2A Proteinase of Human Rhinovirus Cleaves Cytokeratin 8 in Infected HeLa Cells*

Rhino- and enteroviruses encode two proteinases, 2A and 3C, which are responsible for the processing of the viral polyprotein and for cleavage of several cellular proteins. To identify further targets of the 2A proteinase of human rhinovirus serotype 2 (HRV2), an in vitro cleavage assay followed by two-dimensional electrophoresis was employed. Cytokeratin 8, a member of the intermediate filament group of proteins, was found to be proteolytically cleaved in vitro by the 2A proteinase of HRV2 and of coxsackievirus B4 and in vivo during HRV2 infection of HeLa cells. The cleavage results in removal of 14 amino acids from the N-terminal head domain of cytokeratin 8. However, other intermediate filament proteins (cytokeratins 7 and 18 and vimentin) were not cleaved in the course of the HRV2 infection. Compared with the processing of the eucaryotic translation initiation factors 4GI and 4GII, cleavage of cytokeratin 8 occurs late in the infection cycle at the time of the onset of the cytopathic effect. and for modification of host cell proteins. By proteolytically attack-ing host proteins, viruses interfere with cellular metabolism and structural organization to promote expression of their genome.

Upon infection, many viruses express specific proteinases that are essential both for maturation of viral polypeptides and for modification of host cell proteins. By proteolytically attacking host proteins, viruses interfere with cellular metabolism and structural organization to promote expression of their own genome. Proteolytic processing is essential for viral replication, since picornaviruses are small positive strand RNA viruses that contain a single long open reading frame. The RNA is translated into a large polyprotein precursor, which is cleaved into the mature viral proteins by a sequence of proteolytic cleavages. In the case of rhino-and enteroviruses, at least two viral proteinases are involved in processing of the primary translation product. The first proteolytic step is the cleavage of the polyprotein by 2A proteinase (2A pro ), 1 which cleaves between the C terminus of VP1 and its own N terminus to separate the capsid protein region from the nonstructural protein precursor (1). All remaining cleavages are carried out by the 3C proteinase (3C pro ) or its precursor 3CD except for the cleavage of VP0 to the capsid proteins VP4 and VP2, which occurs during viral maturation.
Since these proteinases are highly specific (2,3), it is to be expected that only a small number of cellular proteins will be cleaved by these proteinases. This is indeed the case, as was observed in extracts of poliovirus-infected HeLa cells examined by two-dimensional gel electrophoresis (4). Using purified enzymes, it has been demonstrated that poliovirus 3C pro cleaves transcription factors TFIIIC and TFIID (the TATA-binding protein) (5,6), CREB (7), and the transcription activator Oct-1 (8). 3C pro also cleaves the microtubule-associated protein MAP-4, which is thought to contribute to the collapse of microtubules late in the infection (9,10). In addition, the 3C pro of foot-and-mouth-disease virus was shown to cleave histone H3 (11).
Action of 2A pro on host cell proteins early during infection has been studied extensively. The 2A pro of rhino-, polio-, and coxsackieviruses are responsible for the cleavage of the eucaryotic translation initiation factor 4G (eIF4G, formerly p220) leading to shut-off of host cell protein synthesis (12)(13)(14). Cleavage by 2A pro of these viruses occurs at the identical amino acid sequence of eIF4G. It results in separation of the N-terminal domain of eIF4G, which binds the cap-binding protein eIF4E, from the C-terminal region, which binds eIF4A, eIF3, and hence the 40 S ribosomal complex (for review, see Ref. 15). As a consequence, cellular cap-dependent initiation is inhibited. Translation of uncapped viral RNA, however, remains intact and is even stimulated as viral protein synthesis is initiated internally at the internal ribosome entry segment (IRES) (16 -18). Indeed, IRES-driven translation appears stimulated under conditions in which eIF4G is cleaved (19 -21). During foot-andmouth-disease virus infection, cleavage of eIF4G is mediated by the leader proteinase, which cuts at a different amino acid sequence at a site close to that of 2A pro (22,23).
Recently, a homologue of eIF4G has been identified and termed eIF4GII (24). eIF4G has therefore been renamed eIF4GI. In HeLa cells infected with human rhinovirus serotype 14 (HRV14) or poliovirus, cleavage of eIF4GII lags behind that of eIF4GI (25). Cleavage of both eIF4GI and eIF4GII is required for complete inhibition of cap-dependent cellular translation (26). Another cellular target of 2A pro is the 70-kDa poly(A)-binding protein (PABP). In HeLa cells infected with coxsackievirus B3 (CVB3), cleavage of PABP starts at a time when cleavage of eIF4GI has been completed (27). PABP is also specifically degraded during poliovirus infection and can be cleaved in vitro by CVB3 2A pro and by both 2A pro and 3C pro of poliovirus (28). Recently, it was demonstrated that 2A pro of CVB4 cleaves dystrophin of heart muscle myocytes in vitro. Cleavage was also found in the heart of mice infected with CVB3 (29). This is proposed to lead to the disruption of the cytoskeleton of heart muscle myocytes by disconnecting actin filaments from the membrane-bound dystrophin.
In this paper, we describe the identification of cytokeratin 8 (K8) as a target of proteolytic cleavage in HeLa cells infected with human rhinovirus serotype 2 (HRV2). This cleavage is also observed in cytoplasmic extracts incubated with highly purified 2A pro from HRV2 or from CVB4. Scission with both proteinases occurs at the same site in the head domain of K8 at a distance of 14 amino acids from the N terminus. Compared with proteolysis of the eIF4G homologues, K8 is cleaved late in the infection cycle. Other intermediate filament proteins such as K18 and vimentin remain uncleaved upon HRV2 infection. In this context, it is important to emphasize that the N-terminal domain of K18, the dimerization partner of K8, is cleaved during adenovirus infection, whereas K8 is not affected by the adenoviral proteinase (30). Cleavage of K18 has been shown to result in dramatic changes of cell morphology, leading to the cytopathic effect of adenovirus infection. Since K8 and K18 are involved in pair formation during filament assembly, cleavage within the N-terminal region of K8 by 2A pro of HRV2 might similarly contribute to changes in the cytoskeletal network late in HRV2 infection and in this way facilitate virus release from the infected cell.

EXPERIMENTAL PROCEDURES
Cells and Viruses-HeLa cells, strain Ohio (ATCC CCL2.2) were grown in Dulbecco's modified Eagle's medium containing 10% heatinactivated fetal calf serum (Life Technologies, Inc.). Infection of HeLa cells and preparation of HRV2 was carried out as described (31).
HeLa Cell Cytoplasmic Extracts-All manipulations were performed at 4°C. 8 ϫ 10 7 HeLa cells grown in four 15-cm Petri dishes were washed twice with ice-cold phosphate-buffered saline containing 1.5 mM CaCl 2 and 1.5 mM MgCl 2 and once with phosphate-buffered saline without Ca 2ϩ and Mg 2ϩ . The monolayer was scraped off, taken up in phosphate-buffered saline, and centrifuged at 300 ϫ g for 5 min. The pellet was washed with 2 ml of homogenization buffer (250 mM sucrose, 3 mM imidazole, pH 7.4) and centrifuged again. Cells were resuspended in 1 ml of homogenization buffer using a 1-ml pipette tip and were lysed by passing through a 22-gauge needle attached to a syringe several times. Disruption of cells was monitored by phase-contrast microscopy. Nuclei were pelleted by centrifugation at 1000 ϫ g for 10 min. The supernatant was then centrifuged for 30 min at 100,000 ϫ g in a Beckman TLA100.3 rotor. The clear cytoplasmic supernatant was frozen at Ϫ70°C or used for in vitro cleavage assays directly. Protein concentration was determined using the Pierce BCA system (32).
2A pro Digestion of HeLa Cell Cytoplasmic Extract and Two-dimensional Gel Electrophoresis-The expression and purification of recombinant 2A pro of HRV2 and 2A pro of CVB4 was as described (21). Cytoplasmic extract containing 1 mg of total protein was diluted 1:1 with buffer A (50 mM NaCl, 50 mM Tris-HCl, pH 8.0, 1 mM EDTA), brought to a final concentration of 5 mM dithiothreitol, and incubated for 2 h at 37°C with 20 g of purified 2A pro of HRV2 or 2A pro of CVB4 in a total volume of 600 l. After cleavage, proteins were precipitated with methanol/chloroform (33). High resolution two-dimensional gel electrophoresis was carried out using the Protean II electrophoresis system (Bio-Rad). Samples containing cytoplasmic proteins were dissolved in 10 M urea, 4% CHAPS, 0.5% SDS, 100 mM dithiothreitol supplemented with 2% (v/v) ampholytes (Merck). Insoluble material was removed by centrifugation at 14,000 ϫ g. Isoelectric focusing was performed at 15,500 V-h in a stepwise fashion (2 h at 200 V; 3 h at 500 V; 17 h at 800 V) in 4% acrylamide (Gerbu, Germany), 0.1% piperazine diacrylamide (Bio-Rad) in 1.5 mm ϫ 16 cm tube gels. The gel buffer contained 0.035% Nonidet P-40 and 2% ampholytes (1 volume pH 3.5-10, 1 volume pH 4 -8, 2 volumes pH 5-7). Degassed 20 mM NaOH served as catholyte, and 6 mM H 3 PO 4 served as anolyte. For SDS-PAGE, the extruded tube gels were equilibrated for 3 min in 2.9% SDS, 70 mM Tris-HCl, pH 6.8, 0.003% bromphenol blue and were placed on top of 1.5-mm thick 10% polyacrylamide slab gels. Gels were silver-stained (34) and scanned for evaluation. Protein identification was done by comparing molecular weight/pI data with those of previous experiments (35). For preparative two-dimensional gel electrophoresis, 1 mg of protein was loaded onto a 2.3-mm tube gel. After electrophoresis, the gel was blotted onto an Immobilon-polyvinylidene fluoride membrane (Millipore Corp.) and stained with Coomassie R-250. Spots were cut out and subjected to automated N-terminal sequencing using an ABI 476A sequencer.
Data Base Searching-Peptide sequence manipulations and data base searching were performed using routines supplied by the Swiss Institute of Bioinformatics (available on the World Wide Web) (36).
In Vivo Cleavage Kinetic and Western Blotting-300,000 HeLa cells per well of a six-well plate were infected with HRV2 at a multiplicity of infection of 200 plaque-forming units in minimal essential medium (Dulbecco's modified Eagle's medium containing 2% fetal calf serum and 1.5 mM MgCl 2 ) at 37°C. At the times indicated, the medium was removed, and the cells were lysed by the addition of 100 l of protein sample buffer 3% SDS, 5% ␤-mercaptoethanol, 10% glycerol, 0.04% bromphenol blue 60 mM Tris-HCl, pH 6.8). Proteins were subjected to SDS-PAGE and electroblotted onto polyvinylidene fluoride membranes. Blocking and incubation with antibodies was done using 0.2% Tween 20 and 0.2% I-block (Tropix) in phosphate-buffered saline. Staining using alkaline phosphatase reaction was as described (21).
Peptide Cleavage in Vitro-Cleavage of peptide substrates with 2A pro of HRV2 and 2A pro of CVB4 was carried out as described (3). In competition assays, the peptide eIF4GI (see Table I) was used as the reference substrate. The relative cleavage efficiency (V max /K m ) rel value was calculated as described (37).

Identification of Proteins Susceptible to 2A pro Cleavage in
Cellular Extracts-In order to identify cellular proteins that are cleaved by 2A pro of HRV2, an in vitro cleavage system was used. Cytoplasmic HeLa cell extracts were incubated with purified recombinant 2A pro . A control sample lacking the proteinase was incubated in parallel. The enzymatic activity of 2A pro was monitored by cleavage of eIF4GI on SDS-PAGE followed by immunoblotting. eIF4GI was completely cleaved after 2 h of incubation at 37°C under these conditions (data not shown). At this point, total protein was precipitated and subjected to twodimensional gel electrophoretic analysis. After silver staining, the two-dimensional protein pattern of the 2A pro -treated extract was compared with that of the untreated sample.
As expected, no gross change in the two-dimensional pattern was observed in extracts incubated with 2A pro , indicating that most cytoplasmic proteins are refractory to this proteinase. This is in agreement with the high degree of cleavage specificity of 2A pro of HRV2 observed in previous studies (2,3). However, upon close inspection of the pattern, certain specific changes were noted when comparing the sample incubated with 2A pro with the control sample (2A pro has a molecular mass of 16.2 kDa and migrates out of the gel under these conditions). Fig. 1 shows a selected region of a typical gel pattern. An arrow denotes a prominent new spot in the 2A pro -treated sample corresponding to a polypeptide of a molecular mass of about 52 kDa and an isoelectric point of 5.3. This new species obviously represents the cleavage product of a cellular protein. This result was reproducibly obtained in four independent 2A pro cleavage experiments. To identify the corresponding polypeptide, 2A pro -treated extract was subjected to preparative twodimensional gel electrophoresis, blotted onto a polyvinylidene fluoride membrane, and stained with Coomassie Brilliant Blue. The spot marked in Fig. 1 was cut out and sequenced as described under "Experimental Procedures." The resulting Nterminal peptide sequence XPRAFSSRSY was compared with the Swiss-Prot database and was unambiguously identified as corresponding to amino acids 16 -24 of human cytokeratin 8 (accession no. P05787), a member of the intermediate filament proteins. Since this sequence was generated by N-terminal sequencing, it should represent the PЈ part of the 2A pro cleavage site of the corresponding protein (the nomenclature Pn-P12P1Ј-PnЈ of substrate is that of Schechter and Berger (39), with the scissile bond lying between P1 and P1Ј). Indeed, the derived cleavage site of K8 is VSTS2GPRAFSSRSY, resembling the known consensus cleavage site of 2A pro of HRV2, which is (L/I)xTx2GP (x designates amino acids that are not critical for specificity) (2, 3).
The same experiment was performed using 2A pro of CVB4. At the corresponding position marked in Fig. 1, an identical spot appeared upon incubation with the CVB4 proteinase. This spot was again blotted and cut out. The resulting N-terminal sequence (GPRAFS) identified the corresponding protein as K8 and clearly showed that in vitro 2A pro of CVB4 cleaves K8 at the identical site as 2A pro of HRV2.
K8 Is Cleaved in HeLa Cells during HRV2 Infection-To assess whether K8 is also cleaved in vivo, HeLa cells were infected with HRV2 and harvested at different times from 1 to 10 h postinfection. Cells were lysed in the presence of SDS, and extracts were subjected to SDS-PAGE. Western blots of these samples obtained with antibodies against eIF4GI, eIF4GII, K8, K18, vimentin, and 2A pro of HRV2, respectively, are shown in Fig. 2. Under these conditions, eIF4GI and eIF4GII are cleaved completely at 4 h postinfection (Fig. 2, A and B). There is no discernible difference in the rate of cleavage of the two proteins. The cleavage of K8 starts at 6 h after infection and continues up to 10 h, when approximately 50% of K8 is cleaved (Fig. 2C). This corresponds to the time frame between the onset of the cytopathic effect, when the infected cells start to round up, and the final stage, when cells detach and lyse, releasing virus into the medium. By 10 h, a large percentage of the cells is already lysed, resulting in the progressive loss of intracellular proteins (data not shown). Cleavage of K8 (molecular mass of 53.5 kDa) by 2A pro removes 14 amino acids of the head domain, resulting in a cleavage product with a calculated molecular mass of 51.9 kDa, which agrees with the size of the band obtained in the Western blot. As a control, the same blots were probed with antibodies against K18 (Fig. 2D) and the intermediate filament protein vimentin (Fig. 2E). These proteins were clearly not cleaved by 2A pro during infection. Similarly, a control employing antibody 8.13, which is capable of recognizing K1, K5, K6, K7, K8, K10, and K11, indicated cleavage of only K8 (data not shown). Fig. 2F shows the amount of 2A pro produced during infection. At 6 h after infection, a clear band is visible, representing about 20 ng of proteinase as judged from comparison with 2A pro -spiked cellular extracts in Western blotting (data not shown).
Thus, in contrast to cleavage of eIF4GI and eIF4GII, which starts at an early time when 2A pro could not even be detected by our anti-2A pro antibody, cleavage of the cytokeratin K8 commences at a later point when a high amount of 2A pro has accumulated in the infected cell. Obviously, the translation initiation factors eIF4GI and eIF4GII, which are the targets for the shut-off of host cell protein synthesis, are much more susceptible to 2A pro than K8, which is a component of the cytoskeletal network.
Kinetics of in Vitro Cleavage of K8 by 2A pro of HRV2-Due to the limited resolution of the two-dimensional gel electrophoretic system, no spot corresponding to the uncleaved K8 could be unambiguously identified in the protein pattern. In order to quantify the extent of in vitro cleavage of K8 by 2A pro and to correlate it with the time course of eIF4GI cleavage, SDS-PAGE of 2A pro digests of cell extracts was performed, followed by Western blotting with anti-K8 antibody. Fig. 3 shows the kinetics of the cleavage reaction using 60 g of HeLa cell cytoplasmic extract and 500 ng of purified 2A pro of HRV2 in a total volume of 60 l. Cleavage of K8 starts at 4 h of incubation. When the amount of 2A pro was raised to 2 g, most of the K8 was processed after 24 h of incubation. This cleavage was specific for 2A pro , since a 24-h incubation with buffer alone did not lead to degradation of K8. Again, controls using antibodies against K18, K7, or vimentin did not show any cleavage of these proteins (data not shown). As seen in Fig. 3B, cleavage of eIF4GI is visible already at 5 min of incubation, leading to a total cleavage of eIF4GI after 1 h. The order of in vitro cleavages thus resembles that observed in the HRV2-infected cells. It is noteworthy that after a 24-h incubation with the high amount of 2A pro , two of the primary cleavage products of eIF4GI are further processed and appear to co-migrate with the fastest running species. This has been observed previously (23).
Similar digestion experiments were also carried out with 2A pro of CVB4. Previous studies have shown that both 2A pro of HRV2 and CVB4 cleave eIF4GI at the identical site of the amino acid sequence (14). In order to compare the two protein-ases with respect to their cleavage activity for K8, HeLa cell extracts were incubated with purified 2A pro of HRV2 and 2A pro of CVB4, respectively. Indeed, both 2A pro of HRV2 and 2A pro of CVB4 were found to produce the same size K8 cleavage product as indicated by SDS-PAGE (Fig. 4). N-terminal sequencing of the K8 fragment (see above) clearly showed that in fact both proteinases use the same cleavage site on K8.
Kinetic Measurement of Substrate Specificity in Peptide Cleavage Assays-In order to determine whether the difference in cleavage efficiency between eIF4GI and K8 is defined by the amino acid sequence at the cleavage site, synthetic peptides were employed as substrates for 2A pro of HRV2 and for 2A pro of CVB4. The peptides used are summarized in Table I. Peptide eIF4GI (termed p220-1 in Ref. 2) spans the 2A pro cleavage site of human eIF4GI. This peptide was taken as a reference in competition experiments. The PABP peptide represents the amino acid sequence of the cleavage site of PABP for 2A pro of CVB3 (28) (see Table I). The K8 peptide corresponds to the cleavage site of K8 as determined in this paper.
Peptides K8 and PABP were used in competition experiments against the eIF4GI peptide as the reference substrate. Fig. 5 shows the results. The efficiency of cleavage by both proteinases drops from the eIF4GI peptide ((V max /K m )rel ϭ 1.0) to K8 (0.76 for 2A pro of HRV2, 0.54 for 2A pro of CVB4), whereas the PABP peptide is hardly cleaved under these conditions (Ͻ0.03 for both 2A pro of HRV2 and 2A pro of CVB4). Subtle differences in substrate specificities between 2A pro of HRV2 and 2A pro of CVB4 are observed, but the order of cleavage rates is the same for both enzymes. Thus, the peptide corresponding to the cleavage site of K8 is processed rather efficiently, although not as rapidly as that corresponding to the cleavage site of eIF4GI. DISCUSSION In the life cycle of picornaviruses, a well ordered sequence of cleavages of viral and cellular proteins occurs. The hierarchy of cleavages is responsible for consecutive processing of the viral protein precursors as well as for the pleiotropic effects on the host cell, e.g. the reduction of cellular transcription and the shut-off of host cell protein synthesis.
In previous studies of poliovirus infection of HeLa cells, it has been demonstrated by two-dimensional gel electrophoretic analysis that only nine acidic and five basic cellular proteins were degraded (4). By following the protein pattern of infected cells, it was, however, impossible to discriminate whether these cleavages were carried out by the 2A pro or by the 3C pro enzyme. In order to identify cellular targets of 2A pro of HRV2, we have performed digestion experiments of a HeLa cell extract using highly purified recombinant 2A pro followed by two-dimensional gel electrophoretic analysis. As expected, the overall two-dimensional protein pattern did not change greatly, and only a few changes were observed following 2A pro treatment. How- ever, one strong spot reproducibly appeared in the two-dimensional pattern of the 2A pro digest, indicating the formation of the cleavage product of a major cellular protein. Using Nterminal sequencing, this spot was identified as a fragment of cytokeratin 8, one of the intermediate filament proteins. As a result of the cleavage by 2A pro , 14 amino acids of the head domain of K8 were removed. Even more significantly, it was demonstrated that the same size K8 fragment was also produced late in HRV2 infection of HeLa cells. Both 2A pro of HRV2 and 2A pro of CVB4 yield a K8 fragment of the same size upon digestion of a HeLa cell extract. The cleavage site of 2A pro for both proteinases as derived from the N-terminal sequence of the K8 fragment is VSTS2GP. It matches the known consensus sequence ((L/I)xTx2GP) for cleavage at the positions P2, P1Ј, and P2Ј. Position P4 in the K8 cleavage site is occupied by valine, which is chemically similar to leucine/isoleucine at P4 of the 2A pro consensus cleavage sequence. Amino acid changes at P1 and P3 are known to have only minor effects on 2A pro cleavage efficiency (2,3). Because of the conserved amino acid pattern, it is likely that cleavage of K8 is caused by a direct action of 2A pro rather than by an indirect one via activation of a cellular proteinase. Comparison of the processing rates of peptides corresponding to the cleavage sites indicates that the eIF4GI peptide is cleaved somewhat more efficiently than the K8 peptide. However, this small difference is clearly not sufficient to explain the large discrepancy between the early cleavage of eIF4GI and the late cleavage of K8 in the infected cell. This indicates that the relative rates of protein cleavage are also determined by factors other than the amino acid sequences at the respective cleavage sites.
At early times in HRV2 infection, comparison of the cleavages of eIF4GI and of eIF4GII shows that both homologues are cleaved at similar rates. Complete cleavage of both eIF4GI and eIF4GII has occurred by 4 h postinfection. This is in contrast to results obtained upon infection of HeLa cells by poliovirus or by HRV14, where eIF4GI is cleaved much faster than eIF4GII (25). However, it is not surprising to find differences in the kinetics of cleavage between 2A pro of HRV2 and 2A pro of HRV14, since the latter resembles the polioviral enzyme in terms of its substrate specificity. In fact, the proteinases of both HRVs differ to the extent that 2A pro of HRV2 cannot process a cleavage site peptide containing the PЈ site of HRV14 (3,40).
Cleavage of cytokeratin 8 starts at the time of the onset of the cytopathic effect when cells round up and start to detach from the surface. Cytokeratins are members of the intermediate filament family that together with actin filaments and microtubules form the cytoskeleton. Depending on the type of epithelial cells, a combination of acidic type I cytokeratins K9 -K20 and of the basic type II cytokeratins K1-K8 is found. Cytokeratins contain a conserved ␣-helical central region (rod domain, ϳ310 amino acids) and nonhelical head and tail domains varying greatly in size and sequence (Fig. 6). The mechanism of filament polymerization is a stepwise process; acidic cytokeratins align with basic cytokeratins in a 1:1 molar ratio to form a coiled-coil dimer. Tetramers are generated by side by side aggregation of dimers. The tetramers then polymerize end to end to form a protofilament, with eight protofilaments building up the 10-nm filament. It is of importance that cytokeratins with N-terminal head deletions are still capable of coiled-coil interactions and higher lateral interactions but are deficient in protofilament elongation (41). In HeLa cells, K8 and K18 constitute the major cytokeratin species. Since the basic type II K8 is known to form a pair with the acidic type I K18, the K8-K18 heterodimer therefore constitutes the main building block of the HeLa cell cytokeratin filaments.
Previous work on the effect of viral infection on the cytoskeleton of the host cell has been carried out mostly with adenovirus. It has been observed that during the late phase of the adenovirus infection, cytokeratin K18 is cleaved by the L3 23-kDa proteinase (30). It has been shown previously that the N-terminal deletion of 83 amino acids from K18 prevents filament elongation (42). The adenoviral proteinase removes 73 amino acids from the N terminus of the head domain of K18 (Fig. 6). The collapse of the intermediate filament network at the end of the replication cycle causes the loss of mechanical stability of the cells. This is considered to be an important step in the release process of progeny adenovirus (43). In poliovirus infection of HeLa cells, alterations in the cytoskeletal network were described already in 1979, but the molecular basis of this effect has not been elucidated (44). Recently, it was demonstrated that 2A pro of CVB3 cleaves dystrophin in myocytes and  in the hearts of CVB3-infected mice (29). This cleavage is proposed to disrupt the interaction of the cell membrane with actin filaments. As hereditary forms of dystrophin abnormalities were shown to cause cardiomyopathies, this protease cleavage event is thought to contribute to dilated cardiomyopathy observed in enteroviral infections. Furthermore, the aspartatetype human immunodeficiency virus proteinase was shown to cleave in vitro the intermediate filament proteins vimentin (Fig. 6), glial fibrillary protein, and desmin (45,46). Microinjection of the enzyme into fibroblasts causes the collapse of vimentin filaments (46). Thus, specific cleavages introduced in cytoskeletal proteins by viral proteinases may be a frequent mechanism of promoting viral infection.
The cleavage of the cytokeratin K8 by 2A pro during infection of HeLa cells with HRV2 is a highly specific process, since other intermediate filament proteins were not cleaved (e.g. cytokeratin K7, K18, and vimentin). As a result, the head domain of K8 is shortened by 14 amino acids. What is the functional significance of the truncation of K8? Previous experiments with poliovirus-infected HeLa cells have indicated that destruction of the cytoskeletal network by treatment with cytochalasin D and nocodazole has no effect on virus yield in single cycle infections (47). It is therefore unlikely that the cleavage of K8 by 2A pro of HRV2 affects viral replication directly. However, since K8 and K18 form heterodimers, it is tempting to speculate that HRV2 and adenovirus may use similar strategies to destabilize the integrity of the host cell and thus to promote the spread of the virus. The truncation introduced in K18 by the adenoviral proteinase is much more drastic than the cleavage in the head structure of K8 by the 2A pro of HRV2. Nevertheless, it must be kept in mind that at the time of the cytopathic effect about 50% of K8 is cleaved (Fig. 2). Furthermore, due to the shut-off of protein synthesis early in the infection cycle, cleaved cytokeratins cannot be replaced by newly synthesized intact molecules, which may further augment the effect of cleavage of 2A pro on K8 (43). In this context, it may also be of significance that K8 can be phosphorylated at Ser 23 , which is close to the cleavage site of 2A pro . Alternatively, N-terminal truncation of K8 may affect cytoskeletal scaffolding by disrupting the interaction with linker proteins that cross-link cytokeratin filaments and other cytoskeletal components (48).
The data presented prove that 2A pro is indeed a multifunctional enzyme. In addition to its role in processing of the viral polyprotein and in replication (49), it is responsible for the modification of cellular proteins both at early and late stages of the HRV2 infection cycle. The early function of 2A pro in the host cell shut-off of protein synthesis is well established. The new results on the cleavage of cytokeratin 8 indicate that 2A pro also acts late in the infection at the time of the cytopathic effect when major morphologic changes occur in the infected cell. As indicated by the cleavage of K8 by 2A pro of CVB4, this mechanism may also operate in enteroviruses. Proof of the general validity of this concept clearly must await further experimentation.