Identification of two functionally redundant RNA elements in the coding sequence of poliovirus using computer-generated design

Abstract

Genomes of RNA viruses contain multiple functional RNA elements required for translation or RNA replication. We use unique approaches to identify functional RNA elements in the coding sequence of poliovirus (PV), a plus strand RNA virus. The general method is to recode large segments of the genome using synonymous codons, such that protein sequences, codon use, and codon pair bias are conserved but the nucleic acid sequence is changed. Such recoding does not affect the growth of PV unless it destroys the sequence/structure of a functional RNA element. Using genetic analyses and a method called "signal location search," we detected two unique functionally redundant RNA elements (α and β), each about 75 nt long and separated by 150 nt, in the 3'-terminal coding sequence of RNA polymerase, 3D(pol). The presence of wild type (WT) α or β was sufficient for the optimal growth of PV, but the alteration of both segments in the same virus yielded very low titers and tiny plaques. The nucleotide sequences and predicted RNA structures of α and β have no apparent resemblance to each other. In α, we narrowed down the functional domain to a 48-nt-long, highly conserved segment. The primary determinant of function in β is a stable and highly conserved hairpin. Reporter constructs showed that the α- and β-segments are required for RNA replication. Recoding offers a unique and effective method to search for unknown functional RNA elements in coding sequences of RNA viruses, particularly if the signals are redundant in function.

Fig. 1.

Growth properties of polioviruses containing SD in the P1, P2, and P3 domains of the polyprotein. (A) Genomic structure of PV. The genome consists of a long 5′-NTR, a single ORF (polyprotein), a short 3′-NTR, and a poly(A) tail. The polyprotein contains three domains, one structural (P1) and two nonstructural (P2 and P3). The mature proteins of the P3 domain are indicated. The cre structure in P2 is absolutely required for replication. The location of the Bgl II restriction site used for subcloning is shown. AA, poly(A) tail; IRES, internal ribosomal entry site. (B) SD of the 3CD^pro domain. The 3CD^pro coding sequence, lacking 163 nt at the 5′-terminus of 3C, was divided into two parts using an Afl II restriction site: Δ3C + 152 nt of 3D ([Δ3C/5′-3D]^SD) and the remainder of 3D (Δ3D_a). (C) Growth phenotypes of SD polioviruses. Viruses containing chemically synthesized SD segments were characterized by determining the time of CPE, virus titer, and plaque size (Materials and Methods). The nucleotide numbers shown in parentheses in the “constructs” column mean the SD region. The term “0” passage means full CPE after transfection of transcripts into cells. (D) Growth phenotypes of viruses containing the SD segments within the 3CD^pro domain are characterized.

Fig. 2.

Signal location search using four SDs. Most of the 3D^pol-coding sequence (6,140–7,369 nt) was divided into 16 segments with W and S sequences alternating in different combinations. The first 10 segments were 78 nt in length, and the remaining 6 segments were 75 nt in length. Four such designs were made, and the segments were chemically synthesized. The growth phenotypes of the viruses derived from these four constructs were determined as described in Materials and Methods.

Fig. 3.

Identification of two functionally redundant elements, α and β, in the 3′-terminal 450-nt-long segment of the 3D^pol coding sequence. The 3′-terminal 450 nt of the 3D^pol coding sequence (6,920–7,369 nt) was subjected to SD changes as shown in rows 1–6. The growth phenotypes of the viruses derived from these constructs were characterized as described in Materials and Methods. Above the figure, the positions of α and β are shown within the 450-nt fragment.

Fig. 4.

Comparison of RNA replication of Fluc reporter replicons containing WT or SD sequences. (A) Structures of three reporter replicons containing WT or SD sequences. The linker contains an N-terminal 3CD^pro protease cleavage site and a C-terminal 2A^pro protease cleavage site. (B) To determine the level of RNA replication, RNA transcripts were transfected into HeLa R19 cells both in the absence and presence of 2 mM GnHCl. The luciferase activity in the absence and presence of GnHCl was measured. The luciferase data are the average of two independent experiments.

Fig. 5.

Sequence comparison of α-segment in C-cluster enteroviruses and structures of PV α-segment. (A) Nucleotide sequence alignment of α-segment in C-cluster enteroviruses. A relatively highly conserved 48-nt-long sequence is indicated. Predicted RNA structures of the WT α-segment (B) and SD α-segment (C) of PV are shown. Mutated nucleotides are shown with lowercase letters and marked with dots.

Fig. 6.

Mutational analysis of the conserved hairpin in β-segment. MFold was used for the determination of predicted RNA structures. (A) Structure of the WT β-segment. (B) Structure of the SD β-segment. (C) Structure of mut1-β, in which the nucleotide sequence of the β-hairpin was altered. (D) Structure of SD-β37 in which both the sequence and structure of the β-hairpin were changed. (E) Growth phenotypes of viruses derived from the different constructs shown in A to D. (F) Growth phenotypes of viruses in which only the 48-nt conserved sequence of α is scrambled along with fully scrambled β (SD[α48 + β75]) or with an SD β-hairpin (SD[α48 + β37]).