Unpaired 5' ppp-nucleotides, as found in arenavirus double-stranded RNA panhandles, are not recognized by RIG-I

Abstract

Arenavirus and bunyavirus RNA genomes are unusual in that they are found in circular nucleocapsids, presumably due to the annealing of their complementary terminal sequences. Moreover, arenavirus genome synthesis initiates with GTP at position +2 of the template rather than at the precise 3' end (position +1). After formation of a dinucleotide, 5' pppGpC(OH) is then realigned on the template before this primer is extended. The net result of this "prime and realign" mechanism of genome initiation is that 5' pppG is found as an unpaired 5' nucleotide when the complementary genome ends anneal to form a double-stranded (dsRNA) panhandle. Using 5' pppRNA made in vitro and purified so that all dsRNA side products are absent, we have determined that both this 5' nucleotide overhang, as well as mismatches within the dsRNA (as found in some arenavirus genomes), clearly reduce the ability of these model dsRNAs to induce interferon upon transfection into cells. The presence of this unpaired 5' ppp-nucleotide is thus another way that some viruses appear to use to avoid detection by cytoplasmic pattern recognition receptors.

FIGURE 1.

Prime and realign mechanism of initiating arenavirus genome synthesis. Antigenome synthesis is initiated by GTP opposite the template C at position +2 at the 3′ end of the genome (written 3′ to 5′; top level). After its extension to a dinucleotide, 5′ pppGpC_OH is realigned such that its 3′ terminal C_OH is opposite the genome 3′ terminal G (middle level), and the realigned 5′ pppGpC_OH then acts a primer for antigenome synthesis (dashed line; bottom level). The “pseudo-templated” 5′ pppG is somehow not copied during genome replication, as it is specifically cleaved from genome/antigenome dsRNA by single strand-specific RNase T1. An identical process occurs during genome synthesis.

FIGURE 2.

A, RNase III and RNase A sensitivity of 5′ ppp-RNA1 before and after PAGE purification. Left panel, [³²P]CTP-labeled RNA1 was untreated, digested with RNase III, or digested with RNase A in 0.4 m NaCl, as indicated. After phenol extraction and ethanol precipitation, the samples were analyzed on a 15% sequencing gel. The right-hand lane shows RNA length markers. Right panel, [³²P]CTP-labeled RNA1 was electrophoresed on a preparative denaturing 15% polyacrylamide gel and the 54-nt long band was recovered. PAGE-purified RNA1 (5′ ppp-RNA1*) was then digested (or not) with RNase III or RNase A. The sequence of RNA1 is shown above. B, kinetics of synthesis of [³²P]UTP- or [³²P]CTP-labeled RNA1. RNA1 was synthesized with T7 RNA polymerase directed by promoter-containing DNA, and the products were labeled with either [³²P]CTP or [³²P]UTP, as indicated. Samples were taken at various times and the products were examined by denaturing PAGE. A [³²P]UTP-labeled reaction that was primed with PAGE-purified pppRNA1 in place of promoter-containing DNA was also carried out (RNA template). An underexposure of the 54-nt region of the gel is shown at the top. The sequence of RNA1 is shown above.

FIGURE 3.

The effect of E3L^100–190 expression on 5′ pppRNA1-induced IFNβ activation. Parallel cultures of mouse embryo fibroblast were transfected with pIFNβ-(ff)luciferase and pTK-(ren)luciferase, plus and minus pEBS-E3L^100–190. Twenty hours later, the cultures were further transfected with either unpurified or PAGE-purified 5′ pppRNA1. Cell extracts were prepared at 24 hours post transfection and their luciferase activities were determined. Tom refers to tomato, a red fluorescent protein, used as a control.

FIGURE 4.

RNase III treatment of PAGE-purified 5′ pppRNA1. A, PAGE-purified 5′ pppRNA1 and poly(I-C) were either mock-treated or digested with RNase III, as indicated. Samples were then separated by nondenaturing PAGE and the gel was stained with ethidium bromide. DNA restriction fragments were electrophoresed in the left-hand lane as size markers. B, samples were also transfected into A549 cells and their ability to activate the IFNβ promoter was determined.

FIGURE 5.

The importance of including the 5′ pppN within dsRNA. A, doubly purified 5′ ppp-ssRNA1** (54′mer) was annealed (or not) with 18′-mers complementary to positions 1–18, 2–19, and 4–21 (relative to the 5′ ppp-end), as indicated on the left. Samples were then transfected into A549 cells and their ability to activate the IFNβ promoter was determined. Transfection of the 18′-mers by themselves were inactive (supplemental Fig. S4). Line 1 (Ctl) shows the level of IFNβ activation in untransfected cells. B, doubly purified 5′ ppp-ssRNA1** (54′-mer) was annealed with an 18′-mer complementary to positions 1–18, as indicated on the left. Samples were then transfected into either wild-type A549 cells or A549 cells in which RIG-I was knocked down (KD) via miRNA expression (27), and their ability to activate the IFNβ promoter was determined. Ctl shows the level of IFNβ activation in untransfected cells.

FIGURE 6.

The effect of a 3-nt bulge plus the presence (or not) of the 5′ pppN within dsRNA of model RNA1 on PAMP function. 5′ ppp-ssRNA1 (54′-mer) was annealed (or not, line 2) with 20′-mer complementary to positions 1–20 or 2–21, which did or did not contain 2 mismatches at positions 5 and 7, 6 and 8, 7 and 9, or 8 and 10, as indicated on the left. 200 ng (×1) or 1 μg of 54′-mer (×5) were then transfected into A549 cells, and their ability to activate the IFNβ promoter was determined. 3 μg of total RNA was transfected in all cases, the difference being made up with tRNA (transfection of tRNA by itself was neutral). Line 1 (Ctl) shows the level of IFNβ activation in untransfected cells. Transfection of the 20′-mers by themselves were inactive (see supplemental Fig. S4).

FIGURE 7.

The effect of a 3-nt bulge plus the presence (or not) of the 5′ pppN within dsRNA of modified Junin virus RNA on PAMP function. A, the sequence of the 5′ 61 nucleotides of the Junin virus S genome RNA is shown (5′ to 3′, top line), in which capital letters indicate uridines that were changed to the nucleotides shown to generate Jun −1/60^mod RNA. The bottom line shows the sequence of the 3′ terminal 20 nucleotides of the S genome RNA (3′ to 5′). The mismatches in the dsRNA panhandle structure are indicated. B, 5′ ppp Jun −1/60^mod RNA, RNase III treated (light gray) or not (dark gray), was annealed (or not, lane 7) with synthetic oligonucleotides representing positions −1 or +1 to 20 of the S genome panhandle as complementary RNA, maintaining the 3-nt bulge (lines 8 and 9, respectively), or oligonucleotides in which positions 5 and 7 were changed so that a perfect dsRNA panhandle would be formed to resemble to L segment (lines 10 and 11, respectively). The hybridized RNA (see supplemental Fig. S5B) were then transfected into A549 cells, and their ability to activate the IFNβ promoter was determined. 1 μg of 5′ ppp Jun −1/60^mod RNA was transfected in all cases. Lane 1 (Ctl) shows the level of IFNβ activation in untransfected cells. Transfection of tRNA by itself is shown in lane 6. Transfection of the oligonucleotides by themselves is shown in lanes 2–5.