are the best-fit lines for the empty capsid and mature virus, respectively. The heavy lines represent a high-affinity mode of binding, whereas the light lines represent a low-affinity mode of binding.

The two modes of binding are even more evident when the data are replotted in the Scatchard format (Fig. 1B). At lower concentrations of particles, empty capsids and mature virions bind cells with similar high affinities. At higher particle concentrations, a lower-affinity mode of binding predominates. It is difficult to determine the number of high-affinity binding sites based on conventional Scatchard plot analysis (21). A rough estimate is 5,000 to 6,000 binding sites per cell. Since the cross-sectional area of 6,000 capsids is very much less than the surface area of a HeLa cell, the saturability of this binding is not due simply to blanketing of the cell surface with the particles. This is in approximate agreement with the estimated value of 3,000 binding sites per HeLa cells (27).

Preferential binding to poliovirus-susceptible cells.

Whether the empty capsid binds preferentially to cells containing poliovirus receptor was determined by performing attachment assays with HeLa cells (which possess receptor) and L cells (which do not possess receptor [30]). In the attachment assay, the number of particles added was approximately that needed to saturate the high-affinity HeLa cell binding sites. This comparison reveals that the empty capsid displays the same discrimination as mature virus in binding to cells with or without poliovirus receptor (Table 1).

TABLE 1.

Discrimination in binding of empty capsid and poliovirus to cells with poliovirus receptor (HeLa cells) and cells without poliovirus receptor (L cells)

Type of particle	109 particles bound		Relative binding discriminationa
	HeLa cells	L cells
Empty capsid	42.1	2.6	16.2
Poliovirus	40.9	2.9	14.1

^aRelative binding discrimination is the ratio of the number of particles binding to HeLa cells to the number of particles binding to L cells. 

Competition with mature virus for binding sites.

In experiments to determine if empty capsids and mature virions compete for the same cell binding sites, a constant amount of radiolabeled empty capsids and varying amounts of unlabeled mature virions were added in a cell binding assay at a concentration of radiolabeled particles that would result in the saturation of high-affinity binding sites. Attachment was assayed as in the binding curve experiments. As a control, unlabeled mature virus was allowed to compete with labeled mature virus. The results of these experiments show that unlabeled mature virus and empty capsid compete for the same binding sites on the cell surface (Fig. 2). Moreover, the empty capsid competes as efficiently with mature virus as mature virus competes with itself. This provides compelling evidence that the empty capsid is able to bind to the poliovirus receptor and that it does so with an affinity similar to that of mature poliovirus.

FIG. 2.

Competition for cell surface virus receptors by radioactively labeled particles (empty capsids or mature virions) and unlabeled mature virions. Competition for cell-surface poliovirus receptor was assayed by determining the amount of labeled particle attaching to HeLa cells as the concentration of competing unlabeled virus was increased. The effect of increasing amounts of unlabeled virus is presented by plotting the amount of label copelleting with the cells (normalized to the value obtained with no unlabeled virus added) versus the amount of competing unlabeled virus added.

The above set of experiments demonstrates the ability of empty capsids to attach to cells in a receptor-specific manner. These results indicate that the presence of the three additional amino acid residues at the C terminus of VP3 (which occur in or near the proposed receptor binding site [32, 35]) does not interfere with binding. Moreover, these results demonstrate that the encapsidation of RNA, the cleavage of VP0, and the consequent reorganization of the inner network and increase in stability of the mature particle are not required either to potentiate formation of the virus-receptor complex or to stabilize the complex once formed. Whether or not the empty capsid would be internalized by the cell is an intriguing question. However, the appropriate experiment would be quite difficult to perform since the empty capsid is very heat labile and would undergo conversion to a nonnative form under the conditions necessary for poliovirus internalization.

Structural changes in the empty capsid that mimic those in the mature virion.

Next, we determined whether the empty capsid is able to undergo the same type of structural rearrangements which transform the mature virus to the 135S or 80S particle. In these transitions, the N-terminal arm of VP1 is externalized and the entire VP4 polypeptide is expelled (14). Given that in the empty capsid the VP1 N-terminal arm is disordered (but on the inside surface of the protein shell) and VP4 is still covalently attached to VP2, we did not know what the fate of these segments would be during the conversion of the empty capsid to an H antigenic form. The accessibility of these two segments in the heated empty capsid particles was assayed by protease sensitivity and immunoprecipitability. The assays regarding the VP1 N terminus are based on previous demonstrations that the native-to-135S particle conversion exposes a V8 protease-sensitive site at or near residue 31 of VP1 and renders the particle immunoprecipitable by antibodies against peptides corresponding to the N-terminal region of VP1 (14). The assay for exposure of the VP4 portion of VP0 is based on demonstrations that the breathing motions of the mature virus at physiological temperatures expose VP4 and allow anti-VP4 antibody to immunoprecipitate the particle (22).

The 135S and 80S particles can be prepared in vitro by heating the particles under the appropriate conditions. The 135S particles were prepared by heating purified mature virus in 20 mM Tris–2 mM CaCl₂ for 3 min at 50°C, conditions that result in the virtually complete conversion of native virion to 135S particle (7, 37). The 80S particles were prepared by heating purified mature virus at 56°C for 10 min. These conditions produce virtually complete conversion of native particle to 80S particle (4, 14). Native empty capsids were converted to an H antigenic form by heating at 40°C for 1 h. Such incubation completely converts the empty capsid to the H antigenic state, as is evidenced by the particle’s loss of alkaline dissociability into pentamers (2, 34).

Sensitivity to V8 protease activity.

Experiments with V8 protease (which cleaves preferentially at Asp and Glu residues) have demonstrated that the VP1 capsid protein is resistant to cleavage when in the mature virus form but is susceptible to cleavage when in the 135S or 80S form (14). The sensitivities to V8 protease of the heated and unheated empty capsid were compared to those of native and heated (80S) virus to determine if a similar transition in the VP1 N-terminal arm occurs in the empty capsid upon heating.

The V8 protease (Boehringer Mannheim) has optimum activity at pH 7.8. Since the native empty capsids are unstable at alkaline pH, a suboptimum pH of 7.5 (in PBS) was used for all digests. The amount of V8 and length of digestion time were adjusted to provide a clear signal in the control digest of 80S particles. Approximately 5 μg of V8 protease was added to each 10 μg of particle protein and incubated for 2 h at room temperature.

Cleavage of VP1 by V8 was assayed by Western blot analysis with antibodies anti-pep1 and anti-pep9, which were raised against synthetic peptides corresponding to residues 24 to 40 and 270 to 287 of VP1, respectively (5). Samples were electrophoresed in an SDS–12.5% polyacrylamide gel with 2% cross-linking, and the polypeptides were transferred to nitrocellulose membranes with the Phast transfer system (Pharmacia). The membranes were blocked by incubation in TBST (10 mM Tris [pH 8.0], 150 mM NaCl, 0.05% Tween 20) and 3% dry milk (Carnation) for 30 min at room temperature. This was followed by incubation with antiserum (1:2,000 dilution in TBST) at room temperature for 30 min. Anti-rabbit immunoglobulin G–alkaline phosphatase conjugate (Vector Laboratories) (1:5,000 dilution in TBST) was the secondary antibody. The incubation was at room temperature for 30 min. The bands were visualized with BCIP (5-bromo-4-chloro-3-indolylphosphate)–nitroblue tetrazolium (Promega) according to the manufacturer’s directions.

The results of the V8 sensitivity experiments demonstrate that upon heating of the empty capsid, the N-terminal arm of VP1 becomes susceptible to cleavage by V8 protease (Fig. 3) and thus must be externalized. Moreover, the extent to which the VP1 N-terminal arm is externalized is the same in both the empty capsid and the mature virus, since in both cases Western blot analysis with anti-pep9 (recognizing residues 270 to 287 of VP1) yields the same cleavage pattern as with anti-pep1 (which recognizes residues 24 to 40 of VP1, a sequence that spans the V8 cleavage site).

FIG. 3.

V8 protease sensitivity of heated and unheated empty capsids, native virus, and 80S particles. Native virus, 80S particles, unheated empty capsids, and heated empty capsids were treated with V8 protease. The capsid proteins were separated by SDS-polyacrylamide gel electrophoresis. Fragments containing the antigenic sequence of interest were visualized specifically by immunoblotting and staining with the procedures outlined in the text. As controls, particles not treated with V8 protease were analyzed in parallel. (A) Western blot analysis with polyclonal antibodies binding the pep1 region of VP1 (residues 24 to 40). (B) Western blot analysis with polyclonal antibodies binding the pep9 region of VP1 (residues 270 to 287). PV, poliovirus; EC, empty capsid.

Immunoprecipitation of unheated and heated empty capsids.

The conformation of the exposed VP1 arm and the fate of the VP4 portion of VP0 in heated empty capsids were assayed by immunoprecipitation with polyclonal anti-VP4 antibody raised against synthetic VP4 peptide (22); polyclonal anti-pep0 and anti-pep1 antibodies raised against synthetic peptides corresponding to residues 7 to 24 and 24 to 40 of VP1, respectively (5); and monoclonal anti-pep1 antibody. The monoclonal anti-pep1 antibody has been shown previously to be more reactive with the 135S particle than with the 80S particle (14). This is due presumably to minor differences in the exposure or conformation of the amino terminus of VP1. Antibody binding was quantitated by incubation at room temperature for 60 min of [³H]leucine-labeled particles (∼10,000 cpm) with serial fivefold dilutions of antisera in PBS–0.1% egg albumin (PBSeA). Then, 50 μl of a 10% solution of protein A-Sepharose (Sigma) in PBSeA and 750 μl of PBSeA–0.05% Nonidet P-40 were added to the samples. The samples then were incubated for 120 min at 4°C with continuous gentle mixing. The immunocomplexes were collected by centrifugation at 16,000 × g for 10 min. The radioactivity remaining in the supernatant was removed and counted for radiolabel. The bound radioactivity was released by boiling the pellets in PBS–2% SDS–2% β-mercaptoethanol and then counted together with a 200-μl wash for radiolabel. Percent precipitation was calculated as the ratio of bound counts per minute to total (bound and unbound) counts per minute. The pentamers used for the positive control in the anti-VP4 immunoprecipitation experiment were generated by treating native empty capsids with high pH. Specifically, an equal volume of 0.1 M Tris was added to empty capsids in 25% sucrose in PBS, and the mixture was incubated on ice for 10 min. The pH was then neutralized by adding 4 volumes of 10× PBS.

The results of the immunoprecipitation experiments show that the polyclonal antibodies raised against peptides corresponding to segments in the N-terminal arm of VP1 bind to the heated empty capsid but not to the native empty capsid (Fig. 4A and B). This corroborates the finding of V8 sensitivity experiments that heat-induced conversion of the particle externalizes this arm. However, the heated empty capsid is not bound by a monoclonal anti-pep1 antibody that does bind the externalized arms in the 135S and 80S particles (Fig. 4C). Thus, the externalized arm in the heated empty capsid evidently adopts a different conformation than that in either the 135S or 80S particle.

FIG. 4.

Immunoprecipitation titrations of heated and unheated empty capsids. Titrations were with the following antibodies (Ab). (A) Polyclonal antibodies binding the pep0 region of VP1 (residues 7 to 24). (B) Polyclonal antibodies binding the pep1 region of VP1 (residues 24 to 40). (C) Monoclonal antibody raised against pep1. (D) Polyclonal antibodies binding VP4. In each panel, the percent immunoprecipitated versus the log of the dilution is plotted. Symbols: , empty capsids; ▵, heated empty capsids; ◊, pentamers; , mature virus; ⧫, 135S; ▴, 80S. The dashed line in panel D represents the background level for pentamer immunoprecipitation determined as described in the text. The reactivity of the undiluted anti-pep1 and anti-pep0 polyclonal antibodies toward unheated empty capsids may be due to instability of the empty capsids in serum concentration. The high background in control experiments using pentamers is due to adventitious binding of pentamers to the protein A-Sepharose beads. Mock experiments in which no anti-VP4 antibody was added resulted in precipitation of the pentamers approximately equal to the background level observed in panel D.

The immunoprecipitation experiment with polyclonal antibodies against VP4 showed no reactivity with heated empty capsid (Fig. 4D). The lack of reactivity with the heated empty capsid is unlikely to be due simply to the inaccessibility of a partially exposed VP4 segment. Previous experiments showed that VP4, when even partially exposed during the normal “breathing motion” of the mature virus, can be bound by the anti-VP4 antibody (22). The lack of reactivity is most easily explained by the VP4 portion of VP0 not being externalized at all, perhaps as a consequence of it still being covalently attached to the VP2 portion of VP0. However, it is also possible that this region is externalized transiently and then later completely internalized again during the transition to the 80S equivalent of the empty capsid. In either case, the ability of the N terminus of VP1 to be externalized without the permanent coexternalization of VP4 is significant since it indicates that these two hallmarks of the transition to the cell entry intermediate to some degree can be uncoupled, in contrast to what was previously believed (14).

The immunoprecipitation results, when interpreted in the context of the empty capsid structure, provide hints regarding the mechanism for conversion of the native mature virus to the cell entry intermediate. First, that the extreme N-terminal region of VP1 is externalized in the heated empty capsid even though it is completely disordered in the native empty capsid suggests that this region is not an essential part of the conversion mechanism. Rather, the extreme N-terminal region of VP1 seems to be a passive component in the process. This idea is reinforced by the observation that this region contains one of the most poorly conserved sequences in picornaviruses (33). Thus, the arm is unlikely to make any highly sequence-specific interactions in the virion as part of the conversion mechanism. A passive role in externalization also makes mechanistic sense if the capsid pores thought to open in the receptor–heat-induced transition are situated, as proposed, near the quasi-threefold axis near the center of each protomer (13). Much of the N-terminal arm in the mature virus is positioned near this region (Fig. 5). The N-terminal arm in the empty capsid must also be somewhere in this region because it is tethered nearby to the beta-barrel core of VP1. Thus, if the pores do open by disruption of the intraprotomer contacts at the quasi-threefold axes, then the N-terminal arms might be externalized by virtue of their proximity to these pores. Second, VP4 may also be externalized primarily due to its location. In the mature virus, this polypeptide runs underneath the postulated pores. Third, the experimental results presented here suggest that the mechanism for conversion to the cell entry intermediate does not rely on the rearrangements which occur in the inner surface of the capsid during the final stages of mature virus assembly. Instead, the mechanism seems to be contained entirely within the surface features of the capsid and the beta-barrels which comprise the bulk of the shell.

FIG. 5.

View of the poliovirus protomer. The view is from inside the particle looking out. The N-terminal arm of VP1 (residues 6 to 10 and 20 to 69) and all of VP4 are represented as blue and green ribbons, respectively. Residues 1 to 5 and 11 to 19 are disordered and therefore are not shown. The remaining portions of the protomer are represented by the surfaces, with blue, yellow, and red corresponding to VP1, VP2, and VP3, respectively. The sphere indicates the position of the VP0 scissile bond in the empty capsid. The quasi-threefold axis relating VP1, VP2, and VP3 in the protomer is located near the VP0 scissile bond and is roughly perpendicular to the page.