Viruses require the use of the host cell translation machinery to produce viral proteins in an infected cell. Thus, most viruses have evolved with strategies to effectively compete with cellular mRNAs for ribosomes and a limited number of translation initiation factors. The mechanisms by which caliciviruses manipulate cellular translation had not been established. Here, we report evidence that host cell translation is inhibited during calicivirus replication and demonstrate that PABP, a key protein involved in the translation of polyadenylated cellular mRNAs, can be cleaved and apparently inactivated by the calicivirus 3CLpro
. The NoV 3CLpro
cleaved PABP similarly to PV 3Cpro
, including a recognition of PABP associated with ribosomes that is more efficient than for PABP associated with a less-purified cytoplasmic fraction. This observation indicates that both enzymes selectively target PABP associated with the translation apparatus, possibly to inhibit PABP function in translation. We recently reported that PV 3Cpro
selectively inhibits poly(A)-dependent translation by removal of the CTD of PABP (40
). This cleavage inhibits translation by a unique mechanism without affecting PABP interactions with eIF4GI or the poly(A) tail, since N-terminal cleavage fragments of PABP retain intact RRMs that bind RNA (40
). It was proposed that cleavage of PABP likely inhibits translation by disrupting interactions with eIF4B and eRF3 and possibly affects late events in translation such as ribosome recycling (40
). The striking similarity in the cleavage specificities between the PV and NoV proteinases suggests that study of both virus systems could be useful in further defining common themes used by these viruses to manipulate cellular translation.
We examined a cultivatable calicivirus, FCV, to verify the occurrence of cellular PABP cleavage during replication. Antibodies generated against human PABP cross-reacted with feline PABP, enabling the identification of a PABP cleavage fragment in FCV-infected cell lysates. Of interest, the FCV 3CLpro
(rProPol) cleaved both feline and human PABP similarly, and the cleavage site was provisionally mapped to a site on human PABP rarely utilized by PV 3Cpro
. This alternate cleavage site (3Calt′) is only 24 amino acids away from the 3Calt site, thus a similar effect on PABP function would likely occur with cleavage by either viral proteinase. Examination of cleavage site sequences does not reveal an obvious reason for the differential selection of PABP cleavage sites by the various proteinases. PV 3Cpro
cleavage depends on both conformational constraints and primary sequence which includes a preference for A/V-X-X-Q/G in the P4 through P1′ positions (44
). The sequences surrounding the three PV 3Cpro
cleavage sites on PABP are different; however, all contain Q/G or Q/T dipeptides (Table ) (41
). NoV 3CL recognizes amino acid pairs with E or Q in the P1 position (45
) and, therefore, might share specificity for Q/G pairs in PABP with PV 3Cpro
. The cleavage pair recognized in human PABP by FCV 3CL is likely Q437
/G, although FCV 3CL apparently prefers cleavage sites with E (Table ). However, its ability to recognize cleavage sites with Q in the P1 position has been also demonstrated (57
). In addition, comparison of the sequences adjacent to known cleavage sites of FCV showed a preference for a large, hydrophobic amino acid residue in the P4 position (55
), and the presence of a tryptophan residue in the P4 position of the PABP Q437
/G cleavage site may provide further explanation for efficient recognition of this particular site by FCV 3CL (Table ). Another factor in this recognition could be the folding of the PABP that provides the FCV 3CL proteinase access to this particular cleavage site. It is likely that the cleavage site selection by these proteases requires conformational constraints and perhaps even unknown cofactors. Despite the differences in PABP cleavage site selection, the ability of FCV to cleave PABP strongly suggests that this translation inhibition mechanism occurs during an authentic calicivirus infection. It may prove useful to develop homologous reagents for further study of the effect of FCV replication on cellular proteins in permissive feline cells.
Similar to PV 3Cpro
, the NoV r3CLpro
did not cleave eIF4G in an in vitro assay. It remains possible that eIF4GI may be cleaved during NoV infection by a viral or cellular proteinase, but our in vitro studies suggest that its cleavage is not essential for the translation inhibition associated with the PABP cleavage. We could not rule out the absence of cleavage of eIF4GI during FCV infection because the anti-human and -mouse eIF4G antibodies we tested did not recognize feline eIF4G in CRFK cells (data not shown). Furthermore, it should be noted that FCV (2
) and other caliciviruses such as RHDV (33
) induce apoptosis in infected cells, and this process could also affect the cellular translation machinery.
The mechanism of viral mRNA translation initiation during calicivirus infection remains unclear. The calicivirus genome is capable of serving as an mRNA but lacks either a 5′-end cap structure or any identified internal ribosomal entry site. The genome contains a VPg protein covalently linked to the 5′ ends of both the genome and the subgenomic RNAs (13
), which may play a role in translation initiation by recruiting the 40S ribosomal subunit through interactions with translation initiation factor eIF3 (17
). In addition, the calicivirus VPg itself contains sequence homology with certain translation initiation factors, indicating that it might directly compete with cellular host factors for binding to the ribosome (57
). Both genomic and subgenomic calicivirus RNAs possess a poly(A) tail at the 3′ end, and a recent study has shown that the 3′-end nontranslated region of the Norwalk virus genome with a poly(A) tail of 24 nt binds directly to PABP (25
). However, it is not clear whether PABP binding of the calicivirus poly(A) tail stimulates translation of viral mRNAs since there is no 5′ cap structure. PABP may interact with VPg at the 5′ end of calicivirus RNA in an analogous fashion; however, this has not been tested. Since the poly(A) tail itself stimulates translation of capped and internal ribosomal entry site-containing mRNAs, it will be important to determine whether poly(A) tails stimulate VPg-dependent translation of calicivirus mRNAs. However, if the poly(A) tail is indeed important for viral mRNA translation, it might be predicted that 3CLpro
-mediated cleavage of PABP could also inhibit viral protein synthesis. A mechanism would be needed to allow discrimination between polyadenylated cellular and viral RNAs in infected cells so that host cell mRNAs are selectively inhibited during early phases of infection. It is possible that cleavage of PABP may also be a requirement for the switch between translation and RNA replication, converting the viral genome from a translation-competent to replication-competent status. The elucidation of these mechanisms for both caliciviruses and picornaviruses will be relevant to the understanding of how viral messages are translated at such high efficiency in infected cells and yet allow RNA replication to begin.
In summary, we show that calicivirus 3CLpro cleaves PABP in vitro and during infection. We propose that the calicivirus 3CLpro specifically targets a pool of PABP involved in cellular protein translation. Furthermore, this cleavage likely inhibits cellular mRNA translation by removal of CTD from PABP independently of eIF4G cleavage. Cleavage of PABP may be a strategy used by caliciviruses to down-regulate cellular mRNA translation in infected cells to more effectively synthesize viral proteins. The cleavage of PABP by viral proteinases may be a common mechanism used by different viruses to gain control of host cell mRNA translation, and this may be crucial for the successful life cycle of these RNA viruses.