All plant and vertebrate viruses are parasites of the host’s translational machinery because they do not encode any of its components [1
]. Thus, a key host-pathogen interaction in establishing infection is that of viral mRNA with the host’s translation factors and the ribosome. The genomes of positive strand RNA viruses which, by definition, also serve as mRNA, do not replicate in the nucleus where capping of host mRNAs takes place. Thus, they either encode capping enzymes [2
], which is a genetic burden for most RNA viruses, or their RNA remains uncapped. For uncapped viral RNAs to translate efficiently, they harbor a structure in their untranslated region (UTR) that recruits translational machinery in the absence of a 5′ cap. In animal viruses, such a structure is either a genome-linked protein (VPg) covalently attached to the 5′ end in the case of noroviruses [3
], or more often, an internal ribosome entry site (IRES) in the viral RNA. An IRES is a highly structured RNA element of several hundred bases that commandeers host proteins called IRES trans-acting factors (ITAFs) to recruit the ribosome to the IRES, without requirement for scanning from the 5′ end [4
]. In plant viruses, IRESes have been reported, but, with one apparent exception [7
], they are smaller, less structured, and not as efficient as IRESes in animal virus RNAs [8
]. Instead of an IRES, many plant viruses contain a sequence in their 3′ UTR, which allows efficient cap-independent translation initiation via ribosome scanning from the 5′ end of the genome [12
]. These 3′ cap-independent translation elements (3′ CITEs) bind a surface of translation initiation factor heterodimer eIF4F with high affinity and, in most cases, are thought to deliver this and possibly other factors and the 40S ribosomal subunit, to the 5′ end via long-distance base pairing between the 3′ and 5′ UTR [15
3′ CITEs fall into at least seven categories, each with an entirely different structure [9
]. These include (i) the translation enhancer domain (TED) of satellite tobacco necrosis virus, consisting of a stem-loop with multiple bulges, [22
], (ii) the barley yellow dwarf virus (BYDV)-like 3’ CITE (BTE) consisting of two to five stem-loops branching from a stem helix at a single junction, with one stem-loop highly conserved in sequence [24
], (iii) the I-shaped domain of melon necrotic spot virus [18
], (iv) Y-shaped domain of carnation Italian ringspot virus [17
], (v) the tRNA-like T-shaped structure of turnip crinkle virus (TCV) [28
], (vi) the panicum mosaic virus-like CITE (PTE) [15
], and (vii) a pair of stem-loops forming a dumbell-shape in the Xinjiang isolate (but absent in European isolates) of cucurbit aphid-borne yellows virus (CABYV-X) also found in a resistance-breaking strain of Melon necrotic spot virus (MNSV) [33
]. With the exception of CABYV-X, all of the above viruses are in the Tombusviridae
family or the closely related Luteovirus
genus of the Luteoviridae
In plants, the key translation initiation factor complex, eIF4F, consists of two subunits, the cap-binding protein eIF4E, and the scaffolding protein eIF4G, which recruits other factors, including eIF4A and eIF4B, which together have helicase activity that aid in ribosome scanning [34
]. The BTE binds the eIF4G subunit of eIF4F [16
] and does not require eIF4E, eIF4A, or eIF4B, although these factors enhance the interaction of eIF4G with the BTE [16
]. The TED [35
] and Y-shaped structures [17
] appear to bind the eIF4F heterodimer only, with little affinity for either subunit alone, while the TCV T-shaped structure binds directly to the 60S ribosomal subunit [36
]. The I-shaped structure [37
] and the PTE [15
] bind the eIF4E subunit of eIF4F. The I-shaped structure requires presence of at least a portion of eIF4G bound to eIF4E, whereas the PTE binds eIF4E alone [15
]. This is remarkable in that the presence of the m7
G on the cap-structure on mRNA was thought to be required for eIF4E to bind RNA [38
]. All 3’ CITEs require eIF4G for function and bind with higher affinity to eIF4F than to the individual subunits alone. Plants, but not animals, contain two isoforms of the subunits of eIF4F: eIF4F, which consists of eIF4E plus eIF4G, and eIFiso4F, which consists of eIFiso4E plus eIFiso4G [39
]. eIFiso4E has about 50% sequence identity to eIF4E, while eIF4isoG is about 60% of the size of eIF4G with major deletions in domains of unknown function [40
]. eIF4F stimulates cap-dependent translation, and BTE- [16
] and PTE- [15
] mediated translation more efficiently than eIFiso4F. Thus, this paper focuses only on interactions of PTEs with eIF4E.
Previously, secondary structural analysis of PTEs of nine different tombusvirids revealed that they have little sequence conservation, and vary in many ways, but all form a roughly T-shaped structure with branching stem-loops in which a (usually) C-rich joining sequence (4–6 nt) at the branch-point of the stem-loops has potential to base pair to a G-rich bulge (~8–11 nt) in the main stem, forming a pseudoknot (Figure 1
A and Supplementary Figure S1
]. A diagnostic feature of all PTEs is that one G in the G-rich bulge can be hypermodified by SHAPE reagents in a magnesium-dependent manner [32
]. This G is not hypermodified in the absence of magnesium ion, or if the C-rich sequence is mutated to disrupt potential base pairing to the G-rich bulge. These mutations also inactivate the PTE [15
]. Correspondingly, eIF4E protects bases in the C-rich and G-rich regions from modification by SHAPE reagents, indicating that this pseudoknot base-pairing is the likely eIF4E binding site [32
To understand the structural basis for high affinity binding of eIF4E to the PTE in the absence of a 5′ cap, we undertook a comparative approach. eIF4E varies substantially in sequence, but only slightly in three-dimensional structure across kingdoms. Thus, we compared the efficiency of nine different PTEs in facilitating cap-independent translation in mammalian cells and cell lysates, and the ability of human eIF4E to bind two of these PTEs. We found a wide range in efficiency of translation facilitated by different PTEs, ranging from background levels, to half that conferred by one of the most active mammalian virus IRESes known. The efficiency of translation stimulation correlated with binding affinity of the PTE to eIF4E. To our knowledge, this work reveals the first 3′ CITE that functions in mammalian cells. This PTE may be a useful tool for engineered gene expression in mammalian systems, and it opens the possibility that such elements may exist in mammalian viral RNAs.
4. Materials and Methods
4.1. eIF4E Structure Alignment
eIF4E crystal structures from human (representing mammalian eIF4E) and wheat were obtained from Protein Data Bank (accession numbers 1IPC and 2IDV, respectively), overlaid using the “iterative Magic fit” tool, and RMS values calculated using Swiss PDB Viewer v. 4.03 [72
]. Cartoon renderings were produced using Persistence of Vision Raytracer (Pov-Ray) v. 3.62 (PoVP Ltd. 2004).
4.2. Plasmid Constructs and RNA Synthesis
The luciferase constructs used in translation assays consisting of the firefly luciferase gene (luc2, Promega) flanked by the indicated viral genomic 5′ and 3′ UTRs [32
] were linearized with Sma I and transcribed using the MEGAscript kit (Ambion). CAluc, which is a capped transcript containing nonviral UTRs described previously [73
], was linearized with EcoICRI, transcribed with the MEGAscript kit (Ambion) and post-transcriptionally capped using the T7 mScript Standard mRNA Production System (Cell Script).
-inhibition and structure probing experiments, the PTE sequences from the indicated viruses present in the universal SHAPE cassette [15
] were linearized with either HpaI (giving only PTE transcript for trans
-inhibition studies) or SmaI (which adds a 3′-terminal extension on the PTE providing a primer binding site for SHAPE experiments) and transcribed in vitro using the MEGAshortscript kit (Ambion). All transcripts were purified by phenol/chloroform extraction and ethanol precipitation. RNA concentrations were determined spectrophotometrically and integrity was verified by 0.8% agarose gel electrophoresis.
4.3. Recombinant Protein Expression
His-tagged wheat eIF4E in pET23d vector was introduced into E. coli (BL21 cells) and expression was induced at OD 600 nm = 0.8, with 100 mM IPTG. After 4 h of induction, cells were harvested from 1 L of culture by centrifugation at 10,000× g for 10 min. The cells were frozen at −80 °C for at least 1 h and sonicated 12 times for 30 s each with 2 min cooling on ice in binding buffer [25 mM HEPES–KOH at pH 7.6, 100 mM KCl, 2 mM MgCl2, 10% glycerol plus 0.1 mM phenylmethyl–sulphonyl fluoride, 0.1% Soybean trypsin inhibitor and 1 tablet/10 mL of Complete protease inhibitor cocktail, EDTA-free (Roche)]. The homogenate from 1 L of cells was centrifuged at 38,000× g for 20 min at 4 °C and supernatant was applied to 1 mL of Ni-NTA Superflow Cartridge (Qiagen). The cartridge was washed with 10 volumes of binding buffer plus 10mM imidazole and then with 10 volumes of binding buffer plus 20 mM imidazole. The his-tagged proteins were eluted with 250 mM imidazole in the same buffer. Human eIF4E used in structure probing experiments was a kind gift of Franck Martin, Institut de Biologie Moléculaire et Cellulaire Strasbourg, France.
4.4. Translation in Plant Systems
In vitro translation was set up using 2 nM of capped CAluc or uncapped viral luciferase reporter transcripts in either wheat germ extract as described by the manufacturer (Promega) or BY-2 cell extracts as described by [74
] with some modifications [42
]. Translation reactions were incubated for 1 h at 22 °C and the luciferase activity was measured by addition of 2 µL of the translation reaction to 40 µL of the Luciferase Assay Reagent (Promega). Relative light units were obtained using a GloMaxTM20/20 luminometer. Each construct was tested in triplicate in each experiment, and in at least three independent experiments.
For translation in vivo (protoplasts) 2 nM of capped CAluc or uncapped viral luciferase reporter transcript was mixed with 0.2 nM of capped and poly(A)60-tailed Renilla
-luciferase reporter mRNA [75
] and electroporated into approximately 1 million oat (Avena sativa
cv. Stout) protoplasts that were prepared as described previously [76
]. After 4 h incubation at room temperature, protoplasts were harvested and lysed in Passive Lysis Buffer (Promega). Both Renilla
and firefly luciferase activities were measured using Dual-Luciferase reporter assay system (Promega). To account for variations between experiments in electroporation efficiency and protoplast quality, the firefly luciferase activities were normalized to the Renilla
Trans-inhibition of translation.
Non-saturating amounts (2 nM) of uncapped transcript containing 5′ and 3′ UTRs of PEMV2 flanking firefly luciferase ORF, or capped CAluc transcript, pre-mixed with 200 nM of the designated viral PTEs were translated in wge (Promega) for 1 h at 22 °C as described [16
]. The luciferase activity was measured by addition of 2 µL of the translation reaction to 40 µL of the Luciferase Assay Reagent (Promega) followed by measurement of relative light units in a GloMaxTM20/20 luminometer. The relative activity obtained from P2lucP2 or CAluc in the absence of inhibitor was defined as 100%.
4.5. Translation in the Mammalian Systems
T7-BHK cells were cultured as described previously [52
]. Briefly, cell growth media was prepared by supplementing Glasgow Minimal Essential Media (GMEM) with amino acids, 10% fetal bovine serum (FBS), and penicillin-streptomycin. G418 (geneticin), was added every other passage at a concentration of 1 mg/mL to maintain the T7 expression construct. The cells were grown at 37 °C with 7% atmospheric CO2
, and passaged when they reached approximately 80% confluency.
Transfection was performed using Invitrogen Lipofectamine Transfection Reagent using the manufacturer’s protocol on cells that had been seeded onto 6-well plates at a density of 2 × 105 cells/well and allowed to reach approximately 80% confluency. Significant cell death was not observed upon transfection. Twenty-four hours after transfection, the cells were placed on ice, removed from the plate by scraping, and lysed using Lysis Buffer (Promega). Lysate was diluted 1:4 in Lysis Buffer, then 20 µL was added 100 µL of Luciferase Assay Reagent (Promega), and quantified with the GLO-MAX luminometer, or frozen and stored at −80 °C until quantification.
4.6. Electrophoretic Mobility Shift Assay (EMSA)
Prior to electrophoresis, 2000 cpm 32P-labeled RNA (20 fmol) was incubated with indicated concentrations of purified eIF4E in binding buffer: 10 mM HEPES pH 7.5, 20 mM KCl, 1 mM dithiothreitol, 3 mM MgCl2, 1 µg/µL tRNA, 0.5 µg/µL bovine serum albumen, 1 unit/µL RNaseOUT™ Recombinant Ribonuclease Inhibitor (Invitrogen), 5% glycerol, in a total volume of 10 µL for each sample. After 25 min on ice, 3 µL 50% glycerol with bromophenol blue was added, and the RNA-protein mixture was loaded on a 5% polyacrylamide (acrylamide:bis-acrylamide 19:1), Tris-borate/EDTA gel. After electrophoresis at 110 V for 45–60 min (4 °C), gel was dried and exposed to phosphorimager screen overnight. After scanning the screen in a Bio-Rad PhosphorImager, data were analysed using Quantity One software (Bio-Rad) to calculate the apparent Kds. Shifted (bound) and unshifted (unbound) bands were each quantitated from three separate experiments. Standard error was calculated and significance of pairwise comparisons of human vs wheat eIF4E, TPAV versus TPAVm2 PTE, and capped versus uncapped RNAs was determined using Student’s t-test.
4.7. RNA Structure Probing and Footprinting
Chemical and enzymatic RNA structure probing was performed as described previously [42
]. Briefly, 500 ng of refolded RNA alone or pre-incubated for 10 minutes with indicated proteins was treated with 10% (v/v) of benzoyl cyanide (Sigma-Aldrich) and incubated for 30 s at 22 °C. As a control, RNA refolded in the presence of 3 mM Mg2+
was treated with 10% (v/v) of DMSO in place of chemical reagent. RNA was then purified by phenol–chloroform extraction and ethanol precipitation. Reactions were resolved on an 8% denaturing polyacrylamide gel and dried following primer extension. Dried gels were exposed to a storage phosphor screen as described previously [32
]. Each experiment was repeated at least three times.