The Hepatitis C Virus Internal Ribosome Entry Site Adopts an Ion-dependent Tertiary Fold

Mechanism of ribosome recruitment by hepatitis C IRES RNA

“…Eukaryotic cells and their viruses have evolved at least two mechanisms for recruiting and positioning ribosomes at the start sites for translation of RNA messages+ The primary mechanism involves recognition of a 7-methyl guanosine cap on the 59 terminus of the mRNA by a set of canonical initiation factors that recruit the 43S particle-including the 40S ribosomal subunit and eukaryotic initiation factor 3 (eIF3)-forming the 48S preinitiation complex (Fig+ 1A; for review, see Merrick & Hershey, 1996;Pain, 1996;Sachs et al+, 1997)+ Alternatively, numerous viruses and some eukaryotic mRNAs utilize a cap-independent pathway in which an RNA element, the internal ribosome entry site (IRES), drives preinitiation complex formation by positioning the ribosome on the message, either at or just upstream of the start site+ In hepatitis C virus (HCV), the major infectious agent leading to non-A, non-B hepatitis, the minimum IRES includes nearly the entire 59 untranslated region (UTR) of the message (for review, see Rijnbrand & Lemon, 2000)+ The secondary structure of the HCV IRES RNA, one of the most conserved regions of the entire viral genome, is critical for translation initiation, and is similar to that of the related pestiviruses and GB virus B (Brown et al+, 1992;Wang et al+, 1994Wang et al+, , 1995Le et al+, 1995;Rijnbrand et al+, 1995;Honda et al+, 1996aHonda et al+, , 1996bHonda et al+, , 1999Pickering et al+, 1997;Varaklioti et al+, 1998;Psaridi et al+, 1999;Tang et al+, 1999) + We have previously shown that the HCV IRES RNA adopts a specific three-dimensional fold in the presence of physiological concentrations of metal ions (Kieft et al+, 1999)+ Rather than forming a tightly packed globular structure, the RNA helices extend from two folded helical junctions, JIIIabc and JIIIef (Fig+ 1B)+ This suggests that the IRES RNA acts as a structural scaffold in which specifically placed recognition sites recruit the translational machinery+ This is supported by the observation that eIF3 and the 40S ribosomal subunit, the two largest components of the 43S particle, bind directly to the HCV IRES RNA (Pestova et al+, 1998)+ Unlike IRESs found in some other RNA viruses, such as poliovirus, the IRES RNA•40S•eIF3 ternary pre-initiation complex forms without the involvement of other cellular factors (Fig+ 1A;Pestova et al+, 1998)+ Although several other proteins appear to interact with the HCV IRES RNA, they are not required for 43S binding to the IRES …”

“…Schematic of HCV IRES RNA ribosome recruitment and secondary structure+ A: Examples of strategies used to recruit the 43S particle to the mRNA start codon+ In eukaryotes, the mRNA cap binds eIF4E and eIF4G, which leads to 43S particle binding through the eIF3 component+ In poliovirus, binding of the 43S particle uses a cleaved eIF4G to recruit the 43S particle, followed by movement to the start codon+ In HCV, 43S particles bind directly to the IRES RNA through the 40S subunit and eIF3, and no scanning is involved+ B: The secondary structure of the minimum HCV IRES RNA sequence that is able to initiate translation is depicted in black, with the location of point mutations used in specificity assays in red; secondary structural elements are labeled+ Red circles indicate the two independently folded four-way junctions that fold in the presence of metal ions+ Domain II also folds independently as assayed by RNase T1 probing (data not shown)+ Single-stranded regions that remain flexible upon IRES RNA folding are indicated with green+ The start codon is located in domain IV+ variety of salt conditions+ We also verified the 1:1 ratio of the IRES RNA•40S complex using stoichiometric binding assays (data not shown)+ Because previous experiments demonstrated that IRES RNA folding requires less than 1 mM magnesium ion, we measured the K d of the IRES RNA•40S complex as a function of magnesium ion concentration (Fig+ 2B)+ Binding is almost nonexistent in the absence of added magnesium, but increases with increasing magnesium ion concentration up to ;2 mM+ The magnesium ion dependence of IRES RNA•40S complex formation is very similar to that for IRES RNA tertiary structure formation and for maximal HCV IRES driven translation initiation in vitro (Borman et al+, 1995;Kieft et al+, 1999)+ This result suggests that binding of the 40S subunit to the IRES RNA and subsequent translation initiation requires the formation of the magnesium-induced IRES tertiary structure, although magnesium-dependent ribosome stability may also play a role+ To explore the specificity of the IRES RNA•40S interaction, we measured the affinity of the 40S subunit for HCV IRES RNA point mutants that have been previously characterized in terms of folding and in vitro translation initiation activity (Kieft et al+, 1999)+ Point mutations in loop IIId (U264A, U265A, U269A, G266C, and G267C) have been shown to affect both secondary structure formation and in vitro translation initiation activity (Fig+ 1B)+ Mutations of the uracils within loop IIId that do not change the RNA secondary structure, as assayed by RNase T1 probing (data not shown), have an intermediate effect on IRES activity, and do not affect 40S subunit binding (Table 1)+ However, mutations of guanosines in loop IIId have stronger deleterious effects on both translation initiation and 40S subunit binding, and in the case of G266C, alter the ion-dependent structure of the RNA (Kieft et al+, 1999)+ A single mutation at U228 that prevents the folding of junction JIIIabc (U228C) initiates translation at only 5% FIGURE 2. Binding of wild-type and mutant HCV IRES RNAs to components of the 43S particle+ A: Binding isotherm of wild-type HCV IRES RNA to purified 40S subunits and H+ marismortui 30S subunits+ B: Magnesium dependence of the K d of IRES RNA•40S complex formation is plotted in red; the curve for IRES RNA folding as a function of magnesium ion concentration is plotted in black (for clarity, the data points for the fraction folded curve have been omitted; Kieft et al+, 1999) (Garcia-Barrio et al+, 1995)…”

“…The elements of the IRES RNA secondary and tertiary structure that confer 40S subunit affinity and specificity were identified using a combination of footprinting and modification interference experiments+ To determine the footprint of specifically bound 40S subunit on the IRES RNA, we probed the complex using ribonuclease (RNase) T1 and phosphorothioate footprinting+ Upon folding, recognition loops in the IRES RNA remain sensitive to cleavage by RNase T1 (Kieft et al+, 1999 (Kolupaeva et al+, 2000)+ To probe the entire RNA, independent of sequence or secondary structure, we used iodine cleavage of HCV IRES RNA molecules that were transcribed with 5% incorporation of phosphorothioate linkages (Christian & Yarus, 1992 Footprinting identifies areas of intermolecular contact, but does not yield information about which nucleotides within the footprint determine binding affinity+ To address this, we used modification interference, a technique in which functional RNAs are selected from a pool of randomly modified molecules (Conway & Wickens, 1989)+ Radiolabeled HCV IRES RNA was modified with diethylpyrocarbonate (DEPC), which ethylates adenosine bases at the N7 position, to produce a pool of molecules in which each RNA contained, on average, one modified adenosine+ This RNA was incubated with 40S subunits and the bound IRES RNA was recovered and analyzed to identify locations where modification interferes with 40S subunit binding+ This analysis reveals that adenosine bases involved in 40S subunit recognition are clustered around JIIIef and loop IIIa (Fig+ 3C,D)+ Loop IIIa is not involved in forming the ion-induced fold of the IRES RNA as assayed by RNase T1 probing (data not shown); thus it may represent a specific recognition element for the IRES RNA•40S interaction+ Loop IIIe bases are protected from RNase T1 by the ion-induced fold of the RNA; thus modifying these bases may affect folding or disrupt a direct contact to the 40S subunit (Kieft et al+, 1999)+ Adenosine bases located in both helices of a proposed pseudoknot (IIIf; Fig+ 1B) may also be important for stabilizing this tertiary interaction or for directly interacting with the 40S subunit+ No critical adenosine residues were identified in domains II, IIIb, or IV, indicating that while these regions are proximal to the 40S subunit, they do not contain adenosines whose modification affects 40S subunit recognition+ Apical loops IIIc and IIId are included in the 40S subunit footprint, but as they do not contain adenosine residues, their participation in binding cannot be assessed by DEPC modification+…”

“…a Activity is reported as percentage of wild-type translation initiation activity+ b Previously reported value (Kieft et al+, 1999)+ c n+d+: not done+ FIGURE 3. Footprinting and modification interference of binding of the IRES RNA-40S subunit complex+ A: RNase T1 footprinting of the HCV IRES RNA with bound 40S subunit+ RNase T1 cuts after guanosine residues in single-stranded RNA+ Lane 5 contains RNA without added magnesium, lane 6 contains folded RNA, and lane 7 contains folded RNA with bound 40S subunit+ The location of various secondary structural domains and their associated cleavage sites are shown with arrows+ IIId, IIIa, and IIb are among those loops that are protected+ B: Portion of one gel from thiophosphate footprinting of 40S subunit onto the HCV IRES RNA, using a-thio-G+ Lanes 1 and 2 contain RNase U2 and T1 sequencing reactions, lanes 3 and 4 contain RNA that was not treated with molecular iodine+ Lane 5 contains RNA that was denatured and immediately treated with iodine to yield a thiophosphate-G sequencing lane+ The footprint is observed by comparing the subtle protections observed in lane 7 with lane 6+ Examples of sites of protection are labeled with arrows; sites of background degradation are indicated with asterisks+ Reported sites of protection were observed in multiple experimental repetitions+ Typical protection factors were between 1+5 and 3+ C: Portion of a DEPC modification interference of 40S binding to the HCV IRES RNA experimental gel+ Lane 1 contains starting material, lanes 2 and 3 are RNase sequencing reactions, and lane 4 is a partial alkaline hydrolysis ladder+ Interferences are observed by comparing lanes 5 and 6; sites of interference are indicated with arrows+ D: Summary of footprinting and modification interference data+ In the presence of bound 40S subunits, sites that were protected from RNase T1 are indicated with asterisks, sites protected from iodine cleavage are circled, and adenosine nucleotides implicated by modification interference are shown with arrows+ structed a series of IRES sequences with mutated apical loops to quantitate the thermodynamic contribution of each IRES RNA stem-loop to complex formation (Fig+ 4A)+ In several cases, the apical loops were mutated to both their Watson-Crick complements and to an ultrastable GAAA tetraloop, to control for secondary structural changes or binding effects unique to one type of mutant+ Mutants IIIc_GAAA, IIIb_BVDV, and IIIa_comp were probed with RNase T1 to verify that they did not alter the ion-induced fold (data not shown)+ As predicted, mutations in stem-loops identified by modification interference (IIIa and IIIe) caused a reduction in binding affinity, as did mutations in other loops within the 40S subunit footprint (IIIc, IIId; Table 2)+ Translation initiation activity of these mutants correlates with ability to bind to the 40S subunit, as they are all severely inhibited+ The loop IIIb mutant bound with wild-type affinity, and also was able to initiate translation at 50% of wild-type activity+ Loop IIIf is proposed to form a pseudoknot by base pairing to nucleotides just upstream of the stem-loop that contains the start codon (domain IV; Wang et al+, 1995)+ To explore the possible role of this structure in preinitiation complex formation, we prepared a series of pseudoknot mutants (Fig+ 4B), and measured their 40S subunit binding ability+ Abrogation of the potential pseudoknot interaction by mutation of either nt 325-330 or by mutation of loop IIIf results in a two-to threefold reduction in 40S subunit binding affinity (Table 2), yet IRES translation initiation activity is lost+ A somewhat stronger effect is observed with mutants in which the potential pseudoknot interaction is partially or fully restored through compensatory mutations (Fig+ 4B)+ (Goss & Rou...…”

“…Plasmids encoding the desired RNA sequences were generated using standard cloning techniques+ The construction of plasmids containing wild-type genotype 1b HCV IRES RNA sequence nt 40-372 (p1b35+2), and mutants G266C, G267C, G268C, G(266-268)C, and U288C have been previously described (Kieft et al+, 1999)+ Mutants were generated with either PCR with p1b35+2 as the original template, followed by ligation into the EcoR1/BamH1 site of pUC 19, or using Kunkel site-directed mutagenesis (Stratagene)+ All RNAs contained a hammerhead and hepatitis delta virus ribozyme on the 59 and 39 ends, respectively+ RNA transcription, purification, and end labeling RNA was transcribed in vitro, purified, and 59 end labeled as previously described (Kieft et al+, 1999)+ To 39 end label RNA, the 29-39 cyclic phosphate was first removed by treatment with T4 polynucleotide kinase (PNK) and calf intestinal phosphatase (CIP) (Cameron & Uhlenbeck, 1977)+ Briefly, approximately 100 mg of RNA were treated with 20 mL PNK (NEB) and 10 mL CIP (NEB) in 40 mM MES-NaOH, pH 6+0, 10 mM MgCl 2 , 5 mM DTT, and DEPC-treated water to 100 mL final volume+ After incubation at 37 8C for 2-3 h, the reaction was phenol/chloroform extracted, ethanol precipitated, washed with 70% ethanol, dried, and resuspended in 1ϫ TE buffer+ RNA was then 39 end labeled with 32 P-pCp and T4 RNA ligase (England & Uhlenbeck, 1978)+ The RNA labeling reaction contained 5-10 mg 39 dephosphorylated RNA, 2 mL T4 RNA ligase buffer (NEB), 4 mL T4 RNA ligase (NEB), 2 mL DMSO, and 8 mL 32 P-pCp in 20 mL total volume+ The reaction was incubated for 16 h at 16 8C, then purified as described (Kieft et al+, 1999)+ Isolation and purification of 40S ribosomal subunits and eIF3 from rabbit reticulocyte lysate 40S ribosomal subunits were isolated from rabbit reticulocyte lysate (RRL; Green Hectares) essentially as described (Pestova et al+, 1996)+ Briefly, RRL was thawed on ice in the presence of PMSF and leupeptin and spun in a Beckman ultracentrifuge at 4 8C at 30,000 rpm in a Ti 55+2 rotor for 4 h+ The ribosome pellet was resuspended in buffer D (5 mM Tris-HCl, 0+25 M sucrose, 0+1 mM EDTA, 1+0 mM DTT, pH 7+5), using 2 mL of buffer per centrifuge tube+ We slowly added 4 M KCl to the polysome suspension to yield a final concentration of 0+5 M, followed by gentle rocking for 30 min, then by ultracentrifugation in a Beckman 55+2 Ti rotor at 4 8C at 40,000 rpm for 4 h+ The 0+5 M KCl ribosomal salt wash (0+5 M RSW) supernatant was decanted and used to obtain eIF3 (see below)+ The pellet was resuspended in buffer A (20 mM TrisHCl, 50 mM KCl, 4 mM MgCl 2 , 2 mM DTT, pH 7+5) to a concentration of 100 ODU 260nm /mL+ Puromycin in buffer A was added to a concentration of 1 mM, and the ribosome suspension was incubated for 10 min on ice, then for 10 min at 37 8C, followed by the addition of 4 M KCl to a final concentration of 0+5 M+ The ribosome suspension was then layered onto a 10-30% sucrose gradient in buffer B (20 mM Tris-HCl, 0+5 M KCl, 3 mM MgCl 2 , 2 mM DTT, pH 7+5) and spun at 22,000 rpm in a Beckman SW28 rotor at 4 8C for 17 h+ The 40S subunits were recovered by fractionating the gradient and detecting the absorbance at 280 nm+ Fractions containing the 40S subunit were pooled, concentrated, and exchanged into buffer C (0+25 M sucrose, 20 mM Tris-HCl, pH 7+5, 10 mM KCl, 1 mM MgCl 2 , 1 mM DTT, 0+1 mM EDTA) using Centricon 50 spin concentrators (Amicon) and stored at Ϫ20 8C+ Concentration of 40S subunits was determined by spectrophotometry, using the conversion 1 ODU 260nm ϭ 50 pmol+ The concentration of 40S subunits capable of binding the IRES RNA (active concentration) ...…”