Aichi Virus 2A Protein Is Involved in Viral RNA Replication

Abstract

The Aichi virus 2A protein is not a protease, unlike many other picornavirus 2A proteins, and it is related to a cellular protein, H-rev107. Here, we examined the replication properties of two 2A mutants in Vero cells and a cell-free translation/replication system. In one mutant, amino acids 36 to 126 were replaced with an unrelated amino acid sequence. In the other mutant, the NC motif conserved in the H-rev107 family of proteins was changed to alanine residues. The two mutations abolished virus replication in cells. The mutations affected both negative- and positive-strand synthesis, the defect in positive-strand synthesis being more severe than that in negative-strand synthesis.

The picornavirus nonstructural 2A protein varies among viruses in amino acid sequence and function. In entero- and rhinoviruses, 2A is a protease responsible for cleavage of the polyprotein at its own N terminus (14, 15). It also cleaves cellular proteins, including eukaryotic translation initiation factor 4G (5). In addition, poliovirus 2A is involved in regulation of viral RNA stability, translation, and negative-strand synthesis (4). Aphthovirus 2A is ∼18-amino-acids (aa) long, and cardiovirus 2A is about 140-aa long. The conserved amino acids, Asn-Pro-Gly (NPG), at the C termini of the aphtho- and cardiovirus 2A proteins, together with a proline at the N terminus of 2B, are required for the processing at the 2A/2B junction through a mechanism different from a proteolytic reaction (2). It has been reported for Theiler's murine encephalomyelitis virus, a cardiovirus, that a large deletion within the 2A-coding region does not affect RNA replication significantly (6). Parechovirus 2A, which has no proteolytic activity (12) nor the NPGP motif, shows specific binding activity to both single- and double-stranded forms of the 3′ untranslated region (UTR), suggesting its involvement in viral RNA replication (10).

Aichi virus (AiV), which is associated with acute gastroenteritis in humans (17), is a member of the genus Kobuvirus of the family Picornaviridae (18). AiV 2A, which is 136-aa long, does not have the protease motif characteristic of enterovirus 2A or the NPGP motif. AiV 2A, as well as parechovirus and avian encephalomyelitis virus 2A, has been reported to be related to a cellular protein, H-rev107, a candidate tumor suppressor protein (3, 13).

First, we investigated whether AiV 2A has a proteolytic activity required for the polyprotein processing. We had previously constructed a plasmid, pMAL-3CDmut, which contains the 3CD-coding region with mutations T6492G and G6493C to abolish the 3C protease activity (9). A Csp45I-PstI fragment (nucleotides [nt] 6480 to 6771) of pMAL-3CDmut was substituted for the corresponding fragment of an Aichi virus replicon, pAV-FL-Luc-5′rzm, in which the capsid-coding region was replaced with a firefly luciferase (Luc) gene and a hammerhead ribozyme sequence was inserted upstream of the viral sequence (7, 8), yielding pAV-FL-Luc-5′rzm-3Cmut (Fig. 1A). pAV-FL-Luc-5′rzm and pAV-FL-Luc-5′rzm-3Cmut were subjected to in vitro translation in the presence of l-[³⁵S]methionine and l-[³⁵S]cysteine (Amersham), using a TNT quick coupled transcription/translation system (Promega). After being incubated at 30°C for 90 min, translation products were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and radioactive signals were detected with a BAS2000 bioimaging analyzer (Fujifilm). The predicted molecular mass of the polyprotein is approximately 240 kDa. As shown in Fig. 1B, the polyprotein processing was observed for pAV-FL-Luc-5′rzm but not for pAV-FL-Luc-5′rzm-3Cmut. This result indicates that in AiV, 3C is the only protease involved in the polyprotein processing and that 2A is not a protease.

FIG. 1.

(A) Schematic diagram of pAV-FL-Luc-5′rzm and pAV-FL-Luc-5′rzm-3Cmut. The virus sequences were cloned downstream of the T7 promoter. An asterisk indicates mutations (T6492G and G6493C) in the 3C-coding region. (B) In vitro transcription/translation of pAV-FL-Luc-5′rzm and pAV-FL-Luc-5′rzm-3Cmut in rabbit reticulocyte lysate. The translation products labeled with l-[³⁵S]methionine and l-[³⁵S]cysteine were analyzed by SDS-PAGE, and radioactive signals were detected. The positions of the molecular weight markers are indicated on the left.

To investigate the function of AiV 2A in virus replication, we constructed two kinds of 2A mutants, using an infectious cDNA clone, pAV-FL (11), and a replicon, pAV-FL-Luc-5′rzm (Fig. 2A). Of the two kinds of introduced mutations, one is a frameshift mutation within the 2A-coding region caused by a 1-nt deletion of nt 3895 and a 1-nt insertion between nt 4170 and 4171. By these mutations, aa 36 to 126 were replaced with an unrelated amino acid sequence encoded by another reading frame. The other mutation is a change of the NC (Asn-Cys) motif, one of the motifs of the H-rev107 family of proteins, to AA (Ala-Ala).

FIG. 2.

(A) Organization of pAV-FL, pAV-FL-Luc-5′rzm, and pAV-FL-Luc-mut9. pAV-FL is an AiV infectious cDNA clone. pAV-FL-Luc-5′rzm is a replicon harboring the Luc gene and a hammerhead ribozyme sequence (shaded box). pAV-FL-Luc-mut9 is a replicon containing the indicated mutations at the 5′ end of the genome. The thick lines and open boxes show the UTRs and coding regions, respectively. The thin lines indicate the vector sequence. The virus sequences were cloned downstream of the T7 promoter. Two mutations were introduced into the 2A-coding region of pAV-FL, pAV-FL-Luc-5′rzm, and pAV-FL-Luc-mut9. In the 2Afs mutant, the sequence unrelated to the wild-type 2A sequence is indicated by a filled box. In the NC-AA mutant, an asterisk indicates the position of the mutated amino acids. (B) Replication of the replicon RNAs in Vero cells. Lysates of Vero cells transfected with each RNA were prepared at the indicated time points after transfection, and the Luc activity in each lysate was measured. Error bars represent the standard deviation for triplicate experiments. RLU, relative light units.

A SacI-XhoI fragment of pAV-FL-Luc-5′rzm was subcloned into pGEM-11Zf, and PCR-based mutagenesis was performed using the derived clone. For the frameshift mutation, a DNA fragment derived by inverse PCR with primers 3894 M (5′-GGCCACCTTGCGGATGGCCCAGTG-3′; nt 3894 to 3871) and 4171P (5′-GTGAAAGCGCTCCCAGGCATCAGG-3′; nt 4171 to 4191) and a DNA fragment amplified by PCR with 3896P (5′-CCGACGGCAGTGCCAAACAGATCT-3′; nt 3896 to 3919) and G4170M (5′-CGGCAACAGCAGCCGAGGCTGCGAT-3′; nt 4170 to 4147; the inserted nucleotide is underlined) were ligated. For the NC-to-AA mutation, inverse PCR was performed using primers 4056 M (5′-GTTGGTGGCGCTGTACTCCCACTTG-3′; nt 4056 to 4032) and NC-AA4057P (5′-GCCGCTACCCACTTCGTCAGCTCCATCACT-3′; nt 4057 to 4086; mutated nucleotides are underlined), and the PCR product was self-ligated. After the nucleotide sequences of the derived plasmids had been checked, the BclI fragment and the SacI-XhoI fragment with each mutation were replaced with the corresponding regions of pAV-FL and pAV-FL-Luc-5′rzm, respectively, yielding pAV-FL-2Afs, pAV-FL-NC-AA, pAV-FL-Luc-5′rzm-2Afs, and pAV-FL-Luc-5′rzm-NC-AA (Fig. 2A). In addition, the EcoRI fragments of pAV-FL-Luc-5′rzm, pAV-FL-Luc-5′rzm-2Afs, and pAV-FL-Luc-5′rzm-NC-AA, which contain the T7 promoter, the hammerhead ribozyme sequence, and the 5′-end 391 nt of the genome, were replaced with the corresponding fragment of pAV-FL-mut9 (11), in which the 6-nt stretches nt 3 to 8 and nt 39 to 44 were exchanged with each other. The generated plasmids were called pAV-FL-Luc-mut9, pAV-FL-Luc-mut9-2Afs, and pAV-FL-Luc-mut9-NC-AA, respectively (Fig. 2A). As negative controls for RNA replication, pAV-FL-3Dmut and pAV-FL-Luc-5′rzm-3Dmut, in which 3D RNA polymerase was inactivated (8), were used. The plasmids were linearized by digestion with HindIII, and in vitro transcripts were synthesized with T7 RNA polymerase.

The growth properties of these 2A mutants were examined. First, the abilities of the mutants to generate viable viruses were investigated. One microgram of an in vitro transcript derived from pAV-FL, pAV-FL-2Afs, or pAV-FL-NC-AA was transfected into Vero cells, using a lipofectin reagent (Invitrogen), and the virus titer at 72 h after transfection was determined by plaque assay as described previously (11). AV-FL RNA generated viable viruses at a titer of 10⁵ PFU/ml, whereas the two mutant RNAs produced no plaques (data not shown).

Next, we examined RNA replication of the 2A mutants in Vero cells. Ten micrograms of AV-FL-Luc-5′rzm RNA, AV-FL-Luc-5′rzm-2Afs RNA, AV-FL-Luc-5′rzm-NC-AA RNA, or AV-FL-Luc-5′rzm-3Dmut RNA was electroporated into Vero cells as described previously (11), and the Luc activities of cell lysates prepared at various times were measured by using a luminometer (Lumat LB9507; Berthold) (Fig. 2B). At 1 h, no difference in the Luc activities between AV-FL-Luc-5′rzm and AV-FL-Luc-5′rzm-3Dmut was found, indicating that at this time point, RNA replication had not been initiated and that the Luc activity at this time represents the translation efficiency of the RNA. AV-FL-Luc-5′rzm-NC-AA RNA showed almost the same translation efficiency as AV-FL-Luc-5′rzm RNA. On the other hand, the Luc activity of AV-FL-Luc5′rzm-2Afs RNA at 1 h was approximately 50% lower than those of other RNAs. At 2 h, the Luc activity of AV-FL-Luc-5′rzm RNA was increased, whereas those of AV-FL-Luc-5′rzm-2Afs RNA and AV-FL-Luc-5′rzm-NC-AA RNA, as well as of AV-FL-Luc-5′rzm-3Dmut RNA, were gradually decreased, showing that the two mutants did not replicate in transfected cells.

Furthermore, translation and negative- and positive-strand syntheses of the mutants were examined by using a cell-free translation/replication system as described previously (8). The translation reaction mixture with a mixture of ∼70% l-[³⁵S]methionine and ∼30% l-[³⁵S]cysteine (Amersham) was incubated for 1.5 or 3 h, and then the labeled translation products were analyzed by SDS-PAGE and detected as described above (Fig. 3A and B). Additionally, proteins in the translation reaction mixtures without labeled amino acids were separated by SDS-PAGE, blotted onto a polyvinylidene difluoride membrane, immunodetected with rabbit antiserum raised against recombinant His-tagged 2A expressed in Escherichia coli, and then visualized by chemiluminescence (Fig. 3C). The predicted molecular mass of 2A is 14.4 kDa, and 2A must be found between 2B (17.5 kDa) and 3AB (13.7 kDa) on SDS-PAGE. However, the predicted amino acid sequences of wild-type 2A and 2A with the 2Afs mutation contain no methionine and only one cysteine, and for 2A with the NC-AA mutation, the cysteine was changed to an alanine. Probably because of this, 2A and the 2A mutant proteins could not be detected when labeled with [³⁵S]methionine and [³⁵S]cysteine (Fig. 3A and B). Upon immunoblot analysis (Fig. 3C), wild-type 2A and 2A with the NC-AA mutation were detected (Fig. 3A and B, lanes 1, 2, 5 to 8, 11, and 12), but the processed 2A with the frameshift mutation was not (Fig. 3A and B, lanes 3, 4, 9, and 10).

FIG. 3.

Translation of AV-FL-Luc-5′rzm RNA and its mutant RNAs (A), and AV-FL RNA and its mutant RNAs (B), using Vero cell S10 extracts. For each translation reaction, aliquots collected at 1.5 and 3 h after incubation were analyzed by SDS-PAGE. (C) Immunoblot analysis of in vitro translation products, using anti-2A antiserum. The positions of the molecular weight markers and viral proteins are indicated on the right and left of each panel, respectively. An arrowhead in panel (A) indicates Luc-2A, and asterisks in panel (B) show the protein bands unique to AV-FL-2Afs.

There was no significant difference in the amount of the processing products such as 3D, 2C, 2B, or 3AB among the RNAs (Fig. 3A and B), indicating that the translation efficiencies of the two 2A mutant RNAs were comparable to that of the wild-type RNA. For AV-FL-Luc-5′rzm-2Afs RNA, polyprotein processing was slower than in the other RNAs (Fig. 3A). In addition, the amount of Luc was decreased, and an approximately 80-kDa protein was accumulated (Fig. 3A, lanes 3 and 4). Also for AV-FL-2Afs RNA, an abnormal processing pattern was observed: a reduced amount of VP1 and two unique protein bands were found (Fig. 3B, lanes 3 and 4). Of the two unique bands, the smaller one (approximately 45 kDa) exhibited an electrophoretic mobility (41.4 kDa) corresponding to that of VP1 (27 kDa) plus 2A (14.4 kDa). Upon immunoblotting, the 80-kDa protein was recognized by anti-2A antiserum (Fig. 3C, lane 4), and the 45-kDa protein was also detected on prolonged exposure (data not shown). These results suggest that the 2Afs mutation affects cleavage at the N terminus of 2A. The 80-kDa protein from AV-FL-Luc-5′rzm-2Afs RNA would be Luc-2A (63 kDa plus 14.4 kDa), and the 45-kDa protein from AV-FL-2Afs RNA would be VP1-2A. The failure to detect the processed 2A with the 2Afs mutation upon immunoblot analysis was probably due to inefficient processing as well as the reduced reactivity of the antiserum to the 2A protein containing the mutation. The decrease in the Luc activity observed in AV-FL-Luc-5′rzm-2Afs RNA-transfected cells (Fig. 2B) may have resulted from the reduced amount of properly processed Luc. The protein that is slightly smaller than L-P1 and that is accumulated only in AV-FL-2Afs (Fig. 3B, lanes 3 and 4) would be P1-2A. The accumulation of P1-2A suggests that the proper processing of P1 requires cleavage at the P1/2A junction beforehand. A conformational change in 2A would affect cleavage at the VP1/2A junction; in turn, the accessibility of the cleavage sites to 3C in the resulting P1-2A may be reduced compared with that in P1. It has been reported for mengovirus that mutations introduced into 2A affect the processing at the VP1/2A junction and the processing of P1-2A (19).

RNAs transcribed from plasmids harboring the ribozyme sequence were subjected to the cell-free translation/replication reaction to analyze RNA replication (Fig. 4A). RNA synthesized in the cell-free translation/replication reaction was labeled with [α-³²P]CTP at 3 to 5 h after the start of the reaction. Then total RNA was extracted and analyzed by nondenaturing agarose gel electrophoresis, and radioactive signals were detected. Negative- and positive-strand syntheses were evaluated as the production of the double-stranded replicative form and single-stranded RNA, respectively. In addition, AV-FL-Luc-mut9 RNA, AV-FL-Luc-mut9-2Afs RNA, and AV-FL-Luc-mut9-NC-AA RNA were analyzed to compare the efficiency of negative-strand synthesis among the RNAs in detail (Fig. 4B). The mutation introduced into mut9 has been shown to abolish positive-strand synthesis without affecting negative-strand synthesis (8). Negative-strand synthesis in the 2Afs mutant and the NC-AA mutant was decreased to 18% and 42% of that in the wild type, respectively (Fig. 4B). On the other hand, positive-strand synthesis in the 2Afs mutant was not detected and that in the NC-AA mutant was markedly reduced (Fig. 4A). Thus, the two mutations affected both negative- and positive-strand synthesis, the defect in positive-strand synthesis being more severe than that in negative-strand synthesis. Since the 2Afs mutation prevented cleavage at the N terminus of 2A (Fig. 3), in addition to the loss of function of 2A caused by the mutation, the decrease in the amount of the properly processed 2A may affect RNA synthesis. On the other hand, the mutation of the NC motif had only a moderate effect on negative-strand synthesis and mainly impaired the function of 2A required for positive-strand synthesis.

FIG. 4.

Negative- and positive-strand synthesis of AV-FL-Luc-5′rzm RNA and its mutant RNAs (A) and AV-FL-Luc-mut9 RNA and its mutant RNAs (B) in the cell-free translation/replication system. The positions of replicative form (RF) and single-strand RNA (ssRNA) are indicated. In panel B, the signal intensities of products are quantitated and expressed as percentages of the product in the wild type.

Of the picornavirus 2A proteins related to H-rev107, parechovirus 2A has been studied as to its biochemical properties and its intracellular localization in infected cells (10), but direct evidence of the importance of 2A in virus replication has not been reported. This study showed that AiV 2A is essential for virus replication. AiV 2A was involved in both negative- and positive-strand synthesis, and the two mutations examined affected positive-strand synthesis more severely than negative-strand synthesis. Parechovirus 2A has been reported to interact with the 3′ UTR of the genome (10). There have been studies showing that the picornavirus 3′ UTR is involved in negative-strand synthesis (16) and positive-strand synthesis (1). It is possible that AiV 2A plays roles in negative- and positive-strand synthesis through interaction with the 3′ UTR. To understand the roles of AiV 2A in negative- and positive-strand synthesis, it will be necessary to study the interaction not only with viral RNA but also with viral and cellular proteins. The NC-AA mutant obtained in this study may be useful for such analyses because of its positive strand-specific synthesis defect.