Enterovirus 71 3C Protease Cleaves a Novel Target CstF-64 and Inhibits Cellular Polyadenylation

1Enterovirus 71 (EV71) belongs to the family of Picornaviridae, to which poliovirus also belongs. EV71 infection usually causes childhood herpagina or exanthema, which is also called hand, foot, and mouth disease (HFMD). Acute EV71 infection may also cause a severe polio-like neurological disease and significant mortality. EV71-related neurological complications include aseptic meningitis, brainstem and/or cerebellar encephalitis, acute flaccid paralysis (AFP), myocarditis, and rapid fatal pulmonary edema and hemorrhage, which have been observed during outbreaks in Taiwan, mainland China, Malaysia, Brunei, Singapore, western Australia, the United States, and Europe –.

2Numerous host machineries are affected by picornaviral infection, including host cap-dependent translation, and transcription,. Viral proteases are responsible for inhibitory effects. Picornaviruses typically encode two viral proteases, 2Apro and 3Cpro, which are important for viral polypeptide processing. The major catalytic sites of type I poliovirus and EV71 3Cpro are His40, Glu71, and Cys147 –. In general, enterovirus 3C proteinases cleave at Gln/Gly scissile pairs. Picornaviral 3Cpro can enter nuclei through its precursor 3CD′ or 3CD, which contains a nuclear localization sequence (NLS),, and can cleave some cellular transcriptional factors or regulators, such as TATA-box binding protein, p53, Histone H3 and transcription factor IIIC,, –, offering insight into the effects of these factors or regulators on the transcriptional machinery of the host. Therefore, the identification of other cellular proteins that are cleaved by EV71 3Cpro in nuclei is of interest, as it may help us to understand unknown viral-host interactions during EV71 infection. In this work, a novel nuclear factor, CstF-64, that is cleaved by EV71 3Cpro, is identified.

3Poly(A) tails are important for both cellular mRNA and picornaviral mRNA. The poly(A) tail in eukaryotic cells is suggested to confer mRNA stability, promote the translational efficiency of mRNA, and the transport mRNA from the nucleus to the cytoplasm –. The poly(A) tail of poliovirus is critical to viral replication,. However, poly(A) for picornaviruses is obtained differently from that for host cells. The poly(A) tail of picornaviruses is encoded by the genome, copied into a 5′ poly(U) tract in the negative sense form, then back into a poly(A) tail during the virus RNA replication cycle. Polyadenylation of most eukaryotes involves a series of steps. Prior to poly(A) synthesis, a polyadenylation signal located at the 3′ end of precursor mRNA (pre-mRNA) must be recognized by multiple factors such as cleavage factors I and II (CF I and II), cleavage/polyadenylation specificity factor (CPSF) and cleavage stimulation factor (CstF) complex, which causes endonucleolytical cleavage at the site of polyadenylation,. One of these factors, cleavage stimulation factor, 3′ pre-RNA, subunit 2, 64 kDa (CstF-64) is responsible for recognizing the second polyadenylation sequence element, a G/U-rich motif which is down-stream of the polyadenylation sites on pre-mRNA. In vitro and in vivo blockage of CstF-64 impairs the 3′-end pre-mRNA processing and polyadenylation,.

4This work identified eight cellular nuclear proteins as potential targets for EV71 3Cpro cleavage by two-dimensional (2D) electrophoresis and MALDI-TOF analysis. CstF-64, the target selected from these eight proteins, is further examined. CstF-64 in the nuclear extract was verified by western blot to be degraded upon EV71 3Cpro treatment. The reduction and cleavage pattern of CstF-64 in EV71-infected cells were studied. The EV71 3Cpro cleavage sites in CstF-64 were also mapped by performing an in vitro cleavage assay. The 3′-end pre-mRNA processing in host cells was monitored during EV71 infection. An in vitro assay was also conducted to test the effect of 3Cpro on the cellular 3′-end pre-mRNA processing and polyadenylation. The findings herein are the first to demonstrate that a picornaviral protease interferes with host gene expression at the machinery of polyadenylation.

5Picornavirus 3Cpro proteins can enter the nuclei of host cells,. To identify potential substrates of EV71 3Cpro in nuclei, in vitro 3Cpro cleavage assay was conducted and the proteomic technique was utilized to identify the cellular targets. First, the catalytic activities of purified recombinant EV71 wild-type 3Cpro or mutant 3Cpro(C147S) were verified using a [35S]-labeled peptide substrate, which contains a 3Cpro cleavage site, as described elsewhere. The result shows that the wild-type, but not the C147S mutant of 3Cpro cleaves the viral substrate (Fig. 1A). The recombinant EV71 wild-type 3Cpro or mutant 3Cpro were then added to nuclear extracts from SF268 cells. Following incubation for 4 hours at 37°C, the reactions were subjected to 2D electrophoresis. The experiments were performed six times and the results were analyzed using PDquest 7.0 (Bio-rad). Fig. 1B presents one of these six 2D electrophoresis experiments. Proteins that appeared in the mutant 3Cpro-treated reactants but in at least half of the amounts of the wild-type 3Cpro-treated reactants, based on silver staining, were regarded as potential targets. From the six pairs of gels in these 2D gel experiments, eight proteins that yielded similar results at least three times were selected. These proteins were then identified by MALDI-TOF mass spectrometry and the results were summarized in Table 1. In contrast with that of mutant 3C-treated nuclear extract, CstF-64 of the wild-type 3Cpro-treated nuclear extract was decreased to nearly undetectable levels in five individual 2D electrophoresis gel (Fig. 1C). Arrows in Figs. 1B and 1C indicate the locations of CstF-64 on 2D gels. These results suggest that CstF-64 is a potential substrate for EV71 3Cpro.

6The reduction of CstF-64 by proteomic assay was confirmed by using an anti-CstF-64 antibody to detect the level of CstF-64 in nuclear extracts from SF268 cells, which were incubated with wild-type or mutant recombinant 3Cpro. The amount of CstF-64 in SF268 nuclear extract dramatically declined following incubation with wild-type 3Cpro (Fig. 2A, lane 3), and the intact form of CstF-64 was clearly detected in mutant 3Cpro-treated nuclear extract (lane 2). One potential cleavage product of 55 kDa was also detected using CstF-64 antibody in wild-type 3Cpro-incubated nuclear extract (lanes 3, 6 and 9, indicated by *). CstF-64 levels in nuclear extracts from other cell lines, RD and HeLa cells, also declined upon incubation with wild-type 3Cpro and the cleavage pattern is similar to that in SF268 nuclear extracts. (Compare lanes 6 and 9 with results for mutant 3Cpro-treated nuclear extracts in lanes 5 and 8). The wild-type 3Cpro-induced CstF-64 reduction in nuclear extract depends on dose of recombinant 3Cpro (Fig. 2B, lanes 3–7). The intensities between the full-length CstF-64 and its 55 kDa cleavage product were not the same as expected. This discrepancy is attributed to that later in the experiments, the cleavage sites were mapped to position 251 and multiple positions close to the C-terminus (around position 500) of CstF-64. The cleavage of CstF-64 at positions around 500 produce a product of 55 kDa, which contains another 3Cpro cleavage site at its Gln251. When position 251 was mutated (Gln to Ala), the mutant CstF-64 was cleaved into a 55 kDa product with the same intensity as that of full-length CsfF-64. Fig. 2A shows a band faster than 24 KDa (lane 2, 5 and 8), which we speculate is a cleavage product from a degraded form of CstF-64. The degraded form of CstF-64 might contain the cleavage sites for 3Cpro as well as the CstF-64 antibody recognition sites. Therefore, western blot analysis detected the cleaved product from the degraded form only in mutant 3Cpro, but not in wild-type 3Cpro-treated nuclear extract because the cleavage destroyed the antibody recognition of CstF-64. The other EV71 viral protease, 2Apro was also tested to determine whether it could degrade CstF-64 in vitro. The catalytic activities of recombinant EV71 2Apro were firstly elucidated by treatment with RD cell lysate and were verified by western blot for a known 2Apro substrate, eIF-4GI, to be able to be cleaved in vitro (Fig. 2C, lane 1–3). The western blot for CstF-64 in the nuclear extracts that were treated by purified recombinant 2Apro demonstrated the inability of 2Apro to induce CstF-64 degradation in vitro (Fig. 2C, lane 4–6). These results indicate that viral 3Cpro, but not 2Apro is responsible for reducing CstF-64 in vitro.

7In vitro assays demonstrated that EV71 3Cpro can cleave CstF-64; the degradation of CstF-64 in EV71-infected cells was, therefore, examined. A time-course study was conducted to monitor the levels of CstF-64 in RD cells infected with EV71 at a multiplicity of infection (m.o.i.) of 40. Total cell lysates harvested from EV71-infected RD cells at various hours post-infection (h.p.i.) were analyzed by western blot assay. Infection reduced the amount of host CstF-64 proteins from 6 to 10 hours post-infection (h.p.i.) (Fig. 3A, lanes 4, 6 and 8). One potential CstF-64 cleavage product of 55 kDa (indicated by *) which resembles that obtained by in vitro cleavage, began to be detected at 6 h.p.i. (lane 4). EV71 3Cpro in the same infected cells was also observed at 6 h.p.i. (Fig. 3A, lane 4), and its level increased markedly from 8 to 10 h.p.i. (Fig. 3A, lane 6–8), in a manner that is related to the degradation times of CstF-64 in infected cells. The reduction of CstF-64 and the appearance of 55-kDa product were also observed in cells infected with a lower titer (m.o.i of 1) of EV71 (Fig. S1), but with a 2 hour delay. To detect other cleavage products from CstF-64, a CstF-64 that was fused with FLAG at its N-terminal was overexpressed in EV71-infected cells and detected using FLAG antibody at 6 and 8 h.p.i.. During EV71 infection, FLAG-CstF-64 was cleaved into products of size 55 kDa and 30 kDa (CP1 and CP2 in Fig. 3B). The 55 kDa products from FLAG-CstF-64 reduced at late point of EV71 infection (Fig. 3B, CP1, compare lanes 2 and 4), similar to the result from endogenous CstF-64 in EV71-infected cells (Fig. 3A, the bands marked as * in lane 4 and 6). On the other hand, another cleavage product of about 30 kDa was increased during virus infection (Fig. 3B, CP2 compare lanes 2 and 4). The 55 kDa intermediate cleavage products suggest that CstF-64 may be cleaved at more than one cleavage site in EV71-infected cells.

8Picornavirus infection can relocalize numerous nuclear proteins to the cytoplasm –. To determine whether CstF-64 also undergoes changes in sub-cellular localization, the location of CstF-64 in EV71-infected was monitored by immunohistostaining using a confocal microscope. The result demonstrated that most CstF-64 remains in the nucleus of the infected cells at 6 to 10 h.p.i. (Fig. 3D), whereas the other nuclear factor, heterogeneous nuclear ribonucleoprotein K (hnRNP K), is redistributed to the cytoplasm (Fig. S2). Our 3C Ab was unable to be used in confocal experiment; therefore to determine whether EV71 3Cpro was also present in the nucleus, EV71-infected cells were fractionated into cytoplasmic and nucleus fractions. Western blot demonstrated that EV71 3Cpro is equally distributed across cytoplasmic and nuclear fractions (Fig. 3C, lanes 2 and 4), indicating that EV71 3Cpro can enter into a host nucleus during virus infection, as can other picornaviruses. CstF-64 was detected to be localized primarily in the nuclei of uninfected cells (Fig. 3C, lane 3 compared to lane 1), consistent with previous findings and confocal data presented herein (Fig. 3D). However, CstF-64 in both nuclei and cytoplasm were cleaved during virus infection (Fig. 3C, lanes 2 and 4).

9The results presented above suggest that the CstF-64 was reduced in wild-type 3Cpro-treated nuclear extracts but not in mutant 3Cpro-treated nuclear extracts (Fig. 2A). Therefore, 3Cpro is reasonably hypothesized to cause the reduction of CstF-64 by proteolytic cleavage.

10Details of the cleavage were elucidated using a [35S]-labeled CstF-64 generated by in vitro transcription and translation (TNT) as a substrate in the in vitro 3Cpro cleavage assay. After it had incubated with recombinant 3Cpro, [35S]-labeled CstF-64 can be cleaved into fragments of approximately 25 and 30 kDa (Fig. 4A, lane 3). Since the size of these two peptides do not correlate with the full-length size of CstF-64 (64 kDa compared to 25 kDa+30 kDa = 55 kDa), at least one cleavage product smaller than 10 kDa that could not be detected in PAGEs was expected to exist. To identify the 3Cpro cleavage sites on CstF-64, [35S]-labeled CstF-64 and a series of truncated CstF-64 peptides that were designed according to functional domains of CstF-64 (Fig. 4B) were generated in vitro by TNT for 3Cpro cleavage assay. Fig. 4B summarizes the size of these cleavage products, estimated from Fig. 4A. The peptides 221–409 and 410–557 were both cleaved suggesting that CstF-64 included at least two cleavage sites. The 20 kDa peptide 410–557 was cleaved into the product of 10 kDa, suggesting that a cleavage site was located close to the C-terminus of CstF-64. Peptide 1−220 (25 kDa) was not cleaved but peptides 221−409 and 110−409 were both cleaved into a product of about 17 kDa, suggesting that one cleavage site can produce a product of 30 kDa from the N-terminal of CstF-64. The cleavage of peptide 221−469 supported this conclusion. In summary, these cleavage results suggest the existence at least two potential cleavage sites around the amino acid position 250 and 500 of the CstF-64 protein.

11A high amount of 3Cpro (5 µg) was used in above in vitro cleavage assay, and the incubation time was fixed at 3 hours (Fig. 4A). An attempt was made to detect the cleavage intermediates by performing in vitro kinetics analysis of [35S]-labeled CstF-64 cleavage by recombinant 3Cpro. According to Fig. 4C, three cleavage products began to appear at 15 min. A 55 kDa product resulted from cleavage at positions around 500, while two products of 30 (front) and 35 (rear) kDa both resulted from cleavage at position around 250. A new 25 kDa product began to appear at 60 min, in addition to the three previously detected products. This 25 kDa segment is attributed to the further cleaving of either the aforementioned 35 (rear) kDa product at positions around 500 or the 55 kDa product at position around 250. As the incubation time exceeded 2 hours, only the 25 and 30 kDa products dominated and became the final cleavage products. Various concentrations of 3Cpro were added to the reactions (Fig. S3), and similar cleavage products were observed, as shown in Fig. 4C.

12Picornavirus 3Cpro generally cleaves peptides at the Gln/Gly junction. Based on the cleavage results presented above, an analysis of the amino acid sequence of CstF-64 was conducted, revealing that numerous Gln/Gly junctions may have potential 3Cpro cleavage sites on CstF-64 (Fig. 5A). They are Gln251, corresponding to the predicted cleavage site in the region of amino acid 250, and residues Gln483, 496, 505, 510, 515 at around the amino acid position 500 on CstF-64. To determine which Gln/Gly junction on CstF-64 is the actual 3Cpro cleavage site, an attempt was made to produce mutant [35S]-labeled CstF-64 at putative Gln/X cleavage sites. Since two cleavage sites around the amino acids positions 250 and 500 of CstF-64 were estimated, cleavage at position around 250 is expected to produce a product of ∼55 KDa. However, as is predicted, a CstF-64 in which both 250 and 500 sites are mutated would be fully withstand 3Cpro cleavage (Fig. 5B). Following 3Cpro-treatment, the CstF-64 with the mutation of Gln251 into Ala (Q251A) yielded a cleavage product of 55 KDa (Fig. 5C, lanes 1–3), which is consistent with the predicted product when the cleavage site around amino acid position 250 is blocked (Fig. 5B). This result suggests that the Gln251 of CstF-64 is the only 3Cpro cleavage site around the position 250 in vitro. To identify the other cleavage sites around the amino acid position 500 of CstF-64, CstF-64(Q251A) was further mutated at the Gln483, 496, 505, 510 and 515 amino acids into Ala (Q483A, Q496A, Q505A, Q510A and Q515A, respectively). The results indicate that CstF-64(Q251A) peptides with single mutations on these potential cleavage sites are not completely resistant to wild-type (WT) 3Cpro treatment (Fig. 5C, lanes 4–18). A mutant CstF-64(Q251A) with Q483A, Q496A, Q505A, Q510A, and Q515A is fully resistant to 3Cpro treatment (Fig. 5D, lanes 4–6), suggesting that the existence of multiple cleavage sites around the amino acid position 500 of CstF-64 in the five residues, Gln483, 496, 505, 510 and 515. Moreover, the cleavage pattern was examined in a mutant CstF-64 (CstF-64-5 m), in which all Gln/Gly sites around position 500 were mutated. This observation indicates that CstF-64-5 m was cleaved into two detectable products (Fig. S4): a) 30 kDa product, i.e. the N-terminal part of CstF-64 after 3Cpro cleaves it at Gln/Gly of position 251 and b) the 35 KDa product, i.e. the C-terminal part of CstF-64 when 3Cpro cleaves it at Gln/Gly of position 251, but not at Gln/Gly sites near position 500. Taken together, above results demonstrate that Gln to Ala mutations at position 251 and multiple sites around position 500 let CstF-64 become fully resistant to EV71 3Cpro cleavage.

13CstF-64 is an important factor for host 3′ pre-mRNA processing and RNA polyadenylation,. Results presented above demonstrate the reduction of CstF-64 protein in EV71-infected cells; therefore, whether the machinery of host cell polyadenylation is affected by EV71 infection was further examined. Initially, the pre-mRNA processing in EV71-infected cells was assessed using an artificial exogenous expressed GFP RNA. The pEGFP plasmid contained a SV40 polyadenylation signal sequence, which was commonly used in the study of cellular mRNA processing and polyadenylation. Primers that target GFP pre-mRNA and an oligo-dT primer that targets GFP poly(A) mRNA were used in RT-PCR for detecting RNA species (pre-mRNA and mature mRNA with poly(A) tail) in EV71-infected cells (Fig. 6A). A set of primers that target the GFP coding region, which can theoretically detect both pre-mRNA and mRNA with a poly(A) tail, were also utilized as controls to demonstrate the total amount of GFP RNA in cells. RD cells were infected with EV71 after they were transfected with eGFP plasmid. Additionally, by adopting those primers designed in Fig. 6A, RT-PCR analysis was performed for semi-quantifying different GFP RNA species in these infected cells (Fig. 6B). After they were calibrated to the total amount of GFP, the results revealed the accumulation of unprocessed pre-mRNA and a reduction in polyadenylated mRNA in EV71-infected cells.

14An attempt was made to identify if EV71 infection affected endogeneous pre-mRNA processing. Interlekin 10 receptor beta (IL-10RB) was selected as the target because, in an earlier study, the poly(A) mRNA of IL-10RB was found to be decreased in EV71-infected cells by cDNA microarray. The relative amounts of IL-10RB pre-mRNA and poly(A) mRNA were monitored by real-time RT-PCR. According to those results, the IL-10RB pre-mRNA in EV71-infected cells decreased to 62.95% of that in mock-infected cells, which correlates with our previous micorarray results. Additionally, the relative amount of IL-10RB pre-mRNA increased to 156% of that in mock-infected cells (Fig. 6C). Above results suggest that EV71 infection impairs the CstF-64-related 3′ pre-mRNA processing mechanism.

15The results herein suggest that EV71 3Cpro cleaves CstF-64, potentially inhibiting cellular pre-mRNA 3′-end formation and polyadenylation. An in vitro assay was conducted to test the effect of EV71 3Cpro on HeLa nuclear extract. The substrate employed in this assay is a capped pre-mRNA, which contains an SV40 late gene polyadenylational cleavage site. A comparison with the mutant 3Cpro-treated nuclear extracts (Fig. 7A, lane 5 and Fig. 7B, lane 5) or untreated nuclear extracts (Fig. 7A, lanes 2, 3 and Fig. 7B, lanes 2, 3) shows that the cleavage of pre-mRNA and polyadenylation proceeded efficiently. However, upon wild-type 3Cpro treatment, the treated nuclear extracts lost the ability to perform pre-mRNA cleavage (Fig. 7A, lane 4) and polyadenylation (Fig. 7B, lane 4). Moreover, the impairments were rescued by adding purified recombinant CstF-64 protein (Fig. 7A, lane 6 and Fig. 7B, lane 6). The results suggest that EV71 3Cpro targeting to CstF-64 is a factor that inhibits cellular 3′-end pre-mRNA processing and polyadenylation.

16Picornaviral 3Cpro reportedly influences numerous cellular functions by cleaving various host proteins, which mechanism is important in shutting off cellular gene expression in transcriptional and translational levels,, –. Here we show that EV71 3Cpro may affect another cellular mechanism, polyadenylation, since a novel substrate, CstF-64, was identified using the proteomic approach. CstF-64 is a crucial element in cellular mRNA maturation. This 64 kDa component plays a role in the CstF complex that binds to the polyadenylation signal region of pre-mRNA. Our in vitro cleavage studies revealed that 3Cpro cleaves CstF-64 at amino acid position 251 and around 500 close to the C-terminal domain (Figs. 4 and 5). The in vitro data also demonstrated the blockage of pre-mRNA 3′-end processing and the polyadenylayion activities of the nuclear extract upon 3Cpro-treatment (Fig. 7). The inhibitory effect of EV-71 3Cpro resembles that of the blockage of CstF-64 by the specific antibody. The results together suggest that EV71 3Cpro cleaves CstF-64 and thus inhibits the 3′-end processing of cellular pre-mRNA. Our semi-quantitative RT-PCR results support this claim because the cellular 3′-end pre-mRNA processing was inhibited in EV71-infected cells (Fig. 6B). Pre-mRNA and mature poly(A) mRNA in Fig. 6B were detected individually by using two pairs of primers. Different primers have varying sensitivities, making it difficult to achieve a quantitative correlation between the increase in precursor and decrease in polyadenylated RNA. However, a quantivitive real-time RT-PCR result (Fig. 6C) provides further evidence to demonstrate that the pre-mRNA processing or polyadenylation (either an exogenous RNA or an endogenous RNA) was impaired in EV71-infected cells.

17Other proteins than CstF-64 may be involved in EV71-induced inhibition of pre-mRNA processing in virus-infected cells even though the levels of poly(A) polymerase in EV71-infected cells were found to be similar to those of the mock-infected cells (Fig. S5). However, our in vitro results in Fig. 7 demonstrate that CstF-64 is the target factor that caused 3Cpro to affect 3′-end processing and polyadenylation. Adding recombinant CstF-64 can restore the function of 3C-incubated nuclear extracts on 3′-end pre-mRNA processing and polyadenylation (Lane 6 in both Fig. 7A and B).

18Before the surge of the proteomic method, a substrate for a protease is commonly identified using western blot analysis. The proteomic approach has the advantages of the ability to identify multiple candidates simultaneously with no limitation by antibody availability. However, 2D electrophoresis and the silver staining are occasionally unable to detect proteins visible in western blot analysis. This study identified eight novel proteins as the substrates for EV71 3Cpro via the proteomic approach, in which no previously known substrates for other picornaviral 3Cpro were available, such as TATA-box binding protein, p53, Histone H3 and transcription factor IIIC,,,. This might be due to that silver staining and western blot differ in sensitivity. Moreover, the strategy adopted in Fig. 1 is set only in a single condition, i.e. the recombinant 3Cpro was incubated with nuclear extract from SF268 cells for 4 hours. More potential substrates could be identified if several interaction conditions were tested, e.g., varying in incubation time and nuclear extracts from other cell lines.

19The in vitro cleavage assay revealed that 3Cpro cleaves [35S]-labeled CstF-64 protein into detectable products of 25 kDa and 30 kDa (Fig. 4B). However, CstF-64 in 3Cpro-treated nuclear extract yielded 55 kDa products detected by western blot (Fig. 2). A product similar to that of 55 kDa was also present in EV71-infected cells (Fig. 3A). The size of the 55 kDa products above is correlated to that of the 3Cpro cleavage product when the cleavage site of the amino acid position 251 on CstF-64 is mutated (Fig. 5C). The detection of FLAG in FLAG-CstF-64 overexpressed cells revealed that the proportion of 30 kDa increased and that of 55 kDa declined upon EV71-infection (Fig. 3B). A FLAG-CstF-64 with Q251A mutation was transfected into EV71-infected cells and the results indicated the accumulation of 55 kDa products but no 30 KDa product was detected (data not shown), suggesting that the 55 kDa component was an intermediate cleavage product when 3Cpro cleaved CstF-64 at an amino acid position ∼500. The production of multiple 55 kDa bands in infected cells (Figs. 3A and B) was also observed, explaining the multiple cleavage sites around the amino acid position 500 of CstF-64. In addition, the results of in vitro kinetics of CstF-64 cleavage by 3Cpro (Fig. 4C and Fig. S3) showed a cleavage intermediate of 55 kDa. In summary, these results showed the similarity between the cleavage patterns obtained in the in vitro and in vivo assays suggests that 3Cpro may have reduced CstF-64 levels via proteolytic activity in infected cells.

20Single mutation at position around 500 in CstF-64 still made it susceptible to EV71 3Cpro cleavage, while multiple mutations cause CstF-64 become resistant to the cleavage. Analytical results indicate the C-terminal of CstF-64 may contain multiple cleavage sites; otherwise, these mutations could lead to a loss of exposure of the cleavage site.

21This study also attempted to examine whether C147S mutation in EV71 impairs the viral inhibition of cellular pre-mRNA processing and polyadenylation. However, C147S mutation rendered EV71 lethal. Alternatively, transcripts derived from EV71 infectious clones, i.e. wild-type versus C147S mutant, were transfected into cells to examine how they affect cellular pre-mRNA processing and polyadenylation. In wild-type transcript -transfected cells, CstF-64 was cleaved and cellular pre-mRNA processing was inhibited (data not shown), the same phenomenon was observed in EV71 virus-infected cells. However, in C147S mutant transcript -transfected cells, no cleavage occurred in CstF-64 and no effect on cellular pre-mRNA processing was observed, which is similar to the results in mock-infected cells. Still, no detectable viral protein (mature 3Cpro) was found. We believe that phenomenon is owing to that C147S mutation destroys the proteolytic activity of 3Cpro and inhibits viral protein processing and maturation. Therefore, we can not conclude that the C147S mutant in EV71-infected cells impairs the viral inhibition of cellular pre-mRNA processing.

22CstF-64 reportedly contains at its N-terminal an RNA-binding domain and a hinge domain, which are responsible for the reorganization of RNA and its binding to CstF-77 or other factors –. No 3Cpro cleavage was detected in this region (1–220), suggesting that the cleaved CstF-64 maintains functions RNA binding and interaction with CstF-77 (Fig. 4B). However, the Pro/Gly-rich domains of the C-terminal region of CstF-64 form an extended α-helix structure and are suspected to be able to interact with several transcription factors, including binding of the transcription factor PC4 via the 100 amino acids of C-terminal CstF-64. One study noted that the truncation of the C-terminal domain (529–577) of CstF-64 inhibited cellular 3′ pre-mRNA processing. Accordingly, multiple cleavage by 3Cpro in C-terminal domain (around 500th a.a.) of CstF-64 is expected to destroy the cellular pre-mRNA processing functions of CstF-64.

23The RNA virus-induced inhibition of 3′-end pre-mRNA processing has been observed in the influenza A virus. NS1A of influenza A reportedly inhibits the cellular 3′-end pre-mRNA processing by interacting with CPSF-30 which is within the same machinery co-factor as CstF-64. Similar to those in EV71-infected cells, the inhibition of the cellular 3′-end pre-mRNA processing by influenza A virus causes the loss of poly(A) mRNA and the accumulation of un-poly(A) pre-mRNA. This inhibitory effect of NS1A has also been suggested to be involved in the countering of cellular antiviral responses, including virus-induced interferon β production. The capacity of NS1A to bind to CPSF-30 may also be crucial to viral replication because of high interferon-β mRNA production in NS1A mutant virus-infected cells.

24Picornaviruses influence host-cell gene expression by the inhibition of cellular transcription and cap-dependent mRNA translation –. However, several cellular genes could escape the inhibition of gene expression by picornavirus infection. Our previous cDNA microarray analysis for total cellular RNA demonstrated that the level of some RNAs, related to chemokines, protein degradation, complement proteins and proapoptosis proteins increased upon EV-71 infection, suggesting leakage from the inhibition of transcription by EV71. Translations of of c-myc, Bip, and eIF4G mRNA have been found to be increased in poliovirus-infected cells as the cap-dependent translation shuts down, because of the presence of internal ribosome entry sites (IRES) –. This work proposes a novel mechanism by which picornavirus inhibits cellular gene expression in the 3′-end pre-mRNA processing step, which would probably inhibit polyadenylation of the surviving host RNAs. It has been reported that the poly(A) tail of poliovirus is essential to viral replication,, which suggests several host poly(A)-associated factors are important for viral growth. Inhibition of cellular poly(A) RNA synthesis may provide yet an additional advantage for virus replication.

25In conclusion, a novel mechanism by which picornavirus inhibits cellular function was identified. CstF-64 was identified as an EV71 3Cpro substrate. The 3Cpro cleavage sites were mapped onto amino acid position 251 and the C-terminal region of CstF-64. The cleavage of CstF-64 impairs the cellular 3′-end pre-mRNA processing and polyadenylaytion. EV71 utilizes 3Cpro to interfere polyadenylation of host cellular RNA; however, its own poly(A) synthesis is not affected since picornaviral poly(A) tail is genetically encoded,.

26RD and SF268 cells were maintained in DMEM that contained 10% FBS (Glibco) at 37°C. Cells were split in a ratio of 1∶10 in fresh medium every three days. After the RD cells had grown to 80% confluence, they were washed in PBS. EV71/4643/TW viruses with a multiplicity of infection (m.o.i.) of 40 were inoculated into the cells in DMEM without FBS at room temperature for one hour to infect the cells with the virus. Following one hour of incubation, viruses and the medium were replaced by DMEM that contained 2% fetal bovine serum. Infected cells were cultured at 35°C.

27pET-23-EV71-3C and pET-23-EV71-m3C-C147S plasmid constructs, which contain wild-type 3Cpro or mutant 3Cpro (with Cys147 replaced by Ser) cDNA in pET-23a(+) vector for recombinant EV71 3Cpro production have been constructed elsewhere. For CstF-64 recombinant production, cDNA of CstF-64 was cloned from the Cst-64 coding region that contained the plasmid, pZ64–18, which was a gift from Dr. James L. Manely, and then inserted into vector pET23a(+) using EcoRV and NotI. For the production of [35S]-labeled substrates by TNT assay, full-length and partial cDNA of CstF-64 were cloned and inserted into pcDNA3.1(+) using EcoRV and NotI. For the production of FLAG-CstF-64 in transfected cells, cDNA of CstF-64 was cloned into pFLAG-CMV-2 using EcoRV and NotI. The mutant CstF-64 construct was produced using a QuikChange Site-Directed Mutagenesis Kit (Stratagene). Plasmid pG3SVL-A, which contained sequence -142 to +138 of SV40 late gene poly(A) cleavage sites was kindly provided by Dr. James L. Manely and Dr. Yoshio Takagaki, for the in vitro 3′-end pre-mRNA processing and polyadenylation assay.

28PVDF membrane was blocked wih Tris-buffered saline/0.1% (vol/vol) Tween 20 that contained 5% non-fat dry milk and probed with the indicated antibody. Antibodies against CstF-64 (1∶200; from Dr. Clinton C. MacDonald or 1∶2000; Santa Cruz), FLAG (1∶2000; SIGMA), PCNA (1∶2000; Santa Cruz), HDAC (1∶2000; Santa Cruz),Poly(A) polymerase (1∶2000; Santa Cruz), actin (1∶4000; Chemicon) and EV71 3Cpro monoclonal antibody which generated from recombinant 3Cpro protein by our lab (1∶50) were used. Following washing, the membranes were incubated with HRP-conjugated anti-mouse or HRP-conjugated anti-rabbit (1∶2000). HRP was detected using a Lighting Chemiluminescence reagent (Amersham Pharmacia).

29To perform the cleavage assay in a cell-free system, SF268, HeLa or RD cells were washed in PBS and scraped out. Packed cells were re-suspended in hypotonic buffer (10 mM HEPES pH 7.9, 1.5 mM MgCl2, 10 mM KCl and 0.5 mM DTT) and homogenized using a 25G needle. Homogenized cell fragments were centrifuged at 3300 g to remove cytoplasmic proteins. The pellet was washed twice in hypotonic buffer and the soluble nuclear proteins were extracted by adding buffer with a graded conc. of KCl (20 mM HEPES, 25% glycerol, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT, 0.02 to 1.6 M KCl at pH = 7.9). After centrifuged at 25000 g, the extracted nuclear proteins in supernatant were harvested and dialyzed in buffer (20 m M HEPES, 20% glycerol, 100 mM KCl,0.2 mM EDTA and 0.5 mM DTT at pH = 7.9) and then stored at −80°C.

30Commercial reagents for nuclear extraction (CMN compartment protein extraction kit; Biochain) were applied to prepare the nuclear extract from EV71-infected cells.

31Commercial HeLa nuclear extract (Santa Cruz) was used in in vitro 3′-end pre-mRNA processing and polyadenylation assay.

32To produce recombinant 3Cpro and CstF-64 proteins, plasmid pET-23-EV71-3C, pET-23-EV71-m3C-C147S and pET-23-CstF-64, were introduced into competent E. coli BL21 (DE3 pLysS) and protein expression was induced using 40 µM isopropyl b-D-thiogalactopyranoside. 3C-His fusion proteins were purified using a HiTrap kit (Pharmacia).

33To produce [35S]-labeled protein, TNT Coupled Reticulocyte Lysate Systems (Promega) was used in the in vitro transcription and translation reaction.

34In the proteomic approach to screening potential 3Cpro substrates of, 30 µg of 3Cpro was incubated with 250 µg nuclear extracts in digestion buffer (20 mM HEPES, 20% glycerol and 100 mM KCl at pH = 7.9) with total volume of 100 µl at 37°C for four hours. In western blot assay, 10 µg of 3Cpro was incubated with 50 µg nuclear extract under the aforementioned conditions. To cleave recombinant CstF-64, 10 µg recombinant CstF-64 proteins interacted with 10 µg of 3Cpro for four hours in the same buffer as in the earlier experiments. To cleave the [35S]-labeled substrate, 4 µl of labeled protein from one TNT assay reaction was incubated with 5 µg of 3Cpro in buffer (20 mM HEPES, 20% glycerol and 100 mM KCl at pH = 7.9) with a total volume of 15 µL at 37°C for two hours.

352D gel electrophoresis was conducted using the IPGphor system (Amersham Biosciences). After the nuclear extracts had incubated with 3Cpro and precipitated using 20% trichloroacetic acid (TCA), the sample was dissolved in 8 M Urea, 2% CHAPS, 0.5% IPG buffer (Amersham Bioscience) and bromophenol blue. These samples were loaded on a cup-loading system (Amersham Biosciences). Isoelectric focusing was performed at 45,000 Vh in a stepwise fashion (1500 V, gradient for 2 hours; 4000 V, gradient for 3500 V.hours; 8000 V, step-on-hold until the end) in Immobiline Drystrip IEF gels with the range of pH 3–10 (Amersham Biosciences). For SDS-PAGE, the IEF gels were loaded on gels with graded acryamide conc. of 9% to 16%. Proteins on 2D electrophoresis gels were visualized by silver staining and the results were scanned as TIF files. Protein spots on scanned gels were analyzed using PDquest v7.0 (Bio-rad) software. Six pairs of gels with a similarity of over 60% were selected for further analysis. Spots that appeared in the mutant 3Cpro-treated nuclear extracts disappeared or became at least 50% smaller upon silver staining of the wild-type 3Cpro-treated nuclear extracts were selected as the potential target substrates for wild-type 3Cpro. Following this selection strategy, proteins that yielded similar results to each other more than three times were identified by in-gel digestion and analyzed by Bruker Ultraflex MALDI-TOF mass spectrometry. Mass lists were performed peptide mass fingerprinting by Biotool 2.0 software and the algorithm of Masscot (http://www.matrixscience.com).

36RD cells were seeded onto 22 mm diameter coverslips to 80% confluence and were infected with enterovirus 71 (strain 4643/TW/1998) at a multiplicity of infection of 40. At each time point, the culture medium was removed and the cells were washed with PBS, and fixed with 3.7% formaldehyde for 20 min at room temperature. The cells were then washed with PBS and permeabilized using 0.3% Triton X-100 for 5 min at room temperature. Cells were washed once using PBS containing 2% FBS, incubated in blocking solution for 1 hr at room temperature. In CstF-64, hnRNP K and EV71 2B immunostaining, cells were incubated with anti-CstF-64 (diluted 1∶100; Santa Cruz), anti-EV71 2B (diluted 1∶200; Provided by Dr. Jim-Tong Horng) for 8 hours at 4°C or anti-hnRNP K (diluted 1∶200; Santa Cruz) for 1.5 hr at room temperature. After they had been washed three times with PBS, coverslips were incubated with FITC-conjugated goat anti-mouse IgG or goat anti-rabbit lgG, and Alexa Fluor 568 goat anti-rat IgG (Invitrogen) for 1 hr at room temperature. The coverslips were then washed once with PBS; treated with nuclear stain Hoechst 33258 (diluted 1∶500) for 15 min; washed three times with PBS, and mounted on glass slides with mounting fluid (75% glycerol in PBS). The images were obtained under a confocal laser-scanning microscope (Zeiss; LSM 510 NLO).

37After they had been transfected with pEGFP-N1 for 1 hour, RD cells were infected with EV71 virus. The total RNA of these infected cells were isolated by TRIZOL Reagent (Invitrogen) and purified by phenol/chloroform extraction. To detect EGFP pre-mRNA, a set of primers 5′- CCGGAATTCTGAGCAAAGACCCCAACGAG -3′ and 5′- CCCAAGCTTAAAATATTAACGCTTACAAT-3′, which target the sequence that was removed from pre-mRNA following polyadenylation was used. To detect poly(A) EGFP mRNA, a set of primers 5′- CCGGAATTCTGAGCAAAGACCCCAACGAG -3′ and 5′- TTTTTTTTTTTTTTTGCAGT-3′, which target the poly(A) tail and polyadenylation site of EGFP mRNA, were used. To detect total EGFP RNA, a set of primers 5′-ATGGTGAGCAAGGGCGAGGA-3′ and 5′-CTTGTACAGCTCGTCCATGC-3′, which target the coding region of EGFP was used. The RT-PCR products were placed in 2% agarose gel with ethidium bromide, and semi-quantified using software FUJIFILM Science Lab 2005 (Fuji).

38Capped RNA substrate with a poly(A) signal of the SV40 late gene was synthesized using a Megascript SP6 kit (Ambion) and Cap analog (Ambion). Plasmid pG3SVL-6, flanked by DraI, was utilized as the template in the in vitro transcription reaction. The synthesized capped RNA substrate was purified by phenol/chloroform extraction before it was used in in vitro 3′-end pre-mRNA processing and polyadenylation assay.

39The nuclear extract was treated with recombinant 3Cpro at 37°C for two hours before it was used in 3′ pre-mRNA processing and polyadenylation assay. After the treatment, the 3Cpro activity was stopped by adding protease inhibit cocktail (Roche). The volume of each 3′ pre-mRNA processing reaction with 20 µl contained 25% (v/v) nuclear extract (25 µg in total), and 2 ng of RNA subtrates with added 1 mM EDTA, 0.5 mM dATP, 10 mM creatine phosphate, 2.5% (v/v) of polyvinyl alcohol. In in vitro polyadenylation, dATP was replaced by ATP, and the conc. of EDTA declined to 0.08 mM with the addition of 1 mM MgCl2. 4 µg of recombinant CstF-64 was added into one of the 3Cpro-treated reaction. After incubation for two hours at 37°C, the assayed RNA was purified phenol/chloroform extraction. The purified RNA was analyzed in 1.5% agarose gel and denatured using formaldehyde.

40Real-time RT-PCR was performed to detect levels of IL-10RB RNA. The High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) was applied for reverse transcription. To differentiate pre-mRNA and polyadenylated IL-10RB RNA, different primers were used in reverse transcription step. A specific primer 5′-CTGAGAGCTCCCAGATGACTGA-3′ that targets the region that contains polyadenylation cleavage signal on IL-10RB was used to detect the pre-mRNA. On the other hand, an oligo-d(T) was used as the primer to detect poly(A)-mRNA. For total mRNA (including pre-mRNA and poly(A)-mRNA), a random hexomer was used an the primer in reverse transcription step. For PCR step, 5′-GAGGGATCAGGGCAGCAA-3′ and 5′-CAGGGTCTGGGAGTTCTAGATGTG-3′ designed by Primer Express software (Applied Biosystems) targeting IL-10RB coding region were used. The PCR reaction was performed on 7500 Real-Time PCR System (Applied Biosystem) using SYBR Green Core Kit (Applied Biosystem).