A Predicted Secondary Structural Domain within the Internal Ribosome Entry Site of Echovirus 12 Mediates a Cell-Type-Specific Block to Viral Replication

Abstract

The enterovirus 5′ nontranslated region (NTR) contains an internal ribosome entry site (IRES), which facilitates translation initiation of the viral open reading frame in a 5′ (m⁷GpppN) cap-independent manner, and cis-acting signals for positive-strand RNA replication. For several enteroviruses, the 5′ NTR has been shown to determine the virulence phenotype. We have constructed a chimera consisting of the putative IRES element from the Travis strain of echovirus 12 (ECV12), a wild-type, relatively nonvirulent human enterovirus, exchanged with the homologous region of a full-length infectious clone of coxsackievirus B3 (CBV3). The resulting chimera, known as ECV12(5′NTR)CBV3, replicates similarly to CBV3 in human and simian cell lines yet, unlike CBV3, is completely restricted for growth on two primary murine cell lines at 37°C. By utilizing a reverse-genetics approach, the growth restriction phenotype was localized to the predicted stem-loop II within the IRES of ECV12. In addition, a revertant of ECV12(5′NTR)CBV3 was isolated which possessed three transition mutations and had restored capability for replication in the utilized murine cell lines. Assays for cardiovirulence indicated that the ECV12 IRES is responsible for a noncardiovirulent phenotype in a murine model for acute myocarditis. The results indicate that the 5′ NTRs of ECV12 and CBV3 exhibit variable intracellular requirements for function and serve as secondary determinants of tissue or species tropism.

The Enterovirus genus within the family Picornaviridae contains human pathogens responsible for clinical syndromes involving multiple tissues and organ systems (reviewed in reference 37). The enteroviruses, like all members of the family Picornaviridae, are positive-stranded, nonenveloped RNA viruses which possess a characteristic icosahedral capsid. This genus is subdivided into five groups: the polioviruses, coxsackievirus groups A and B, echoviruses, and the numbered enteroviruses. The ∼7.4-kb enterovirus genome is structurally similar to eukaryotic mRNA in several respects: a single open reading frame encoding structural and nonstructural proteins is flanked by 5′ and 3′ nontranslated regions (NTRs) and a 3′ polyadenylated tail. The 5′ NTR is highly structured, containing six predicted secondary structural domains, or stem-loops (SLs) (44, 52), and carries out functions crucial to the life cycle of the virus (2, 42). In addition, the 5′ NTR has been shown to influence the virulence phenotypes of several enteroviruses (16, 20, 32, 41).

SL I, also known as the cloverleaf, within the poliovirus (PV) 5′ NTR is a cis element essential for stimulation of positive-sense RNA transcription (2, 3). A discontinuous region encompassing SLII through SLVI constitutes a complex element known as the internal ribosome entry site (IRES) (42). The IRES stimulates translation initiation of the open reading frame by a mechanism that is unlike that utilized by the majority of cellular mRNAs. Specifically, the enterovirus genome lacks a 5′ (m⁷GpppN) cap structure (26, 40) which, through its interaction with the cap-binding complex (eIF-4F), recruits the 40S ribosomal subunit to the 5′ end of the mRNA. A salient feature of the enterovirus IRES is the functional requirement for interaction with noncanonical initiation factors. For example, poly(rC) binding protein 2 (PCBP2) interacts with the cloverleaf and SLIV of the PV 5′ NTR (8, 9, 18). Depletion of PCBP2 from cellular extracts utilized for in vitro translation analysis significantly restricts the activity of the PV IRES (18, 57).

The primary determinant of tropism for PV, the prototypic picornavirus, is the availability of a cell surface receptor. Early studies conducted by Holland indicated a strong correlation between the ability of various simian tissues to bind PV and their ability to support PV replication (28). However, certain tissues capable of binding PV were found to be resistant to infection, suggesting that factors other than receptor availability contribute to PV tissue tropism.

Primates are the only naturally susceptible hosts for the PVs. However, mice can be made susceptible to infection by the wild-type PVs through the construction of transgenic strains expressing the gene for human PV receptor (PVR) (38, 47). This indicates that receptor availability is the primary determinant for the PV host range. While PVR transgenic mice develop paralytic disease in response to inoculation with PV1 Mahoney (PV1M), replication was found to be restricted to a specific subset of tissues expressing PVR (46). In addition, targeted expression of the PVR transgene within intestinal epithelial cells of mice failed to confer PV susceptibility to the cells (61). Taken together, these observations further suggest that PV tropism is not determined solely by receptor availability.

Experimental data derived from the characterization of engineered PV mutants and chimeras between PV and other picornaviruses has provided evidence that the 5′ NTR can affect host range and cell-type-specific replication in addition to virulence. Mutations within SLII of the PV1M 5′ NTR were found to attenuate virus replication and virulence specifically in cultured murine PVR-expressing cell lines and PVR mice, respectively (51). Molecular characterization of the SLII mutant viruses indicated a defect in translation initiation in a murine context. A chimera containing the IRES of human rhinovirus 2 in the background of the PV1M genome [PV1(RIPO)] exhibited decreased replication kinetics on a neuroblastoma cell line (SK-N-MC) compared to parental PV1M and was attenuated for neurovirulence in PVR mice and cynomolgus monkeys (20). These and other findings (34, 54) indicate that the enterovirus 5′ NTR may influence cell-type-specific replication as a function of IRES efficiency within different cell types.

We present here the construction and characterization of a chimeric enterovirus between the Travis strain of echovirus 12 (ECV12), an enterovirus rarely associated with severe human disease (53), and the 0 (zero) laboratory strain of coxsackievirus B3 [CBV3(0)] (12). This chimera contains the putative IRES element of ECV12, SLII through the true initiation codon, within the full-length background of the CBV3 genome. The ECV12(5′NTR)CBV3 chimera exhibited unique in vitro replication properties relative to the parental CBV3 strain and was attenuated for virulence in a murine model for CBV3 disease (56). Both CBV3 and the chimera replicated efficiently on human and simian cell lines. However, unlike CBV3, ECV12(5′NTR)CBV3 was completely restricted for growth on two primary murine cell lines: murine fetal heart fibroblasts (MFHF) and BNL CL.2 embryonic liver cells. The naturally-occurring ECV12 SLII was found to be responsible for the MFHF-BNL CL.2 replication-restricted phenotype. In addition, specific mutations within the ECV12 5′ NTR were found to restore viability on both MFHF and BNL CL.2 cells. These observations indicate that an element within a naturally occurring enterovirus 5′ NTR can influence tissue or species tropism in a receptor-independent manner.

MATERIALS AND METHODS

Cells and viruses.

HeLa, LL-CMK2 (monkey kidney), BNL CL.2 (mouse liver) (American Type Culture Collection, Manassas, Va.), and MFHF (59) cells were propagated at 37°C as monolayers in minimal essential medium (MEM) supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 25.5 mM sodium bicarbonate, and 50 μg of gentamicin/ml. The complete sequences of the wild-type ECV12 Travis strain (American Type Culture Collection) and the CBV3(0) laboratory strain have been reported (12, 33). The full-length infectious clone of CBV3(0) has been well characterized and produces a noncardiovirulent virus for which the virulence determinant (nucleotide 234) has been mapped (56). Stocks of CBV3 and recombinant viruses were propagated on HeLa cells. ECV12 (Travis strain) was propagated using LL-CMK2 cells, and titers were determined. For the determination of the viral titer, serial dilutions (10⁻¹ through 10⁻⁸) of virus were plated on HeLa cells in 96-well plates (10³ cells/well), incubated for 5 days at 37°C, and examined for cytopathic effect. Titers were calculated (13) and expressed as 50% tissue culture infective doses (TCID₅₀) per milliliter.

Construction and generation of ECV12-CBV3 chimeras.

The following primers were utilized for reverse transcription (RT)-PCR, overlap fusion PCR mutagenesis, and/or sequencing of chimeric viruses: 1 (JRp64), ACGGTACCTTTGTGCGCCTGTTTTA; 2 (MD90), ATTGTCACCATAAGCAGCCA; 3 (MD91), CCTCCGGCCCCTGAATGCGGCTAAT; 4 (DREV), GCGTTGATACTTGAGCTCCC; 5 (ECV12-1+), CCTCCCCCAACTGTAACCTAGAAGTTCATCAC; 6 (ECV12-1−), GTTACAGTTGGGGGAGGG; 7 (CBV30-int+), CTAGAAGTAACACACACCTGATCAATAGTTAGCTCAAC; 8 (CBV30-int−), GATCAGGTGTGTGTTACTTCTAGGTTACTGAGGG; 9 (CBV30A+), GTGGCACACCAGCCACGTTTTGATCAAGCACTTCTGTTA; 10 (CBV30A−), CGTGGCTGGTGTGCCACGCTGACTGTTGATCATTTGTGATGAAC; 11 (SLII+), GCACTTCTGTTACCCCGG; 12 (SLII−), CCGGGGTAACAGAAGTGC; 13 (12/0−468), GGATTAGCCGCATTCAGGGGCCGGA; 14 (12/0+446), CCTCCGGCCCCTGAATGCGGCTAAT; and 15 (ATG), ACTTGAGCTCCCATT (49). Primer 1 contains a KpnI site (nucleotide [nt] 69 relative to ECV12 numbering), whereas primers 4 and 15 possess sites for SacI (nt 750).

To generate the full-length ECV12(5′NTR)CBV3 clone, an ECV12 5′ NTR subclone was initially constructed from two different RT-PCR amplifications of ECV12 using primer pairs 1-2 and 3-15. This yielded DNA fragments containing ECV12 sequences from nt 63 to 599 (primers 1 and 2) and 449 to 764 (primers 3 and 15). Both PCR fragments were blunt-end ligated into the SrfI-linearized pCRscript vector (Stratagene, San Diego, Calif.). Subsequently, an ECV12 downstream 5′ NTR clone (nt 460 to 764) in the reverse orientation was digested with SacI. The resulting linearized plasmid was blunted with T4 polymerase and then cut with BsmI (nt 464). This fragment was then inserted into an upstream ECV12 5′ NTR clone (nt 63 to 599) linearized with BsmI and EcoRV (a blunt cutter within the vector backbone). This created the ECV12 5′ NTR subclone. The full-length infectious clone of ECV12(5′NTR)CBV3 was constructed through insertion of the KpnI-SacI fragment from the ECV12 5′ NTR subclone into a vector prepared from the cDNA clone of CBV3(0).

All of the following chimeric 5′ NTR PCR fragments were cloned into the full-length CBV3(0) cDNA using KpnI (nt 69) and SacI (nt 750). ECV12(II-IV)CBV3 was constructed through amplification of the ECV12 upstream (nt 63 to 472) and CBV3 downstream (nt 446 to 764) regions using the primer pairs 1-13 and 4-14, respectively. The resulting PCR products were used as templates for overlap fusion PCR (27) with primers 1 and 4. ECV12(V-AUG)CBV3 was generated using the same primers, but the upstream and downstream PCR products were amplified from CBV3 and ECV12, respectively. ECV12(SLII)CBV3 and CBV3(SLII)ECV12 were constructed by amplification of SLII (nt 63 to 186) of ECV12 or CBV3 using primers 1 and 12. The resulting PCR product was fused to a downstream fragment of ECV12 or CBV3 (nt 169 to 754), amplified with primers 4 and 11, in an overlap fusion PCR using the primer pair 1-4. CBV3-link-ECV12 was generated by fusion (using primers 1 and 4) of PCR products amplified with primer pairs 1-6 (nt 63 to 111) and 5-4 (nt 93 to 764) using ECV12 as a template. CBV3-int-ECV12 was constructed by amplification of ECV12 using primer pairs 1-8 (nt 63 to 130) and 7-4 (nt 108 to 764), with subsequent fusion with primers 1 and 4. Lastly, CBV3-apical-ECV12 was constructed by fusion of products amplified with primers 1-10 (nt 63 to 158) and 4-9 (nt 142 to 764) from ECV12 using the primer pair 1-4. Overlap PCR products were inserted into the pCRscript vector.

Viral-replication assays.

Monolayers of 10⁵ HeLa, MFHF, LL-CMK2, or BNL CL.2 cells in 35-mm² plates were washed once with phosphate-buffered saline (PBS) and adsorbed with 2 × 10⁶ TCID₅₀ in a volume of 0.5 ml for 30 min at room temperature. Subsequently, the plates were washed twice with PBS, refed with 2 ml of supplemented MEM, and incubated at 37.0°C (unless otherwise indicated) in 5% CO₂. All experiments were performed in duplicate. At various time points, two plates were removed and subjected to three freeze-thaw cycles. The titers of virus were determined using a standard infectivity assay as described above.

Selection of ECV12(5′NTR)CBV3 revertants.

ECV12(5′NTR)CBV3 (2 × 10⁶ TCID₅₀) was incubated on 10⁵ PBS-washed MFHF cells in a 35-mm² plate for 30 min at room temperature. The cells were then washed twice with PBS, refed with 2 ml of supplemented MEM, and incubated at 33.5°C in 5% CO₂ for 8 h to allow limited viral replication. Virus-infected MFHF cells were subsequently shifted to 37.0°C and incubated for 72 h. The plates were then freeze-thawed three times. The medium was removed from the plates and clarified by centrifugation. Two hundred microliters (1/10 volume) of the clarified supernatant was passaged onto 10⁵ fresh MFHF cells and incubated at 37.0°C for 72 h. The virus was blind passaged in this manner a total of five times. Virus from the fifth passage was subjected to RT-PCR amplification of the 5′ NTR, using primer pair 1-4. The amplicon was directly sequenced as described below. The RT-PCR product, encompassing nt 64 through the true AUG, was then cloned into the pCRscript vector. An individual clone was selected and sequenced to verify fidelity to the direct RT-PCR product sequence. The revertant ECV12 5′ NTR was subsequently cloned into the full-length background of the genomic CBV3(0) cDNA using the KpnI and SacI sites.

Viral RNA extraction, RT-PCR, and sequencing.

One hundred microliters (each) of viral stocks were extracted with guanidium isothiocyanate (ORCA Research Inc., Bothel, Wash.), and the resulting viral RNA was ethanol precipitated, washed, and dried. RT-PCR was performed as previously described (16, 49) using primers 1 and 4 (see above) to amplify the 5′ NTR. The RT-PCR products were subjected to 1% agarose gel electrophoresis in 1× Tris-acetate-EDTA buffer, stained with ethidium bromide (0.5 μg/ml), and purified by gel extraction with a gel extraction kit (Qiagen, Valencia, Calif.). The purified products were then directly sequenced with primers 1, 2, 3, and 4 using a terminator cycle-sequencing kit (Thermo Sequenase; Amersham, Cleveland, Ohio). Sequencing reaction mixtures were electrophoresed through an 8% acrylamide–8 M urea–1× Tris-taurine-EDTA gel for 1.8 h at a constant power of 80 W.

Assays for cardiovirulence.

Determination of the cardiovirulence phenotype was performed as described previously (16). Briefly, 3- to 4-week-old male C3H/HeJ mice were injected intraperitoneally with 2 × 10⁵ TCID₅₀ of virus [ECV12(5′NTR)CBV3 or CBV3(20)] in 0.1 ml of unsupplemented medium or with medium alone. Groups of five mice were maintained in separate microisolators for 10 days, at which point the animals were sacrificed and the hearts were excised. Portions of the hearts were subsequently fixed, sectioned, and stained with hemotoxylin and eosin. Subsequently, myocardial sections were examined for histopathological evidence of acute myocarditis. The remaining heart tissue was utilized for determination of the cardiac viral titer as described above.

Computational analysis.

Sequence analysis was performed with the MacVector 6.0 software package (Oxford Molecular, Oxford, United Kingdom). The m-fold algorithm (version 3.0) (35, 62) was used to generate predicted RNA secondary structures. RNA sequences were folded at 37.0° with 5% suboptimality.

RESULTS

ECV12(5′NTR)CBV3 in vitro replication kinetics.

The full-length cDNA clone of the chimeric ECV12(5′NTR)CBV3 (Fig. 1b) was constructed, and infectious virus was obtained as detailed in Materials and Methods. Direct sequencing of the RT-PCR-amplified ECV12 5′ NTR after serial HeLa passage revealed that no mutations had occurred in the overall viral quasispecies, indicating that the chimera was stable. ECV12 and CBV3 exhibit an overall sequence identity of 77% within the 5′ NTR and have similar predicted secondary structures consisting of six structural domains (44, 52). Wild-type ECV12 is incapable of replication on HeLa cells, presumably due to lack of the putative ECV12 cell surface receptor, decay-accelerating factor (DAF) (45). However, the recombinant virus was capable of replication to high titer on HeLa cells, indicating that the ECV12 IRES is functional in this cell line.

FIG. 1.

Linear representation of ECV12-CBV3 5′ NTR chimeras with indicated SL exchanges. CBV3 and ECV12 SLs are open and hatched, respectively. All chimeras contain the CBV3 cloverleaf domain. (a) Parental CBV3 5′ NTR; (b) ECV12(5′NTR)CBV3; (c) ECV12(II-IV)CBV3; (d) ECV12(V-AUG)CBV3; (e) ECV12(SLII)CBV3; (f) CBV3(SLII)ECV12.

The in vitro replication kinetics of the ECV12(5′NTR)CBV3 chimera and the parental CBV3 strain were evaluated on three distinct cell lines: human cervical carcinoma cells (HeLa), simian kidney cells (LL-CMK2), and MFHF cells. Both viruses replicated with similar kinetics on HeLa (Fig. 2a) and LL-CMK2 (Fig. 2b) cells. However, both CBV3 and ECV12(5′NTR)CBV3 exhibited reduced growth kinetics on LL-CMK2 cells compared to those on HeLa cells. The parental CBV3 was capable of growth on MFHF cells (Fig. 2c), albeit with less efficiency than on HeLa or LL-CMK2 cells. Most interestingly, the ECV12(5′NTR)CBV3 chimera showed a complete restriction for replication on MFHF cells. These results suggested that the ECV12 5′ NTR sequences and/or structures, constituting the putative IRES element, were not functional within the MFHF intracellular milieu.

FIG. 2.

Growth curves of selected ECV12-CBV3 chimeras on HeLa (a), LL-CMK2 (b), and MFHF (c) cells at 37°C. HeLa single-cycle growth kinetics were determined for CBV3 (■) and the MFHF growth-restricted chimeras ECV12(5′NTR)CBV3 (⧫), ECV12(II-IV)CBV3 (▴), ECV12(SLII)CBV3 (×), CBV3-int-ECV12 (∗), CBV3-link-ECV12 (●), and CBV3-apical-ECV12 (+). LL-CMK2 single-cycle growth kinetics were assayed for CBV3 and ECV12(5′NTR)CBV3. Extended time points were utilized for analysis of MFHF growth kinetics with CBV3 and ECV12(5′NTR)CBV3. All single-cycle growth experiments were performed in duplicate. p.i., postinfection.

Localization of the ECV12 MFHF growth restriction determinant.

We utilized a reverse-genetics approach to localize the portion of the ECV12 5′ NTR responsible for the MFHF growth restriction phenotype. As a means to this end, two intra-5′ NTR chimeric viruses between ECV12 and CBV3 were constructed to distinguish between potential MFHF growth restrictive effects of the upstream and downstream portions of the ECV12 5′ NTR. One virus, ECV12(II-IV)CBV3, contained the predicted SLs II, III, and IV from ECV12 (nt 88 to 445 relative to ECV12) in the background of CBV3 (Fig. 1c). The second recombinant virus, ECV12(V-AUG)CBV3, was the converse of the first and contained the downstream SLs V and VI through the correct start codon from ECV12 (nt 446 to 744) exchanged with the homologous sequences from CBV3 (Fig. 1d). Similar to the ECV12(5′NTR)CBV3 chimera, both of these viruses yielded high-titer stocks on HeLa cells and were devoid of 5′ NTR mutations after multiple HeLa cell passages, as determined by RT-PCR and sequence analysis.

The ability of these chimeras to replicate on MFHF cells was evaluated using 72-h growth assays. As can be seen in Fig. 3, the chimera containing the upstream IRES domains from ECV12, ECV12(II-IV)CBV3, exhibited a complete block for replication on MFHF cells. The converse chimera, ECV12(V-AUG)CBV3, yielded an increase in viral titer of approximately 3.5 log TCID₅₀/ml over 72 h. These findings implied that the ECV12 5′ NTR sequences responsible for the MFHF growth-restricted phenotype were located within SLs II to IV.

FIG. 3.

Seventy-two-hour MFHF growth experiments at 37°C. Starting titers are shaded, and 72-hour titers are solid. Plates representing starting time points were harvested immediately after the monolayers were washed to remove excess virus. Monolayers of MFHF cells were infected with the indicated viruses as described in Materials and Methods. All experiments were performed in duplicate. Chimeras are abbreviated as follows: ECV12(5′NTR)CBV3, 12-NTR-0; ECV12(II-IV)CBV3, 12(II-IV); ECV12(V-AUG)CBV3, 12(V-AUG); ECV12(SLII)CBV3, SLII(12); CBV3(SLII)ECV12, SLII(0); CBV3-link-ECV12, 0-link; CBV3-int-ECV12, 0-int; CBV3-apical-ECV12, 0-apical; and ECV12(5′NTR)CBV3-REV, 12-0R.

Analysis of sequence identity within the upstream domains of ECV12 and CBV3 revealed significant variation within SLs II and III, with relative conservation of SLIV (77, 80, and 87% sequence identity, respectively). Because SLIII has been shown to be dispensable for IRES function in the PV paradigm (23, 39), we elected to characterize the possible role of SLII in mediating the MFHF growth restriction phenotype. Two chimeras were constructed. The first, named ECV12(SLII)CBV3 (Fig. 1e), resulted from exchange of the CBV3 SLII, including the linker sequence between the cloverleaf and SLII, with that of ECV12 (nt 88 to 181). The second chimera contained the CBV3 SLII within the context of the ECV12 5′ NTR fused to the CBV3 coding region and was termed CBV3(SLII)ECV12 (Fig. 1f). Both chimeras were capable of replication to high titer on HeLa cells and were free of mutation after HeLa cell passage. Seventy-two-hour growth experiments with these chimeras on MFHF cells revealed a growth-restricted phenotype for ECV12(SLII)CBV3 (Fig. 3). In contrast, the converse chimera, CBV3(SLII)ECV12, possessed a restored ability to replicate on MFHF cells. These results indicated that the ECV12 sequences that make up the predicted SLII were responsible for the MFHF growth-restricted phenotype. In addition, these findings suggested that the ECV12 SLII is incapable of performing a vital intracellular replicative function within MFHF cells.

In order to rule out possible subtle effects of domain exchanges on virus replication efficiency, we conducted single-cycle growth curves with all MFHF-restricted chimeras on HeLa cells (Fig. 2a). All viruses replicated with similar kinetics on HeLa cells, although the ECV12(SLII)CBV3 chimera displayed a ∼0.8 log TCID₅₀/ml reduction in titer at 8 h postinfection compared to other viruses. Thus, a gross deficiency in replication on HeLa cells did not correlate with the MFHF growth restriction phenotype.

The m-fold (35, 62) predicted secondary structures of the ECV12 and CBV3 SLIIs were found to be nearly identical despite significant sequence diversity (Fig. 4A and B). Sequence alignment revealed that the ECV12 SLI-SLII linker region is two residues shorter than that of CBV3, and four nucleotide differences are present between the two viruses. The predicted structures of the lower stems were identical, and a single nucleotide difference between the viruses is present within the two-residue 5′ bulge loop. A large internal loop located atop the lower stem was predicted for both viruses. Seven nucleotide differences between ECV12 and CBV3 are present in the 5′ side of this loop. The m-fold algorithm predicted slight differences in the structure of the upper stem loop. Whereas this region of CBV3 contained a single 5′ bulge loop of 5 nt, ECV12 possessed two smaller bulge loops separated by a 3-bp stem. Overall, this region of SLII exhibits 17 nucleotide differences between the two viruses.

FIG. 4.

Predicted SLII secondary structures of CBV3 (A), ECV12 (B), CBV3-link-ECV12 (C), CBV3-int-ECV12 (D), and CBV3-apical-ECV12 (E). Secondary structures were generated using the m-fold algorithm (35, 62). Intra-SLII chimeras were constructed through exchange of homologous regions between ECV12 and CBV3 as described in Materials and Methods. CVB3 sequences are indicated in lowercase letters. CBV3-link-ECV12 resulted from the exchange of nt 88 to 102 (SLI-SLII linker) of ECV12 with nt 88 to 104 of CBV3. Exchange of ECV12 nt 116 to 125 (5′ bulge loop) with nt 118 to 127 of CBV3 yielded CBV3-int-ECV12. Lastly, replacement of ECV12 nt 126 to 165 (apical SL) with CBV3 nt 128 to 166 resulted in the CBV3-apical-ECV12 chimera.

Based on the sequences and predicted structural differences of the CBV3 and ECV12 SLIIs, additional chimeric viruses were constructed to explore the potential contributions of individual regions of the ECV12 SLII to the MFHF growth restriction phenotype. Our strategy was to exchange specific homologous regions of the ECV12 SLII with those of CBV3 in an attempt to restore viability on MFHF cells. Three viruses were engineered (Fig. 4C, D, and E) which contained (i) the SLI-SLII linker (CBV3-link-ECV12), (ii) the 5′ portion of the large internal loop (CBV3-int-ECV12), or (iii) the entire upper SL of CBV3 (CBV3-apical-ECV12) within the background of the ECV12 5′ NTR. All three intra-SLII chimeras were viable and replicated with efficiencies similar to that of CBV3 on HeLa cells (Fig. 2A). However, all three viruses failed to replicate on MFHF cells (Fig. 3). These results indicated that at least two of the selected regions, if not the entire structural domain, must be present from the MFHF-permissive CBV3 SLII in order to rescue the MFHF growth-restricted phenotype.

Selection of ECV12 revertants with restored MFHF viability.

A previously reported chimera between PV1 and CBV3 has been shown to exhibit a temperature-sensitive (ts) phenotype (50). Specifically, this chimera replicates more efficiently at temperatures slightly lower than physiological temperature. To examine whether ECV12(5′NTR)CBV3 exhibited a ts phenotype, we assayed its replication on MFHF cells at 33.5°C. At the standard temperature of 37.0°C, the chimera is completely restricted for replication at 72 h postinfection (Fig. 3). However, over the course of a 12-hour single-cycle growth experiment at 33.5°C, ECV12(5′NTR)CBV3 yielded a 1,000-fold increase in infectious virus (Fig. 5). Similarly, the other MFHF-restricted viruses described above displayed MFHF growth-permissive phenotypes at 33.5°C (data not shown). As demonstrated previously, the parental CBV3 strain utilized for this study exhibited increased replication efficiency on MFHF cells at 33.5°C (58).

FIG. 5.

Results of single-cycle replication assays with MFHF cells at 33.5°C. Growth experiments were conducted with CBV3 (■) and ECV12(5′NTR)CBV3 (⧫) as described in Materials and Methods. All experiments were conducted in duplicate. p.i., postinfection.

The ts phenotype of ECV12(5′NTR)CBV3 was exploited to generate viral quasispecies in MFHF cells from which to select revertants capable of efficient replication in MFHF. ECV12(5′NTR)CBV3 was incubated for 8 h on MFHF cells at 33.5°C and subsequently passaged five times on MFHF cells at 37.0°C. At the fifth passage, widespread virus-induced cytopathic effect was evident. RT-PCR and direct sequence analysis of the ECV12(5′NTR)CBV3 quasispecies revealed the presence of three mutations. Two occurred within SLII at positions 148 (A→G) and 160 (U→C) (numbering relative to ECV12) (Fig. 6). The U160C transition results in predicted secondary structures that differ from the wild-type ECV12 SLII. A third mutation was found in SLIV at position 426 (U→C). To rule out the possibility that unidentified mutations occurring outside of the ECV12 5′ NTR could have contributed to an MFHF replication-competent phenotype, the 5′ NTR containing the three identified transitions was engineered into the background of ECV12(5′NTR)CBV3, yielding the construct ECV12(5′NTR)CBV3-REV. Seventy-two-hour growth experiments revealed that the recombinant ECV12 mutant was capable of replication on MFHF cells (Fig. 3). In addition, the growth properties of ECV12(5′NTR)CBV3-REV and the overall revertant quasispecies were found to be similar on MFHF cells (data not shown), confirming the role of these transition mutations in the restoration of viability on MFHF cells. The specific contribution of each mutation to the MFHF growth-permissive phenotype has not been determined. However, since the reverse-genetics approach identified the importance of SLII in determining the MFHF growth phenotype, the mutations within the ECV12 SLII may be crucial for the reversion phenotype.

FIG. 6.

Predicted SLII secondary structures of wild-type ECV12 and revertant. Mutations (A148G and U160C) are indicated by arrows. Wild-type ECV12 SLII has a predicted ΔG of −20.9 kcal/mol. Two alternative predicted structures are shown for the revertant SLII: ΔGs for the top and bottom revertant structures are −19.7 and −18.9 kcal/mol, respectively.

Replication of ECV12-CBV3 chimeras on additional murine cell lines.

The cellular receptor for the CBVs, known as the human coxsackievirus and adenovirus receptor (hCAR), has been isolated and identified as a member of the immunoglobulin superfamily (6). In addition, the murine homolog of the receptor (mCAR) has been identified (7). Analysis of tissue-specific mCAR expression indicated the presence of mCAR transcript in murine heart, lung, liver, and kidney tissue. We therefore reasoned that other murine cell lines derived from these tissues might be permissive to CBV3 replication. Two murine cell lines were selected for analysis of susceptibility to CBV3: Lewis lung carcinoma (LLC1) and a normal embryonic liver cell line known as BNL CL.2. Both CBV3 and ECV12(5′NTR)CBV3 failed to replicate on LLC1 cells (data not shown), perhaps due to absence of mCAR expression. In contrast, BNL CL.2 cells were capable of supporting CBV3 replication with virus yields similar to that observed for MFHF cells at 72 h (Fig. 7). Although CBV3 was capable of replicating within BNL CL.2 cells, the chimeras ECV12(SLII)CBV3 and ECV12(5′NTR)CBV3 were completely restricted for growth. Furthermore, CBV3(SLII)ECV12 and ECV12(5′NTR)CBV3-REV, viruses that possess a recovered capacity for replication on MFHF cells, were capable of replication on BNL CL.2 cells. Thus, the ECV12 SLII restricts viral replication in both MFHF and BNL CL.2 primary murine cell lines.

FIG. 7.

Results of 72-h growth experiments using murine liver (BNL CL.2) cells. Starting time points are shaded, and 72-h time points are solid. Virus chimeras are abbreviated as follows: ECV12(5′NTR)CBV3, 12-NTR-0; CBV3(SLII)ECV12, SLII(0); ECV12(SLII)CBV3, SLII(12); and ECV12(5′NTR)CBV3-REV, 12-0R.

Cardiovirulence assays of the ECV12(5′NTR)CBV3 chimera.

Coxsackieviruses are known to be capable of infecting the heart, resulting in acute myocarditis (4). Experimental evidence has indicated that, like the PVs, the 5′ NTR of CBV3 harbors the primary cis-acting determinant for virulence (16). We sought to determine whether the 5′ NTR of the relatively nonvirulent ECV12 (Travis strain) mediated an attenuated phenotype in a well-established murine model for cardiovirulence that mimics the disease observed in humans (16, 56). Hearts from C3H/HeJ mice inoculated with the CBV3(20) positive control, a well-characterized cardiovirulent laboratory strain of CBV3 (16), revealed widespread inflammatory lesions with significant necrosis and calcification (Fig. 8). CBV3(20)-infected mice yielded 2.58 × 10⁴ TCID₅₀ per g of heart tissue. In contrast, the myocardia of mice infected with ECV12(5′NTR)CBV3 showed no signs of inflammation and appeared identical to hearts from mock-infected animals. In addition, no live virus could be detected (limit of detection, ∼60 TCID₅₀) in homogenized hearts of ECV12(5′NTR)CBV3-inoculated mice. Therefore, the ECV12 5′ NTR mediates an attenuated cardiovirulence phenotype in the murine model utilized.

FIG. 8.

Representative cardiac histology of mice infected with CBV3(20) (a), medium alone (b), or ECV12(5′NTR)CBV3 (c). Juvenile C3H/HeJ mice were inoculated with 10⁵ TCID₅₀ of ECV12 (5′NTR)CBV3 or the cardiovirulent laboratory strain CBV3(20)/ml. Negative-control mice were mock infected with unsupplemented MEM alone. At 10 days postinoculation, the murine hearts were recovered and examined for evidence of acute myocarditis as described in Materials and Methods.

DISCUSSION

Enterovirus infections have been associated with a myriad of clinical syndromes, including meningitis, encephalitis, myocarditis, and neonatal sepsis. Additionally, for reasons yet to be elucidated, several enteroviruses known to infect humans, including ECV12, are rarely associated with clinical syndromes. The primary determinant of enteroviral tropism is receptor availability. However, a growing body of experimental data suggests that the enterovirus 5′ NTR acts as a secondary, intracellular tropic determinant (20, 51). The exact mechanism(s) by which it influences tropism and disease progression remains elusive. Evidence suggests that the efficiency of IRES-directed translation initiation in specific cell types, as a function of cell-type-specific initiation factor availability or identity, may play a significant role (20, 22, 51, 54, 55).

We have demonstrated that the putative SLII from the naturally occurring, relatively avirulent ECV12 (Travis strain), within the background of CBV3, completely restricted viral replication on two primary murine cell lines: MFHF and BNL CL.2. Substitution of the entire CBV3 SLII for that of ECV12 completely restored the capacity for viral replication on these cell lines. As both of the cell lines were susceptible to infection by the parental CBV3 strain utilized as the background for our constructs, the MFHF-BNL CL.2 growth-restricted phenotype can be considered the result of viral factors independent of receptor availability.

Attempts to further localize the region responsible for the ECV12 MFHF growth phenotype were unsuccessful. Homologous exchange of selected regions within the ECV12 SLII with those from CBV3 was insufficient to reverse the MFHF growth-restricted phenotype. These findings suggest that the entire CBV3 SLII, or multiple SLII regions in combination, are required for expression of an MFHF-BNL CL.2 replication-permissive phenotype. Although CBV3 and ECV12 display 23% sequence diversity within SLII (compared to ∼14% overall 5′ NTR diversity), the exchange of homologous SLII regions between the viruses appeared to have no effect on viability or replication in HeLa cells. Tolerance of the described SLII alterations may reflect a predisposition to genetic plasticity within this region of the enterovirus genome. Indeed, 5′ NTR sequence analysis of the prototypic CBV1-5 has revealed a high level of sequence diversity within the predicted SLI-SLII linker region and SLII relative to that of the remainder of the 5′ NTR (48).

Two of the three ECV12 revertant mutations generated through multiple passages of ECV12(5′NTR)CBV3 on MFHF cells occurred, not unexpectedly, within SLII. The A148G transition is located within the predicted apical loop of SLII. Position 160 is the central residue found within the 3′ side of a predicted 3-bp stem. The U-to-C transition at position 160 is predicted to disrupt a potential canonical base pair with A134. Indeed, m-fold analysis of the mutant SLII predicted two alternative secondary structures that differ from the wild-type ECV12 SLII (Fig. 5). The third mutation was found in SLIV at nucleotide 426 (U→C) and is located in the 3′ side of the lower internal loop in SLIV. This is the only mutation that reverted to the same nucleotide occupying the homologous position in CBV3. Both SLII mutations, A148G and U160C, were found at positions for which CBV3 and ECV12 share nucleotide identity (i.e., CBV3 and ECV12 possess A148 and U160). The significance of this observation is as yet unknown. As mentioned previously, the relative contribution of each mutation to a reversion of the MFHF-BNL CL.2 growth-restricted phenotype is yet to be determined. We hypothesize, however, that the identified SLII mutations are particularly crucial given the results of our reverse-genetics experiments.

The exact mechanism by which the ECV12 SLII restricts virus replication in MFHF and BNL CL.2 cells remains to be determined. However, it is reasonable to hypothesize that the observed growth-restricted phenotype of ECV12(5′NTR)CBV3 results from a defective cell-type-specific ECV12 IRES function. Studies of the PV1M type I IRES element have indicated the necessity for specific protein-IRES interactions in order to promote accurate and efficient translation initiation. Unlike the structurally nonhomologous type II IRES elements of the Apthovirus and Cardiovirus genera, the enterovirus and rhinovirus type I IRESes are incapable of proper translation initiation in rabbit reticulocyte lysate without the addition of HeLa cell extract (10, 15). Several noncanonical translation initiation factors have been identified that interact with the PV 5′ NTR and stimulate IRES activity (reviewed in reference 5). Specifically, addition of polypyrimidine tract binding protein (PTB) (24, 29), PCBP2 (9, 18), or the La protein (1, 36) to translation-competent extracts lacking the corresponding factor results in heightened and accurate PV IRES activity. Studies have also shown that PV IRES activity requires higher levels of PCBP2 than that of type II IRES elements (57), indicating that the levels of these intracellular factors may affect cell-type-specific replication.

As has been demonstrated for PV, SLII is essential for IRES function (39). PTB and eIF-2α have been shown to interact with PV RNA encompassing the putative SLII (14, 25). Inefficient recruitment of factors essential for IRES function might explain the growth-restricted phenotype on MFHF and BNL CL.2 cells. In particular, the SLII of ECV12 may be incapable of direct interaction with specific initiation factors within MFHF and BNL CL.2 cells. Reducing the incubation temperature from 37.0 to 33.5°C restored the replication competence of MFHF growth-restricted ECV12-CBV3 chimeras. This may reflect temperature-dependent stabilization of intra-IRES and/or protein-IRES interactions that facilitate translation initiation.

Alternatively, the ECV12 SLII may have an impact on the viral growth phenotype by negatively affecting genomic RNA replication specifically within the primary murine cells utilized. Indeed, recent evidence derived from the characterization of PV mutants implicates SLII as a cis-acting signal for both RNA synthesis and translation initiation (31). Data exist supporting the involvement of as-yet-unidentified host cell factors in enterovirus RNA synthesis (reviewed in reference 60). Further study is required to determine the exact role of the enterovirus SLII in genomic RNA synthesis.

Chapman et al. recently reported the characterization of a chimera containing the entire PV1M 5′ NTR fused to the genome of a cardiovirulent CBV3 (11). The resulting virus (CPV/49) displayed a cardiovirulence-attenuated phenotype in a murine model of myocarditis. CPV/49 exhibited somewhat retarded replication kinetics on several cell lines, including MFHF cells. In addition, CPV/49 displayed decreased translational efficiency in MFHF cells. Unlike ECV12(5′NTR)CBV3, however, CPV/49 is not MFHF growth restricted, indicating that the PV 5′ NTR is functional in this murine cell type. Complete cell-type-specific growth restriction as a function of IRES identity has not been previously reported. Relative to PV and CBV3, the two enteroviruses focused upon in this study are closely related (30, 43). It was therefore unexpected to observe the dramatic in vitro replication differences between CBV3 and ECV12(5′NTR)CBV3 on the primary murine cell lines utilized.

A wealth of experimental data indicate that the PV 5 NTR is the major determinant of the virulence phenotype (17, 32, 41). Gromeier and colleagues (20, 21) have characterized the neurovirulence phenotype and in vitro growth properties of a chimera containing the human rhinovirus 2 5′ NTR within the background of PV1M [PV1(RIPO)]. This virus replicated significantly more slowly in SK-N-MC cells, a human neuroblastoma cell line, than the parental PV1M, although both viruses replicate efficiently in HeLa cells. Using a reverse-genetics approach, the authors localized specific regions of the PV SLV and -VI as the primary determinant of neurovirulence for both transgenic murine and simian models of polimyelitis (19). A similar strategy was utilized to identify the ECV12 SLII as the determinant for MFHF-BNL CL.2 growth restriction. Taken together, these experiments verify that the identities of specific SLs within the enterovirus IRES can influence cell-type-specific replication and virulence (discussed below).

For CBV3, the 5′ NTR has also been found to be a major determinant of the virulence phenotype. The cardiovirulent phenotype of a laboratory strain was attenuated when the 5′ NTR, but not the capsid coding region, was exchanged with that of a clinical, noncardiovirulent isolate (16). However, the virulent phenotype of the parental laboratory strain was maintained when the 5′ NTR was exchanged with that of a clinical cardiovirulent isolate. Interestingly, further experiments utilizing these clinical CBV3 isolates have indicated that SLII identity is important in determining the cardiovirulence phenotype (unpublished data).

Given that ECV12 is rarely associated with overt clinical disease (53) and the ECV12 5′ NTR restricts in vitro replication in two primary murine cell lines, it is not suprising that the ECV12(5′NTR)CBV3 chimera exhibited an attenuated phenotype in an in vivo murine model for CBV3-induced acute myocarditis. Wild-type ECV12, like the overwhelming majority of echoviruses, is incapable of causing disease in mice, presumably due to a lack of receptor availability. Our experiments suggest that the 5′ NTRs of some echoviruses may be secondary blocks to murine virulence.

Experimental evidence derived from the analysis of engineered PV1M SLII mutants has revealed that murine-specific defects in translation initiation are possible (51). This study provides evidence that a naturally occurring, murine-restricted correlate exists. Although ECV12(5′NTR)CBV3 is noncardiovirulent and virus could not be detected in the murine myocardium, we have not explored the possibility that the ECV12(5′NTR)CBV3 chimera is capable of replication within other tissues (pancreas, central nervous system, and liver) known to be affected by CBV3. Accordingly, the ECV12 5′ NTR may specify altered murine tissue tropism within the confines of mCAR availability. Whatever the case, it is evident that the ECV12 5′ NTR has intracellular requirements that are distinct from those of the closely related CBV3 5′ NTR.

We have demonstrated that the putative SLII within the ECV12 IRES, when present within the CBV3 genome, is responsible for altered in vitro tropism and cardiovirulence phenotype. Our findings support the concept that naturally occurring IRES elements serve as effectors of enterovirus tropism and virulence.