Suppression of Injuries Caused by a Lytic RNA Virus (Mengovirus) and Their Uncoupling from Viral Reproduction by Mutual Cell/Virus Disarmament

Abstract

Viruses often elicit cell injury (cytopathic effect [CPE]), a major cause of viral diseases. CPE is usually considered to be a prerequisite for and/or consequence of efficient viral growth. Recently, we proposed that viral CPE may largely be due to host defensive and viral antidefensive activities. This study aimed to check the validity of this proposal by using as a model HeLa cells infected with mengovirus (MV). As we showed previously, infection of these cells with wild-type MV resulted in necrosis, whereas a mutant with incapacitated antidefensive (“security”) viral leader (L) protein induced apoptosis. Here, we showed that several major morphological and biochemical signs of CPE (e.g., alterations in cellular and nuclear shape, plasma membrane, cytoskeleton, chromatin, and metabolic activity) in cells infected with L⁻ mutants in the presence of an apoptosis inhibitor were strongly suppressed or delayed for long after completion of viral reproduction. These facts demonstrate that the efficient reproduction of a lytic virus may not directly require development of at least some pathological alterations normally accompanying infection. They also imply that L protein is involved in the control of many apparently unrelated functions. The results also suggest that the virus-activated program with competing necrotic and apoptotic branches is host encoded, with the choice between apoptosis and necrosis depending on a variety of intrinsic and extrinsic conditions. Implementation of this defensive suicidal program could be uncoupled from the viral reproduction. The possibility of such uncoupling has significant implications for the pathogenesis and treatment of viral diseases.

INTRODUCTION

Viral reproduction is very often accompanied with host cell damage ranging from minor functional alterations to the complete demise. This damage is a major cause of viral diseases. It is a general belief that the impairment of cellular functions is due to the virus/host competition for substrates, energy sources, and cellular infrastructure. In other words, cell pathology is usually considered to be either a prerequisite for or consequence of efficient replication of viral genomes and production of viral particles (37).

A common outcome of viral infection is death of the host cell. Two types of such death, necrosis and apoptosis, can be distinguished on the basis of morphological and biochemical criteria. Typical gross features of necrotic death are rounding up, permeabilization of plasma membrane, and complete disorganization of the cytoplasmic and nuclear infrastructure. On the other hand, the surface of apoptotic cells exhibits characteristic “blebbing” not accompanied with the plasma membrane permeabilization; chromatin is strongly condensed, the chromosomal DNA undergoes fragmentation to oligonucleosomes, and the cells are disrupted into so-called apoptotic bodies. While apoptotic cytopathic effect (CPE) results from the implementation of a cellular genetically encoded program activated by the infecting agent (8), the virus-triggered necrosis is considered to be a passive and uncontrolled cellular destruction. These two types of death have different biological consequences: in contrast to apoptotic cells, necrotic cells usually trigger inflammatory reactions.

Recently, we proposed that the most severe virus-related cell damage may come from the host defensive antiviral measures as well as from the viral antidefensive activities (4). Such reasoning predicts that suppression of these activities (mutual host/virus disarmament) may result in a marked amelioration of the pathogenic properties of a virus, not necessarily accompanied with a significant impairment of its reproduction. The present study aimed to check the validity of this proposal.

As a model, mengovirus (MV), a strain of encephalomyocarditis virus (EMCV; Cardiovirus genus, Picornaviridae family) was chosen. The virus contains a 7.8-kb single-strand RNA genome of positive polarity encoding the viral polyprotein, which is eventually processed into a dozen mature proteins (Fig. 1A) (1, 47), including two so-called security proteins, the leader (L) and 2A, specifically dedicated to antidefensive functions (4).

Fig 1.

Effects of MV infection on the morphology and plasma membrane permeability of HeLa cells. (A) Schematic representation of the genomes of wt MV and its mutants. The substitutions in the Zn finger motif of the Zf-mut are shown. The deleted amino acids in the ΔL(p7) mutant are highlighted in black. The position of the newly discovered protein 2B* encoded in an alternative reading frame (43) is not shown. (B) Permeabilization of the plasma membrane of infected cells as assayed by PI staining. (C) Scanning electron microscopy of virus-infected and mock-infected cells. Bar, 10 μm. (D) Phase contrast images and PI-stained cells. The same fields are shown for the identical conditions in the two rows. (E) Hoechst 33342-stained nuclei. The time of infection is indicated.

Infection with MV (or other EMCV strains) kills HeLa cells. Their death is preceded by a variety of pathological changes, to which the security proteins of this virus make significant contributions. Thus, the 2A protein is implicated in the shutoff of the cap-dependent translation of host mRNA (29, 56), likely in part through its association with ribosomes (30, 44), possibly due to its RNA-binding activity (27). This protein also appears to be involved in inhibition of host mRNA transcription (5). Likewise, the L protein was reported to inhibit host translation (65) and transcription of certain genes (31) and impairs the nucleocytoplasmic traffic (41, 48) through binding Ran-GTPase (48) and modulation of phosphorylation of nucleoporins (9, 49, 50).

The results obtained in this study have validated the hypothesis according to which the major cause of virus-induced cell pathology may be the struggle between host defensive and viral antidefensive mechanisms. The results also are compatible with a model according to which viral infection may activate a cell-encoded defensive death program, which includes two competing branches, necrotic and apoptotic. The choice between these subroutines depends on the environmental conditions and the properties of hosts and viruses, in particular, on activities of the viral antidefensive security proteins.

MATERIALS AND METHODS

Cells and viruses.

HeLa-B cells (60) and human rhabdomyosarcoma RD cells were grown in Dulbecco's modified Eagle medium (DMEM) with 10% fetal bovine serum (HyClone) at 37°C in 5% CO₂. Wild-type MV was derived from the plasmid pM16.1 (21). The MV mutants encoding L protein with substitutions Cys19→Ala and Cys22→Ala destroying the Zn finger motif (hereinafter Zf-mut) and with Val216→Leu and Thr219→Ser substitutions in the RNA-dependent RNA polymerase 3D^pol (Pol-mut) were described previously (20, 31). The L deletion mutant ΔL(p7) corresponded to the vM16.1Δ L(12-52) mutant (i.e., mutant having a deletion of amino acid residues 12 through 52 of the L protein [65]), which underwent 7 passages in BHK-21 cells in this laboratory.

Infection, single-cycle growth experiments, and plaque titration.

These procedures were carried out essentially as described previously (52). The multiplicity of infection (MOI) in all experiments was ≈40 PFU/cell, except for the time-lapse video microscopy, where it was ≈200 PFU/cell. Plaque assay was performed in RD cells, whereas all other experiments were done with HeLa-B cells. The broad-spectrum caspase inhibitor Q-VD-OPh (QVD; MP Biomedicals), when present, was used at a final concentration of 20 μM.

Tubulin and actin staining.

Virus-infected and mock-infected cells were grown on coverslips at 37°C in 5% CO₂ for the time intervals indicated on the figures and fixed with 3.7% paraformaldehyde in phosphate-buffered saline (PBS) for 20 min at 37°C. The fixed cells were washed thrice with PBS and were incubated in 0.2% Triton X-100 for 5 min. The cells were again washed thrice with PBS and incubated in 3% fat-free milk in PBS for 1 h and then in a solution of anti-α-tubulin antibodies (DM1A; CEDARLANE) diluted 1:100 in 3% fat-free milk for 1 h at 37°C. After being washed thrice with PBS, the cells were incubated with a mixture of 1:100-diluted anti-mouse fluorescein isothiocyanate (FITC)-conjugated IgG antibodies (Invitrogen) and 1:150 diluted tetramethyl rhodamine isocyanate (TRITC)-conjugated phalloidin (Sigma) in 3% fat-free milk in PBS for 1 h at 37°C. After being washed, the cells were incubated with 5 μM Hoechst 33342 at room temperature for 20 min. The coverslips were placed onto a drop of 90% glycerol in 10 mM Tris, pH 9.0.

TUNEL assay.

Virus-infected and mock-infected cells were grown on coverslips at 37°C in 5% CO₂ for the time intervals indicated on Fig. 4, stained with Hoechst 33342, and fixed at room temperature with Safe Fix (Curtin Matheson Scientific) for 20 min. The cells were treated consecutively with 96% and 70% ethanol and stored at −20°C. The assay was performed by using the apoptosis detection DeadEnd Colorimetric terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) system kit (Promega) according to the manufacturer's instructions.

Fig 4.

TUNEL assay of HeLa cells infected with wt and mutant MV. Top row, Hoechst 33342 staining; bottom row, TUNEL staining. The same fields are shown for identical conditions in both rows.

Microscopy.

Microscopy was performed with a Nikon Eclipse Ti-U microscope equipped with the following filters: UV-1A (for Hoechst 33342), B-2A (for FITC), and G-2A (for TRITC and propidium iodide [PI]). Photography was made by using a digital Infinity 3 camera. When appropriate, the contrast and/or brightness of the images were adjusted by using Adobe Photoshop options. In such cases, identical adjustments were made in all images of a given panel.

Scanning electron microscopy.

Cells were fixed with 2.5% glutaraldehyde in Hanks buffer without Ca²⁺ and Mg²⁺ containing 5 mM EDTA and 10 mM HEPES, pH 7.3. Postfixation was done with 1% OsO₄ in 0.1 M Na cacodylate containing 0.1 M sucrose, pH 7.3. Cells were dehydrated in the acetone series, critical point dried with liquid CO₂ as a transitional fluid in a Balzers apparatus, sputter coated with gold-palladium, and observed at 20 kV with a Camscan S-2 scanning electron microscope.

Time-lapse video microscopy.

Cells grown in Delta T dishes (Bioptechs) until 70 to 80% confluence were infected with an appropriate virus in the presence or absence of QVD. After a 30-min adsorption period, the cells were washed with DMEM and further incubated in DMEM overlaid with mineral oil light white (MPI) at 37°C by using the Bioptechs Delta T heating system mounted on a Carl Zeiss Axiovert 200 M microscope equipped with a Plan-Neofluar 40× numerical-aperture (NA) 0.7 LD phase-contrast objective and a Hamamatsu Photonics ORCAII-ERG2 camera. In the case of cells infected with Zf-mut in the presence of QVD, the snapshots were made with 5-min intervals for 20 h postinfection (hpi).

Permeabilization assay.

Cells grown in the plastic dishes with glass-bottom (CELLwiev; Greiner Bio-One) were infected and incubated at 37°C in 5% CO₂. At appropriate time intervals (around 7 and 17 hpi), a mixture of Hoechst 33342 and propidium iodide (MP Biomedicals) (to a final concentrations of 5 μM and 10 μg/ml, respectively) was added to the medium, and the cells were photographed after 5 to 10 min of incubation. The numbers of nuclei stained with each dye were counted, and the proportion of the PI-stained nuclei relative to the Hoechst-stained nuclei was calculated.

Viability assay.

The CellTiter-Blue cell viability assay (Promega) based on the ability of metabolically active cells to catalyze NADH-dependent reduction of resazurin into a highly fluorescent resorufin was used. Cells grown in 12-well plastic dishes to ≈80% confluence were infected with the relevant virus and after being washed with DMEM were incubated in 0.5 ml DMEM at 37°C in 5% CO₂. The incubation medium in appropriate samples contained QVD. At 16 hpi, 180 μl of alamarBlue (to a final concentration of 230 μM) was added to each well. The control well received SDS to a final concentration of 0.13%. Incubation was continued for 2 h. Fluorescence at 600 nm was measured at 20-min intervals with Power Wave XS2 (BioTek).

Protein synthesis.

Cells were grown in 35-mm plastic dishes and infected as described above. At the time intervals indicated on Fig. 3, the growth medium was replaced with the Eagle's medium containing a hydrolysate of ¹⁴C-labeled protein (final concentration, 20 μCi/ml). After a 30-min incubation, the medium was discarded and 0.25 ml of 2× sample buffer (60 mM Tris-HCl [pH 6.8], 2% SDS, 10% glycerol, 5% β-mercaptoethanol, 0.01% bromophenol blue) was added. The lysates were equalized with respect to the actin content by Western blotting and were subjected to electrophoresis in 12% polyacrylamide gels. The gels were fixed with a 10% acetic acid in 20% ethanol. The gels underwent 3 successive 1-h-long soakings in dimethyl sulfoxide (DMSO) followed by soaking overnight in 20% 2,5-diphenyloxazole (PPO) in DMSO, washed with water for 3 h, dried at 70°C, and exposed to films. ¹⁴C-labeled denatured proteins run in a parallel gel were subjected to Western blotting by using the above-mentioned antiactin antibodies from Sigma.

Fig 3.

Reproduction of wt and mutant MV and effects on metabolism of infected HeLa cells. (A and B) Single-cycle reproduction. ○, ●, wt MV; □, ■, Zf-mut; △, ▲, Pol-mut; ♢, ♦, ΔL(p7); open symbols, without QVD; full symbols, with QVD. Standard errors of means are presented. (C) Viability assay. The metabolic activity of cells was measured after 16 hpi. ×, SDS-treated cells; ▽, mock-infected cells; other symbols are as described in panels A and B. The steeper the initial drop of the optical density, the higher the metabolic activity. (D and E) Translational shutoff in the cells infected in the absence (D) and presence (E) of QVD. Electrophoretic patterns of ¹⁴C-labeled proteins of infected and mock-infected cells. The bottom row shows the actin marker as revealed by Western blotting.

Mitochondria staining.

For visualization of mitochondria, HeLa cells expressing the enhanced yellow fluorescent protein fused with the mitochondria-targeting sequence from subunit VIII of human cytochrome c oxidase (EYFP-mito) (52) were used.

RESULTS

Characterization of CPE induced by wild-type MV in HeLa cells.

To have a background for the evaluation of effects of various modifications of experimental conditions, the features of CPE induced in HeLa cells by the wild-type (wt) MV were characterized in some detail. The infected cells exhibited typical signs of necrotic CPE. The cells rounded up (Fig. 1C), and their plasma membrane underwent alterations as evidenced by the increased permeability to propidium iodide (PI; Fig. 1B and D) and formation of filopodia (Fig. 1C) and large balloon-like blisters (Fig. 1D). As evidenced by time-lapse video microscopy, small blisters started to appear at different loci of the cellular surface at ∼4 to 5 hpi and coalesced relatively rapidly (within 10 to 20 min) into entities comparable in size with the whole cells and eventually were ruptured and collapsed (not shown). The mechanism underlying blistering is unknown but is likely related to osmotic effects caused by the permeabilization of the plasma membrane (see below). Nuclei of the infected cells shrunk, and chromatin was condensed (Fig. 1E). The majority of cells developed these signs by 4 to 6 hpi.

The viral infection was also accompanied by cytoskeletal alterations (Fig. 2). The microtubule network, loosely radiated in a fan-shaped manner from the perinuclear space in mock-infected cells, was rearranged into dense cocoon-like bundles in a significant proportion of the wt MV-infected cells already by 4 hpi (Fig. 2A), resembling alterations observed upon poliovirus infection (57). In some cells, thinner ringlike tubulin structures at the cellular periphery were also present. Later on, the microtubule network was degraded and was no longer visible in a significant proportion of cells. The network of actin bundles was also largely rearranged early in infection into dense peripheral circular structures and later became visible as amorphous masses with numerous filopodia (Fig. 2B).

Fig 2.

Cytoskeletal changes in HeLa cells infected with wt and mutant MV. (A) Tubulin revealed by immunofluorescence; (B) TRITC-phalloidin-stained actin. The same fields are shown for the identical conditions in panels A and B.

Reproduction of wt MV in HeLa cells was rather efficient, with the final yield amounting to several hundred PFU per cell. The development of pathological changes in infected cells coincided in time with viral reproduction, which was essentially completed by 6 to 8 hpi (Fig. 3A and B).

Changes in cellular pathology due to partial disarmament of the virus.

Partial disarmament of MV was achieved by the functional inactivation of one of its security proteins, the L protein. As reported previously (52), infection of HeLa cells with an MV mutant encoding L with the destroyed Zn finger motif (Cys19→A and Cys22→A) elicited apoptosis, implying that the wt L exhibited an antiapoptotic activity in these cells. This type of death was manifested by plasma membrane blebbing (Fig. 1C and D) not accompanied, at least by the end of the viral reproduction cycle, with increased permeability to PI (Fig. 1B and D) and with only very rare blistering. Also, a significant proportion of cells exhibited a stronger chromatin condensation, nuclear fragmentation (Fig. 1E and 4), and positive TUNEL signals (Fig. 4). The microtubule and microfilament networks were transformed into dense amorphous entities, with filopodia seen in many cells (Fig. 2).

Although functional inactivation of L resulted in a severalfold decrease in the final yield of infectious virus (Fig. 3A), the pathological changes in the mutant-infected cells did not develop more slowly, and probably developed even faster, than in the wt-virus-infected cells (see Fig. S1 in the supplemental material).

Amelioration of cell pathology upon mutual virus/cell disarmament.

Partial host disarmament was achieved through preventing the development of the defensive apoptotic response by a pan-caspase inhibitor Q-VD-OPh (QVD) (17). The addition of QVD did not appreciably affect the appearance of uninfected cells (not shown) and the time course of reproduction and final harvest of either wt virus or Zf-mut (Fig. 3A). However, when infection with the Zf-mut (i.e., partially disarmed MV) was carried out in the presence of QVD, development of not only the major apoptotic signs but also of necrotic CPE was either suppressed or markedly delayed. The infected cells continued to resemble the mock-infected cells in their overall appearance, nuclear morphology, absence of plasma membrane permeabilization (Fig. 1), the cytoskeletal changes (Fig. 2), and negative TUNEL signals (Fig. 4) well beyond the time of the generation of the infectious virus (Fig. 3A and B). Time-lapse video microscopy demonstrated that many cells exhibited a close-to-normal appearance as late as 20 hpi, though marked vacuolization could be seen at >8 hpi (that is after completion of the replication round) (see the movie in the supplemental material). Also, the movie in the supplemental material shows an apparent increase in the local motility of the cellular surface at the late times. The metabolic activity of the cells infected with Zf-mut in the presence of QVD, as measured by their capacity to catalyze the NADH-dependent reduction, was only slightly lower than that of the mock-infected cells (Fig. 3C). The addition of QVD to the wt MV-infected cells exerted little, if any, effect on the development of morphological signs of CPE, e.g., alterations of the microtubule network and the appearance of Hoechst-stained nuclei (Fig. 5), and resulted in preservation of only a low level of the metabolic activity (Fig. 3C). These observations indicated that the partial disarmament solely of the host was insufficient for the marked amelioration of the virus-induced pathology.

Fig 5.

Effect of QVD on cytopathic changes induced by infection with wt MV. The pattern of tubulin staining and morphology of nuclei (Hoechst 33342 staining) observed in the absence of QVD (A) did not appreciably differ from those in samples infected in the presence of QVD (B). The same fields for identical conditions are shown in the top and bottom rows.

As noted above, the Zf-mut exhibited a somewhat lower (than the wt) level of reproduction in HeLa cells. Therefore, one might argue that it was just the less efficient viral reproduction rather than the inactivation of specific L functions that largely contributed to the suppression of cellular injury in the presence of QVD. Such an argument seemed to be warranted, because it is known that a decrease in the efficiency of poliovirus reproduction may indeed change the response of HeLa cells from necrotic to apoptotic (60). To evaluate the validity of such reasoning in the case of the present system, we used as a control another EMCV mutant, the Pol-mut, with a growth potential comparable to that of Zf-mut (Fig. 3B) but having mutations in the RNA-dependent RNA polymerase 3D^pol (Val216→Leu and Thr219→Ser) (20) rather than in L. The Pol-mut-infected cells exhibited necrotic CPE similar to that observed upon wt MV infection, and no obvious delay in the time course of the development of morphological signs of this pathology was observed in the presence of QVD (Fig. 6). The metabolic activity of these cells was as markedly (or even stronger) inhibited as in the wt-infected cells under identical conditions (Fig. 3C). Thus, the two remarkable features of the Zf-mut infection, apoptosis in the absence of QVD and significantly prolonged cell survival in the presence of the inhibitor, did not appear to be caused by a less efficient reproduction of the mutant but rather were due to the inactivation of L.

Fig 6.

Morphological changes triggered by infection with the Pol-mut. The same fields are shown for the identical conditions in each row.

Incompleteness of the suppression of cellular pathology.

Although the above-described data unambiguously revealed the possibility of uncoupling of the major pathological signs from the viral growth, most of the cells infected with Zf-mut in the presence of QVD eventually died. Their death was preceded by a variety of metabolic and structural alterations. The Zf-mut induced the shutoff of cap-dependent translation with similar kinetics in the presence and absence of QVD and even somewhat faster than in the wt virus infection (Fig. 3D and E). In this mutant, the L protein was inactivated by destruction of its Zn finger domain. However, deletion of nearly the entire L sequence in ΔL(p7) MV (Fig. 1A), associated with a somewhat greater reproduction deficiency (Fig. 3B), resulted in a similar inhibition of host translation (Fig. 3D), confirming that this inhibition was not appreciably dependent on the presence of L. Of note, QVD prevented ΔL(p7)-induced pathology as efficiently as it did in the case of Zf-mut (Fig. 3C and 7). QVD failed to prevent mitochondria fragmentation (not shown), previously demonstrated to occur upon both wt MV and Zf-mut infections (52).

Fig 7.

Effect of QVD on the morphological changes induced by the ΔL(p7) mutant. The mock-infected control cells from this experiment are shown in Fig. 6. The same fields are shown for the identical conditions in each column.

DISCUSSION

Several aspects of the present study are of general interest. First, the results demonstrated that the production of infectious progeny by a lytic RNA virus may, under certain conditions, take place without typical cellular injuries. The uncoupling of cell damage and viral growth caused by partial disarmament of the host (HeLa cells with chemically inhibited apoptotic pathway) and virus (MV expressing incapacitated L protein) implies that at least some of the major pathological alterations are not the prerequisites for, or direct consequences of, viral reproduction but rather represent the outcome of the interactions between the host innate immunity and the virus self-protection. This notion differs from the generally held paradigm considering viral CPE largely as a consequence of the competition for substrates and energy sources or of rearrangement of the intracellular infrastructure to ensure efficient replication of the viral genome. Such competition and rearrangement may exist as well, but they do not appear to be the sole or even the major factors in the virus-induced injuries, at least in the case of certain virus/cell systems. It can be said that the infected cell is injured not because it has been “robbed” but rather because of fighting with a suspicious agent. Of course, the above-described way of thinking does not negate the fact that the extent of virus-induced pathology may to some extent depend on the efficiency of viral growth. Moreover, it cannot be ruled out that the major cell pathology and viral replication may be coupled more tightly in some virus/cell combinations than in the system studied here.

Of note, additional self-damage can be inflicted by mechanisms of the acquired immunity in the case of virus-organism interactions.

The second important aspect is a striking apparent multifunctionality of the small (67-amino-acid-residue) leader protein of EMCV/MV. In addition to the impressive list of its other activities (see the introduction), L was shown here to be involved in the alterations of plasma membrane, cytoskeleton (microtubules and microfilaments), and cellular and nuclear shapes. Indeed, if implementation of the apoptotic program was suppressed by an appropriate drug, the occurrence of these changes depended on the availability of functional L. In addition, the ability of cardiovirus L protein to suppress formation of stress granules in the infected cells has recently been demonstrated (10). Though molecular mechanisms of all these alterations are yet to be elucidated, they obviously involve very different biochemical reactions.

Alteration of the plasma membrane is a common feature of picornavirus-infected cells. It is evidenced by increased permeability to low-molecular-weight inhibitors (14), changes in intracellular ionic composition (16), and several other deviations (reviewed in reference 15). Plasma membrane blistering, observed also upon infection with another cardiovirus (64), seems to be a related phenomenon likely caused by osmotic disequilibrium. It is not a virus-specific event, since similar balloon-like protrusions could also be seen in uninfected cells treated, for example, with anticellular antibodies in the presence of complement (26) or with tumor necrosis factor (39).

Although it has long been known that picornavirus infection leads to a marked rearrangement of the cytoskeleton (40), the underlying biochemical reactions remain largely enigmatic. It was reported that disruption of the cytoskeletal network may be due to the cleavage of certain cellular proteins by the proteases 3C^pro of poliovirus (36) and foot-and-mouth disease virus (7). Another picornaviral protease, rhinovirus 2A^pro (a security protein having no structural/enzymatic counterpart among cardiovirus proteins), was implicated in the destruction of cytokeratin 8 in the virus-infected HeLa cells (53). However, a coherent picture of the events leading to the cytoskeleton collapse (and consequently to change in the cell shape) in cells undergoing picornavirus-triggered CPE is yet to be drawn. Even less is known about the factors affecting in such cells the shape of the nucleus and infrastructure of its interior, although one may speculate that changes in the intricate structure of nucleoskeleton (19, 54) are somehow involved.

It is a challenge to understand how the cardiovirus L protein is involved in the control of all these diverse functions. It cannot be rigorously ruled out that L independently affects numerous unrelated targets but more likely it controls various downstream functions of a hub element (or a few control elements) of a yet-to-be-discovered signaling pathway(s) affecting the whole cellular metabolism and infrastructure and playing a pivotal role in the fate of virus-infected cells. It is appropriate to note in this regard that at least the majority of known activities of cardiovirus L depend on the integrity of its Zn finger motif (9, 22, 31, 41, 52) and involve phosphorylation regulation (9, 22, 48, 50, 66).

As we argued previously (2, 3), picornavirus infection may activate two competing death programs, caspase-dependent apoptosis and caspase-independent necrosis. We hypothesize now that not only the apoptotic but also the necrotic program is host encoded, representing a distinct component of the innate immunity system. The main argument supporting this hypothesis is provided by the observation that major necrotic cytopathic changes induced by a lytic virus may be suppressed by treatments involving manipulations with the innate immunity without decreasing efficiency of viral reproduction.

The proposal on the virus-activated suicidal necrotic cellular program is in line with the recent discovery of physiologically important cell-encoded necrotic programs, cumulatively referred to as necroptosis or regulated necrosis (18, 25, 62). An intracellular complex involving protein kinases RIP1 and RIP3 belonging to the receptor-interacting protein (RIP) family, the ripoptosome, can mediate, depending on conditions, either necrotic or apoptotic death (24, 58). The murine cytomegalovirus (a DNA-containing virus of the herpesvirus family) was reported to induce RIP3-dependent and RIP1-independent necrosis, which can be suppressed by a viral protein (vIRA) (61). Also, RIP3 may mediate necrotic response after activation of the Toll-like receptor 3 (TLR3) by double-stranded RNA (32). However, the bona fide ripoptosome or any other RIP3-dependent mechanisms are hardly directly involved in the effects described here, because HeLa cells, the host cells used in the present study, were reported to be deficient in RIP3 (24, 33). Hence, the search for another functionally similar innate immunity mechanism within the sophisticated multicomponent network of cross-talking cell death pathways (25) should be a challenging aim of further research.

On the basis of the presently reported and previous data, we propose a model, implying that virus infection may induce a defensive suicidal program possessing two branches, necrotic and apoptotic, which are under control by a common upstream element/mechanism (Fig. 8). The implementation of this program depends on a variety of factors. In HeLa cells infected with wt MV, the necrotic branch of the program is dominating due to the antiapoptotic activity of L (Fig. 8B). If L is incapacitated, e.g., by mutations, the apoptotic program is activated, while the development of necrosis is competitively suppressed (Fig. 8C). If the implementation of apoptotic pathway is interrupted by a chemical inhibitor after the fate-determining decision in the L⁻-infected cells is made, the infected cell survives or at least its death is delayed (Fig. 8D).

Fig 8.

A model of the host defensive response to the MV infection. (A) The infection activates a defensive suicidal program possessing two competing branches, apoptotic and necrotic; (B) upon wt virus infection, the necrotic subprogram dominates, owing to the antiapoptotic function of L; (C) in the absence of functional L, the balance is switched in favor of apoptosis; (D) infection with L-deficient MV mutants in the presence of QVD initiates the apoptotic program, but its execution is prevented by the caspase inhibitor, resulting in suppression or delay of pathological changes.

One of the known mechanisms of competition between apoptosis and necrosis consists in the ability of caspase-8 to cleave CYLD (45), a deubiquitylase and a key component of the programmed necrotic pathway (34). In this regard, it is appropriate to note that although infection of HeLa cells with the Zf-mut did result in the activation of caspase-8 (52), the suppression of necrosis in our system could not be ascribed to this activation because it was observed in the presence of a pan-caspase inhibitor QVD.

Surely, the model shown in Fig. 8 presents an idealized and oversimplified picture, because the conditions of the present study obviously did not represent the complete viral and host disarmament. It does not take into account the existence of another security protein, 2A. Some of the pathological changes in the cells infected with the L mutants in the presence of QVD might likely be due to the activity of 2A, which, as already mentioned, was implicated in the inhibition of host cell translation (29, 56). This protein was also reported to exhibit an antiapoptotic activity (13). Very recently, a novel EMCV-specific protein, 2B*, encoded in an alternative reading frame, has been discovered (43). This protein fulfils the proposed criteria for a good security protein candidate (4) and hence could also contribute to the outcome of virus/cell struggle. In addition, viral proteins other than the security proteins (in particular, protease 3C^pro) may be involved in antidefensive cell-damaging functions (reviewed in reference 63). Furthermore, host defenses include numerous potentially harmful activities other than the apoptotic pathway. Elucidation of the roles of these additional viral and cellular functions is a promising goal of further investigations.

It should be kept in mind that some cell damage may be advantageous to viral reproduction and hence may be more directly associated with the efficiency of viral reproduction. Thus, lysis of the infected cells obviously helps externalization of the progeny of nonenveloped viruses, though even such viruses appear to know tricks allowing them to pass through the plasma membrane without cell killing (57).

Considering a more general situation, it is necessary to emphasize that switching between necrotic and apoptotic pathways in virus-infected cells may be controlled by a variety of host and viral factors, environmental conditions, and in particular viral security proteins. In this regard, it is appropriate to note that L protein of another cardiovirus, Theiler's murine encephalomyelitis virus (TMEV), may exhibit either antiapoptotic or proapoptotic activity in the strain-specific, cell-specific, and context-specific manner (23, 46, 55). The ability of poliovirus to elicit apoptotic response upon productive infection is markedly cell specific as well (6, 28, 38, 42, 51, 60). The 2A security protein of poliovirus was reported to trigger apoptotic response upon its individual ectopic expression (12), whereas it appears to be antiapoptotic in the context of the viral infection (11, 35). Extrinsic factors also affect the choice of specific branches of defensive death pathways (59). All these circumstances point to the complexity of the mechanisms of the virus/host struggle and the necessity to take into account their striking virus and host dependence.

Although virus-induced necrotic and apoptotic CPEs appear to involve host cell-encoded pathways, manifestations of these pathologies may vary not only in a cell-dependent but also in virus-dependent manner. In this regard, it is noteworthy that the plasma membrane blistering similar to that described here was observed in cells infected with another cardiovirus (64) but not with L-lacking picornaviruses, e.g., poliovirus. The gross appearance of cells undergoing cytopathic changes upon infection with different viruses may markedly vary. For example, nuclear deformation in HeLa cells productively infected with poliovirus is more profound compared with that in cardiovirus infection (51, 52). It is tempting to hypothesize that a significant contribution to the diversity of CPE patterns is made by the variability in the sets and properties of the viral security proteins (4).

The results reported here have obvious implications for the pathogenesis and treatment of viral diseases. One of the important corollaries is that amelioration of virus-triggered pathology does not obligatorily require inhibition of viral reproduction but may be achieved by disarmament of the virus and/or its host (cell or organism). On the other hand, natural or drug-induced deficiency of host defenses or viral antidefenses may result in viral persistence, which may or may not be accompanied with significant pathology.