Inhibitors of protein synthesis identified by a high throughput multip
http://www.100md.com
《核酸研究医学期刊》
1 Department of Biochemistry, 2 McGill High Throughput Screening Facility and 3 McGill Cancer Center, 3655 Promenade Sir William Osler, McIntyre Medical Sciences Building, McGill University, Montreal, Quebec H3G 1Y6, Canada
*To whom correspondence should be addressed at: McIntyre Medical Sciences Building, Room 810, 3655 Promenade Sir William Osler, McGill University, Montreal, Quebec H3G 1Y6, Canada. Tel: +1 514 398 2323; Fax: +1 514 398 7384; Email: jerry.pelletier@mcgill.ca
ABSTRACT
The use of small molecule inhibitors of cellular processes is a powerful approach to understanding gene function that complements the genetic approach. We have designed a high throughput screen to identify new inhibitors of eukaryotic protein synthesis. We used a bicistronic mRNA reporter to multiplex our assay and simultaneously screen for inhibitors of cap-dependent initiation, internal initiation and translation elongation/termination. Functional screening of >90 000 compounds in an in vitro translation reaction identified 36 inhibitors, 14 of which are known inhibitors of translation and 18 of which are nucleic acid-binding ligands. Our results indicate that intercalators constitute a large class of protein synthesis inhibitors. Four non-intercalating compounds were identified, three of which block elongation and one of which inhibits initiation. The novel inhibitor of initiation affects 5' end-mediated initiation, as well as translation initiated from picornaviral IRESs, but does not significantly affect internal initiation from the hepatitis C virus 5'-untranslated region. This compound should be useful for delineating differences in mechanism of initiation among IRESs.
INTRODUCTION
Inhibition of prokaryotic protein synthesis by small molecule ligands has provided modern medicine with an important arsenal of anti-infective compounds and has helped dissect the pathway of protein synthesis. The cumulative knowledge gained from studying inhibitors of translation demonstrates the power of applying a chemical biology approach to study this process. Pharmacological inhibitors of protein synthesis have allowed the characterization of events leading to the assembly of active polysomes, trapping intermediates of the initiation and elongation cycle and providing insight into the molecular functions of protein factors (1,2).
Surprisingly, no report documents the systematic search for new inhibitors of eukaryotic protein synthesis, with the majority of compounds currently used having been identified several decades ago (1,2). The impetus for establishing new chemical screens stems not only from the usefulness of protein synthesis inhibitors as tools, but from the growing number of studies implicating deregulated translation as a contributor to cancer initiation and progression (3,4). For example, overexpression of some initiation factors can lead to malignant transformation, whereas down-regulation of these same factors can suppress the transformed phenotype (reviewed in 5). Components of the translation apparatus are overexpressed or mutated in cancers (5). The tumor suppressor gene product (pRB) directly impacts on the translation process by affecting the levels of ribosomes (for a review see 6). The PI3K/Akt signaling pathway is deregulated in a large number of cancers and modulates activity of factors regulating initiation of protein synthesis (7). As such, translation represents an interesting, under-explored target for chemotherapeutic intervention.
The majority of known protein synthesis inhibitors bind to ribosomes, with some of these making key contacts with the RNA component (8–11). These findings validate the idea of structured RNA as a biological target for small molecule interdiction. Additionally, recent studies have demonstrated that highly structured domains within prokaryotic mRNAs can serve as metabolite-responsive genetic switches (12). In the cases studied, the metabolite binds to an RNA motif and induces an allosteric change that alters transcription termination (13) or translation initiation (14). In eukaryotes, small molecule–RNA interactions (15,16) or small molecules acting as chemical inducers of dimerization (17) can inhibit protein synthesis by interfering with translation initiation. The potential for regulating gene expression with chemical ligands that modulate mRNA function has remained largely unexplored. Many features of RNA make it an attractive target: it is central to many functions of the cell, contains complex secondary and tertiary structural folds and lacks a cellular repair mechanism. The ability of small molecules to interact with RNA is well recognized (18) and these interactions have the potential to prevent or enhance gene expression, achieve allele-specific modulation of gene expression (when allelic sequence differences result in altered RNA conformations), achieve isoform-specific modulation of gene expression and exhibit allosteric effects.
We have undertaken a high throughput screen designed to identify general inhibitors of eukaryotic protein synthesis, as well as small molecule ligands that inhibit gene-specific expression at the level of mRNA–ribosome interaction. The use of a bicistronic mRNA reporter in an in vitro translation extract allowed us to multiplex the assay, thus reducing our reagent cost and consumption of small molecule ligand collection. This report demonstrates the validation and execution of a functional eukaryotic translation screen of >90 000 distinct compounds. Herein, we have identified several novel protein synthesis inhibitors and characterized their activity profiles.
MATERIALS AND METHODS
Materials and general methods
Restriction endonucleases, calf intestinal alkaline phosphatase, the Klenow fragment of DNA polymerase I, T4 DNA ligase and SP6 RNA polymerase were purchased from New England Biolabs (Beverly, MA). cytidine triphosphate (20.5 Ci/mmol) and methionine (>1000 Ci/mmol) were obtained from Perkin Elmer Life Sciences (Boston, MA). Preparation of plasmid DNA, restriction enzyme digestion, agarose gel electrophoresis of DNA and RNA, DNA ligation and bacterial transformations were carried out using standard methods (19).
Plasmid construction and cell-free extract preparation
The bicistronic vector, pSP/(CAG)33/FF/HCV/Ren·pA51, was generated as follows. A sequence encoding a poly(dA) tail was inserted immediately downstream of the renilla luciferase gene by inserting an oligonucleotide containing 51 dA residues d(CTAGGTTCAAGCTTCGGTGGA51GGGG) into the XbaI and BamHI sites of phRL-null (Promega). This plasmid was called phRL-null·pA51. The HCV 5'-untranslated region (5'-UTR) (nucleotides 14–383) was amplified by PCR using oligonucleotides A and B from an HCV internal ribosome entry site (IRES) bicistronic plasmid (kindly provided by Dr N. Sonenberg, McGill University), digested with EcoRI and XbaI and cloned into phRL-null·pA51 which had been linearized with EcoRI and NheI. This places the HCV ATG initiation codon (nucleotide 342) in-frame with the renilla initiation codon. The resulting vector was called HCV/RL·pA51. Plasmid pGL3-Basic (Promega) was digested with XbaI, Klenow fragment repaired and the firefly luciferase gene excised with BglII and ligated into HCV/RL·pA51 which had been digested with EcoRI (blunted with Klenow fragment) and BglII. The resultant vector was called pFF/HCV/RL·pA51. Thirty-three CAG repeats were cloned into the EcoRI site of pSP64 using an oligonucleotide containing the sequence d(GAATTC15CAATT15CAATT3CAATTC), to generate pSP/(CAG)33. The FF/HCV/RL·pA51 fragment was excised from pFF/HCV/RL·pA51 with BamHI and BglII and inserted into pSP/(CAG)33 which had been digested with BamHI. All cloned PCR products and oligonucleotide inserts were sequenced to ensure the absence of undesirable mutations.
The construction of pKS/FF/EMC/Ren was undertaken by first digesting E·CAT (which contains the 5'-UTR of EMC) with MscI and ligating NcoI linkers d(CAGCCATGGCTG). The EcoRI–NcoI fragment was then transferred using the same sites in phRL-null·pA. The resultant plasmid was digested with EcoRI, blunted with the Klenow fragment of DNA polymerase I and digested with BglII. The EMC/renilla-containing fragment was inserted into the XbaI (blunted) and BglII sites of pGL-3. The BglII–BamHI fragment containing FF/EMC/Ren was then transferred into the BamHI site of pKSII+. To generate plasmid pKS/FF/Ren, phRL-null·pA was digested with EcoRI, blunted with the Klenow fragment of DNA polymerase I and digested with BglII. The released fragment was cloned into the XbaI (blunted) and BglII sites of pGL-3. The BglII–BamHI fragment containing FF/Ren was then transferred into the BamHI site of pKSII+. Plasmid pcDNA3/Ren/P2/FF was a kind gift of Dr N.Sonenberg (McGill University).
In vitro Krebs translation extracts were prepared as follows. Krebs ascites cells were passaged in mice for a period of 8–10 days, after which time they were harvested, washed with cold HNG buffer (35 mM HEPES pH 7.3, 10 mM glucose, 145 mM NaCl) and resuspended in 2 packed cell volumes of hypotonic buffer (25 mM HEPES pH 7.3, 50 mM KCl, 1.5 mM MgCl2, 1 mM DTT). The cells were left to swell on ice for 20 min, transferred to a 40 ml type B Dounce homogenizer (Bellco) and lysed (20–40 strokes) (lysis was monitored by staining cells with Trypan blue and visual inspection by microscopy). Approximately 60–80% lysis was routinely achieved. A one-ninth volume of 10x concentrate (250 mM HEPES pH 7.3, 1 M KOAc, 29 mM MgCl2, 30 mM DTT) was added to the lysate, mixed gently and centrifuged at 17 000 g for 20 min. The supernatant was collected, aliquoted and stored at –80°C. A typical yield of translation extract from 30 mice was 100 ml.
Discrete compound collection, liquid handling and data capture
The discrete compound collection used in the high throughput screen consisted of 40 000 synthetic compounds from ChemBridge (San Diego, CA), 1680 synthetic and natural products from MicroSource Discovery (Gaylordsville, CT), 42 000 synthetic compounds from NCI-DTP, 2056 compounds from TimTec (Neward, DE), 8669 InterBioScreen (IBS) compounds (Moscow, Russia) and 800 purified natural products collected by one of us (J.P.). Each 96-well plate had 80 discrete compounds, four negative controls (DMSO controls), two positive controls (10 μM anisomycin) and two wells containing firefly luciferase protein. Compounds re-ordered from the Drug Synthesis and Chemistry Branch of the National Cancer Institute, IBS, or TimTec were dissolved in DMSO and equivalent amounts of solvent were included in all in vitro control reactions.
All liquid handling steps were performed on a Biomek FXII pipetting station (Beckman). Liquid handling protocols were optimized: (i) for the distribution of programmed translation mix and compound to minimize formation of air bubbles, (ii) to ensure proper reagent mixing; (iii) to eliminate cross-well and nuclease contamination. All data were captured and analyzed using ActivityBaseTM (IDBS, Boston, MA). To determine the inhibitory effect of the compounds tested, data for each compound were internally normalized to the control mean values (translations in the absence of compound) on each plate and expressed as relative inhibition.
In vitro translation assays and secondary assay
In vitro transcriptions were performed essentially as described previously (17). In vitro translations in rabbit reticulocyte lysate and Escherichia coli S30 extracts were performed according to the manufacturer’s recommendations (Promega). In vitro translations in Krebs extracts were prepared as described previously (20), except that they were formatted in 96-well conical bottom plates (Costar 3897). Ten microliter translation reactions were set up containing 4 ng/μl in vitro transcribed mRNA. The addition of compound to a final concentration of 10 μM was performed with a solid pin tool (V&P Scientific FP3 pins). The pin tool was washed in water, ethanol and DMSO, then blotted onto a cleaning pad (V&P Scientific 540D) after each compound transfer to avoid carryover and contamination. Translation plates were sealed and incubated at 30°C for 1 h, after which time 25 μl of luciferase stabilization buffer (70 mM HEPES pH 7.7, 7 mM MgSO4, 3 mM DTT, 1% BSA) was added and the plates frozen at –20°C until processed for luciferase measurements.
For in vitro translations using methionine and exogenously supplied (CAG)33/FF/HCV/Ren·pA51 mRNA, Krebs extracts were first treated with micrococcal nuclease to digest endogenous mRNA, then programmed with mRNA at a final concentration of 15 ng/μl. For gel analysis of translation products, aliquots from translation mixtures were mixed with SDS/sample buffer, boiled for 2 min and applied to a 10% SDS–polyacrylamide gel. Gels were fixed in 10% methanol, 7.5% acetic acid, treated with EN3HANCE (NEN) and exposed to X-OMAT (Kodak) film at –70°C with an intensifying screen. For in vitro translations using methionine incorporation into endogenous protein, translations were set up as above except that the micrococcal nuclease treatment step was omitted. methionine incorporation was assessed by spotting an aliquot of the lysate onto Whatman 3MM paper (that had been pre-blocked with 0.1% methionine), dried and placed in cold 10% TCA for 20 min. Filters were transferred to 5% TCA, boiled for 15 min, washed once with 5% TCA and once with 95% ethanol and dried. Radioactivity was determined by scintillation counting.
A counterscreen was developed to identify compounds that inhibited firefly or renilla luciferase activity non-specifically. Large scale translations (500 μl) in Krebs extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA were performed at 30°C for 1 h. Aliquots of 9 μl were placed in white flat-bottomed flash plates (Costar 3912) and 1 μl of appropriately diluted compound added. Plates were then processed for luciferase measurements.
Dual-luciferase reporter assay
Firefly and renilla luciferase activities were measured using a plate reading luminometer equipped with two reagent injectors (Lmax; Molecular Devices). Dual activity measurements were performed by injecting 50 μl of firefly luciferase reagent (21), waiting for 2 s, followed by a 10 s measurement period. An aliquot of 50 μl of renilla luciferase reagent was then injected, followed by a 2 s pre-measurement delay and a 10 s measurement period.
Ribosome-binding assays
Ribosome-binding assays were performed as described previously (17). Essentially, 32P-labeled CAT mRNA was incubated in 25 μl of extract at 20°C for 10 min in the presence of 0.6 mM cycloheximide or 1 mM GMP-PNP and 50 μM small molecule ligand. The final KOAc concentration was adjusted to 150 mM. Initiation complexes formed were analyzed by sedimentation through 10–30% sucrose gradients. Centrifugation was for 3.5 h at 39 000 r.p.m. at 4°C in an SW40 rotor. Fractions of 500 μl were collected and radioactivity was determined by scintillation counting.
RESULTS
Rationale of the assay
We have devised a multiplex screen based on in vitro translation of a specific mRNA reporter that enables us to simultaneously identify (i) inhibitors of cap-dependent translation (Fig. 1A, step 1), (ii) small molecules that bind to a particular RNA motif in the 5'-UTR of the first cistron and inhibit ribosome scanning (step 2), (iii) inhibitors of elongation or termination (steps 3 and 4) and (iv) inhibitors of IRES-mediated initiation (step 5). The basis of the screen lies in the ability to quantitatively assess the products of the firefly (FF) and renilla (Ren) cistrons, following in vitro translation of the test transcript. Any ligand that inhibits 5'-mediated (cap-dependent) initiation or that binds to a particular RNA motif within the 5'-UTR and inhibits 40S ribosome scanning will decrease expression of firefly luciferase, but should not affect renilla luciferase activity. We engineered 33 CAG repeats within the FF 5'-UTR to screen for compounds that could bind to the imperfect stem–loop structure predicted to be formed by this trinucleotide repeat (Fig. 1B) and subsequently inhibit ribosome scanning. Small molecule ligands interacting with RNA structural motifs within mRNA 5'-UTRs have been previously shown to inhibit translation (15,16) at the level of initiation (16). The (CAG)33 repeat was chosen because of its implication in the pathogenesis of Huntington’s disease through expansion of the repeat within the huntingtin coding region, the goal being to identify a small molecule ligand that could preferentially inhibit translation of the mRNA encoded by the allele carrying the expanded CAG repeat, due to increased target copy number.
Figure 1. Outline of the strategy used to multiplex the high throughput search for inhibitors of protein synthesis. (A) In the configuration of the assay, a bicistronic mRNA is used in which initiation can occur at the 5' end via a cap-dependent mechanism, as well as internally via an internal ribosome entry site (IRES). The firefly (FF) luciferase gene is the first shaded box and the renilla (Ren) luciferase is the second black box. The hepatitis C virus (HCV) IRES is denoted by a thickened line. The plasmid is prepared for in vitro transcription by linearizing with BamHI. Translation initiation of renilla luciferase is driven by the HCV IRES. The position of 33 CAG nucleotide triplets upstream of the firefly luciferase initiation codon is denoted. Inhibition of translation in a cell-free translation extract by a small molecule ligand can occur at one of five sites: by interfering with cap-dependent initiation (Step 1), by interfering with migration of the 40S subunit along the 5'-UTR of the mRNA (Step 2), by inhibiting elongation (Step 3), by inhibiting termination (Step 4) or by inhibiting binding of the 40S subunit to the IRES (Step 5). (B) Predicted secondary structure (G = –38 kcal/mol) formed by the CAG tract within the (CAG)33/FF/HCV/Ren·pA51 5'-UTR using Mfold (48). The bracketed region indicates that this section of the stem is repeated 13 times. (C) Schematic representation of the strategy undertaken to identify novel inhibitors of eukaryotic protein synthesis. See text for details.
Translation initiation on HCV occurs by a cap-independent mechanism and is directed by a highly structured 5'-UTR, consisting of four stem–loops and a pseudoknot structure. It contains a unique cis-acting element known as an IRES that mediates cap-independent internal initiation (22). The hepatitis C virus IRES is significantly different from other IRESs since it does not have a strict requirement for the canonical initiation factors (eIFs), other than eIF2 and eIF3 (23). Hence, a decrease in renilla luciferase activity, but not of firefly luciferase production, should occur if a small molecule ligand inhibits HCV IRES-mediated initiation. A general inhibitor of translation (initiation, elongation or termination) would be expected to reduce both firefly and renilla luciferase production.
The strategy we have established for the multiplexed screen is outlined in Figure 1C. Following the primary in vitro translation screen, positive hits are re-tested in duplicate to ensure that the initial result was a true positive. A counterscreen is then performed to eliminate non-specific inhibitors of luciferase enzyme activity. Candidates are evaluated in a secondary assay in which in vitro translation reactions in micrococcal nuclease-treated Krebs extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA are analyzed by polyacrylamide gel electrophoresis. At this point, compounds that inhibited in the secondary assay are considered candidate inhibitors.
Assay validation
To maximize the sensitivity of the in vitro translation system for detection of protein synthesis inhibitors using the bicistronic luciferase reporter several conditions first had to be met. The choice of the cell-free extract was important to allow both cap-dependent initiation and HCV IRES-mediated internal initiation. For eukaryotic in vitro translation screens, we had the option of using several extracts, derived from wheatgerm, rabbit reticulocyte lysates, Krebs cells or HeLa cells. Wheatgerm extracts were not chosen since they do not support translation initiation from the HCV IRES (J. Pelletier, unpublished data). Lysates prepared from rabbit reticulocytes do not faithfully execute cap-dependent translation and hence could not be used in our particular assay (24). Both HeLa and Kreb cell extracts can initiate in a cap-dependent fashion, as well as internally on the HCV IRES. In our small molecule screen we used extracts prepared from Krebs cells as these could be easily prepared in large quantities.
In vitro transcription from plasmid DNA template using SP6 RNA polymerase was used to generate (CAG)33/FF/HCV/Ren·pA51 mRNA for in vitro translation experiments. The quality of the RNA template was assessed by fractionation on a denaturing formaldehyde/agarose gel and indicated that the mRNA was of the expected 4 kb size (Fig. 2A). To establish the range of mRNA concentration for translation in Krebs extracts, we varied the (CAG)33/FF/HCV/Ren·pA51 mRNA concentration in Krebs extracts from 0.5 to 50 ng/μl (Fig. 2B). The results indicate that for both firefly and renilla luciferase expression, this concentration range was within the linear range. We chose to perform in vitro translations at a final mRNA concentration of 4 ng/μl in order to maintain a good signal-to-noise ratio while minimizing consumption of the mRNA template. To assess the sensitivity of Krebs extracts to DMSO, we titrated DMSO in extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA (Fig. 2C). The results indicate that the extracts can tolerate DMSO concentrations 1%, with a minimal effect of protein synthesis activity. Finally, to ensure that the Krebs extracts were expressing firefly luciferase in a cap-dependent fashion, we performed translations in extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA and containing either 500 μM GDP (Fig. 2D, lane 2) or 500 μM m7GDP (Fig. 2D, lane 3). Since addition of m7GDP, but not GDP, to the reaction significantly altered the firefly to renilla expression ratio (compare lane 3 with lanes 1 and 2), these results indicate that translation of firefly luciferase in Krebs extracts is cap dependent.
Figure 2. Evaluation of the assay format for screening inhibitors of eukaryotic protein synthesis. (A) Formaldehyde/agarose gel analysis of mRNA generated by in vitro transcription of pSP/(CAG)33/FF/HCV/Ren·pA51 linearized with BamHI. In vitro transcriptions were performed as described previously (17). RNA was fractionated on a 1.2% agarose/formaldehyde gel and visualized by staining with SYBR gold (Molecular Probes). The RNA markers are from Gibco BRL and the arrow denotes the position of migration of full-length RNA. (B) Titration of (CAG)33/FF/HCV/Ren·pA51 mRNA into Krebs translation extracts. Both axes are drawn to log scale. Experiments were performed in duplicate and the error of the mean is too small to be visualized. (C) Sensitivity of Krebs extract to DMSO utilizing (CAG)33/FF/HCV/Ren·pA51 mRNA (5 ng/ml). Experiments were performed in duplicate and the error of the mean is shown. (D) Translation in Krebs extracts is cap dependent. Translations were performed in Krebs extracts with (CAG)33/FF/HCV/Ren·pA51 mRNA (5 ng/ml) alone (lane 1), in the presence of 500 μM GDP (lane 2) or with 500 μM m7GDP (lane 3). The ratio of renilla to firefly luciferase is plotted on the ordinate. Experiments were performed in duplicate and the error of the mean is shown. The absolute RLU obtained in the control experiments (lane 1) was 249 022 and 451 008 (for luciferase) and 1 259 448 and 2 630 247 (for renilla).
Using the above assay conditions, in vitro translations were performed in control 96-well plates (with no compounds). The data indicated a well-to-well variation of 4.4% CV, a signal-to-background ratio of 50:1 and a Z' factor of 0.8 (data not shown) (25). These parameters suggest the suitability of this assay for use in a miniaturized format. To validate that this screen could identify translation inhibitors, we chose a set of known inhibitors representing examples of different activity spectra (Supplementary Material, Fig. S1). The behavior of many of these compounds in Krebs extracts has not been previously documented. Anisomycin and bruceantin are inhibitors of peptidyl transferase (26,27), didemnin B binds to ribosome/eEF-1 complexes and blocks eEF-2 binding to pre-translocative ribosomes (28,29), bouvardin inhibits eEF1-dependent binding of aminoacyl-tRNA and eEF2-dependent translocation (30) and verrucarin A inhibits elongation, but is only effective on reinitiating ribosomes (31). The mechanism of action of baccharinol has not been reported, but it is a trichothecin epoxide that likely inhibits elongation. Compounds were titrated into 96-well plates, reformatted, added to translation reactions and tested at assay concentrations ranging from 0.65 nM to 10 μM. The IC50 values of the compounds ranged from 4 (bouvardin) to 300 nM (verrucarin) and demonstrated that different protein synthesis inhibitors with various potencies are readily detected in our assay.
Discrete compound screen
To assess the behavior of our assay in a high throughput screen format, we screened 50 96-well plates of our discrete small molecule collection (Supplementary Material, Fig. S2). The distribution of the firefly and renilla values obtained from the tested compound population indicated a good performance of this assay in the high throughput screen format. Using a cut-off value of 3-fold inhibition of translation (RLU < 0.33), 0.75% of compounds were positive for inhibition of firefly activity and 1.1% of compounds showed inhibition of renilla activity. Of these, five (0.125%) inhibited both firefly and renilla activity.
Based on these results, our discrete compound collection was screened once in 96-well format. The set of small molecule ligands still positive after the selection criteria denoted in Figure 1C are listed in Table 1. Of the 36 compounds identified in our screen (0.05% hit rate), 14 of these were known or structurally related to known protein synthesis inhibitors. These compounds provide an internal validation of the ability of our screen to identify inhibitors of protein synthesis. Since the mechanisms of action of these compounds are well characterized, they were not further studied. As many of the compounds we identified were nucleic acid-binding ligands or intercalators (e.g. acriflavine, ethidium bromide, dequalinium, Hoechst 33258, etc.), we established an assay to quickly determine if a given compound was a nucleic acid-binding ligand. This assay consists of incubating supercoiled DNA with 50 μM compound and monitoring alterations in electrophoretic mobility on agarose gels (Fig. 3A). DNA binding results in loss of negative supercoils producing a decrease in mobility when analyzed by agarose gel electrophoresis (Fig. 3A, compare lane 3 with lane 2). Based on this assay, we conclude that NSC 305831 (lane 4), NSC 22907 (lane 5), NSC 130813 (lane 6), NSC 254681 (lane 7), NSC 101327 (lane 10) and NSC 85701 (lane 12) bind to DNA. In this assay, NSC 119889 (lane 8), suramine (lane 9), NSC 111041 (lane 11) and NSC 115183 (lane 13) did not affect the mobility of supercoiled DNA and for the purposes of this study are not classified as general DNA-binding compounds (Table 1).
Table 1. Small molecule ligands identified in the high throughput screen
Figure 3. (A) Intercalation assay assessing the ability of compounds to alter the mobility of supercoiled plasmid DNA. Supercoiled plasmid DNA was incubated with 50 μM ligand at room temperature for 20 min and electrophoresed into a 0.8% agarose gel. The gel was stained with ethidium bromide (0.5 μg/ml) and visualized using short wavelength UV light. The identity of the compounds incubated with supercoiled DNA is indicated above the panel. The position of migration of supercoiled DNA is indicated by an arrow. (B–D) Relative production of firefly and renilla by translation of (CAG)33/FF/HCV/Ren·pA51 mRNA in the presence of small molecule ligands. Following in vitro translations, methionine-labeled samples were treated in SDS sample buffer and electrophoresed into a 10% SDS–polyacrylamide gel. The gels were treated with EN3Hance, dried and exposed to X-Omat (Kodak) film. The concentration of compound used in the translation mix is indicated above the panel and the identity of the compound indicated below the panel. The position of migration of firefly and Renilla luciferase protein are indicated.
Since the initial high throughput translation assay was set up to identify compounds that could inhibit one of several steps, we undertook a series of experiments to deconvolute their mode of action. Addition of compound to Krebs extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA, followed by SDS–PAGE analysis of methionine products, indicated that the small molecule ligand could be classified into one of three categories, depending on which ORF they preferentially inhibited. (i) At 50 μM, acriflavine (Fig. 3B) and ethidium bromide (A. Malina and J. Pelletier, data not shown) inhibited expression of both firefly and renilla luciferase, however, at 10 μM, translation of renilla luciferase was significantly inhibited (>10-fold) whereas that of firefly luciferase was affected 2-fold (Fig. 3B). (ii) At 50 μM, NSC 119889 completely inhibited firefly luciferase translation, but had little effect on renilla expression (Fig. 3C). (iii) The remaining set of compounds were general inhibitors of translation that equally inhibited synthesis of firefly and renilla luciferase (Fig. 3D). These included NSC 22907, NSC 305831, NSC 130813, NSC 254681, NSC 101327, NSC 85701, NSC 111041 and NSC 115183 (Fig. 3D), as well as quinacrine, Hoechst 33258, mitoxanthrone, chartreusin, nogalamycin, quinolinium and ellipticine derivatives (data not shown). Given the large number of intercalators identified in this study, detailed characterization of their modes of action will be presented elsewhere (A. Malina and J. Pelletier, manuscript in preparation) and for the current study we will focus on the four non-intercalating compounds, NSC 119889, NSC 111041, suramine and NSC 115183 (Fig. 4A).
Figure 4. (A) Structures of NSC 119889, NSC 111041 and suramine. The structure of NSC 115183 is currently not available. (B) Titration of compounds in Krebs extracts. Translations were performed in the presence of the indicated amounts of compound and at final mRNA and K+ concentrations of 5 μg/ml and 100 mM, respectively. Control translation reactions contained 0.5% DMSO, which is equivalent to what was present in reactions containing compound. Luciferase values are normalized to the activity obtained in the control translations (which were set to 1). The relative firefly luciferase values are represented by white bars and the relative renilla values are represented by black bars. Translations were performed three times and the average values are presented along with the error of the mean. (C) Titration of compounds in rabbit reticulocyte lysate. The relative firefly luciferase values are represented by white bars and the relative renilla values are represented by black bars. Translations were performed twice and the average values are presented along with the error of the mean. (D) Effect of compounds (50 μM) on translation in E.coli S30 extracts. Protein synthesis was monitored by assessing the amount of methionine incorporation into TCA-precipitable material by endogenous mRNA. Translations were performed twice and the average values are presented along with the error of the mean.
Titration of these compounds into Krebs translation extracts programmed with (CAG)33/FF/HCV/Ren·pA51 indicated IC50 values of 10, 1 and 0.4 μM for inhibition of both firefly and renilla translation by suramine, NSC 111041 and NSC 115183, respectively. The IC50 values for inhibition of firefly and renilla translation were 10 and 50 μM, respectively, for NSC 119889, indicating a differential effect on expression of firefly and renilla cistrons by this compound (Fig. 4B). The inhibitory activity of these compounds was also assessed in rabbit reticulocyte lysates programmed with (CAG)33/FF/HCV/Ren·pA51 (Fig. 4C) and showed the same relative potency as in Krebs extracts, with suramine being the least active (IC50 50 μM) and NSC 115183 being the most potent compound (IC50 0.2 μM) (Fig. 4C). NSC 119889 maintained the same differential effect on firefly and renilla expression as observed in Krebs extracts (Fig. 4C). To assess if any of these compounds exerted an effect on prokaryotic protein synthesis, they were tested in an E.coli S30 extract. Suramine, NSC 111041 and NSC 115183 did not significantly affect translation when tested to a final concentration of 50 μM, whereas NSC 119889 inhibited translation when present at a final concentration of 50 μM (Fig. 4D).
A compound that inhibits initiation of translation will prevent formation of 80S ribosome–mRNA complexes, whereas inhibitors of elongation or termination will not affect loading of ribosomes onto mRNA templates. Ribosome binding experiments were performed in the presence of cycloheximide (an inhibitor of elongation) and 50 μM compound. Ribosome–mRNA complexes were fractionated on sucrose gradients and the radioactivity in the individual fractions determined by scintillation counting (Fig. 5). Suramine, NSC 111041 and NSC 115183 did not affect the efficiency of 80S ribosome–mRNA complex formation (Fig. 5A and B). These experiments indicate that suramine, NSC 111041 and NSC 115183 exert their effects by inhibiting downstream of initiation. On the other hand, NSC 119889 prevented mRNA–ribosome interaction, indicating that this compound inhibits initiation of protein synthesis (Fig. 5C). A non-hydrolyzable analog of GTP, GMP-PNP, can be used to trap 43S initiation complexes at initiation codons of mRNA templates since it inhibits the release of assembled initiation factors from the small subunit, preventing 60S subunit joining (Fig. 5D). When NSC 119889 was present in the binding reactions with GMP-PNP, 43S complex binding was inhibited, indicating that this compound inhibits ribosome binding to mRNA templates (Fig. 5D).
Figure 5. Effect of compounds on initiation complex assembly in rabbit reticulocyte lysates. 32P-labeled CAT mRNA was incubated in rabbit reticulocyte lysates and initiation complexes resolved by centrifugation on sucrose gradients. Fractions from each sucrose gradient were collected using a Brandel Tube Piercer connected to an ISCO fraction collection and were individually counted. (A) Complex formation in the presence of 600 μM cycloheximide, 600 μM cycloheximide and 50 μM NSC 111041 or 600 μM cycloheximide and 50 μM suramine. The total counts recovered and the percentage mRNA bound to 80S complexes were: CAT mRNA/cycloheximide (320 856 c.p.m., 6% binding); CAT mRNA/cycloheximide + NSC 111041 (304 285c.p.m., 6% binding); CAT mRNA/cycloheximide + suramine (315 307 c.p.m., 3% binding). (B) Complex formation in the presence of 600 μM cycloheximide and 50 μM NSC 115183. The total counts recovered and the percentage mRNA bound to 80S complexes were: CAT mRNA/cycloheximide (109 065 c.p.m., 10% binding); CAT mRNA/cycloheximide + NSC 115183 (112 299 c.p.m., 9% binding). (C) Complex formation in the presence of 600 μM cycloheximide and 50 μM NSC 119889. The total counts recovered and the percentage mRNA bound to 80S complexes were: CAT mRNA/cycloheximide (182 569 c.p.m., 7% binding); CAT mRNA/cycloheximide + NSC 119889 (286 708 c.p.m., 0% binding). (D) Complex formation in the presence of 1 mM 5'-guanylylimidodiphosphate (GMP-PNP) or 1 mM GMP-PNP and 50 μM NSC 119889. Total counts recovered from each gradient and the percentage mRNA bound to 43S complexes were: CAT mRNA/GMP-PNP (98 359 c.p.m., 7% binding); CAT mRNA/GMP-PNP + NSC 119889 (95 225 c.p.m., 0% binding).
To better gauge the inhibitory spectrum of NSC 119889, its effects on expression of several IRESs was assessed (Fig. 6). The IRESs tested were derived from EMC, poliovirus and HCV (Fig. 6A). As previously indicated, translation initiation from the HCV IRES was particularly resistant to inhibition by NSC 119889 , whereas cap-dependent translation was completely inhibited when NSC 119889 was present in the translation reaction at a final concentration of 50 μM. The effect of NSC 119889 on expression of firefly and renilla ORFs from Krebs extracts programmed with FF/EMC/Ren or Ren/P2/FF mRNAs was similar in that expression of both ORFs appeared to be equally inhibited from the two templates. These results indicate that unlike with the HCV IRES, initiation from both EMC and poliovirus IRESs is sensitive to inhibition by NSC 119889 (Fig. 6B and C). Translation of both firefly and renilla ORFs from FF/Ren mRNA was equally affected by NSC 119889, consistent with expression of both ORFs from this mRNA template being cap dependent (Fig. 6B and C). In this particular case, expression of renilla luciferase is expected to arise as a consequence of reinitiation by ribosomes that have completed translation of the firefly ORF. These results are consistent with different initiation mechanisms for ribosome binding between HCV and picornaviral IRESs.
Figure 6. (A) Schematic representation of constructs used to assess the inhibitory potential of NSC 119889 on different IRESs. The names of the plasmids harboring the inserts is provided to the left and the restriction sites used to linearize the plasmids denoted to the right. Note that mRNA derived from pKS/FF/Ren, pSP(CAG)33/FF/HCV/Ren·pA51 and pKS/FF/EMC/Ren will contain a poly(A) tail of 51 adenylate residues. (B) Relative translational efficiencies obtained in Krebs extracts programmed with FF/Ren, (CAG)33/FF/HCV/Ren, FF/EMC/Ren and Ren/P2/FF mRNA in the presence of NSC 119889. Control translation reactions contained 0.5% DMSO. Luciferase values were normalized to the activity obtained in the control translations (which were set at one). The relative firefly luciferase values are represented by white bars and the relative renilla values are represented by black bars. The bar represents the average of two experiments and the error of the mean is denoted. (C) A representative experiment demonstrating the relative translation efficiencies of firefly and renilla production from the different test mRNAs in the presence of NSC 119889. Following in vitro translations in Krebs extracts in the presence of methionine, the indicated amounts of NSC 119889 and 5 ng/ul mRNA, samples were treated in SDS sample buffer and electrophoresed into 10% SDS–polyacrylamide gels. Gels were treated with EN3Hance, dried and exposed to X-Omat (Kodak) film. The concentration of NSC 119889 in the translation reaction is indicated above the panel and the identity of the mRNA used to program the translation reaction is indicated below the panel. The position of migration of firefly and renilla luciferase proteins is denoted. The position of molecular mass markers (NEB) is indicated to the left.
NSC 119889 contains a xanthenyl ring structure relating it to fluorescein or gallein. A small number of structurally related compounds were obtained to identify the functionally important groups of NSC 119889 for translation inhibition. The structure–activity study was performed at a final concentration of 50 μM in Krebs extracts programmed with (CAG)33/FF/HCV/Ren (Fig. 7A). Fluorescein and related compounds (NSC 119892, NSC 122391, NSC 119894, NSC 157411, NSC 122390, NSC 119887 and NSC 119888) were ineffective at inhibiting firefly luciferase (Fig. 7A and B), whereas NSC 119889 and NSC 378139 were active. With the exception of NSC 622475, the addition of two hydroxyl groups to the xanthenyl ring system at positions 4' and 5' yielded compounds (NSC 119915, NSC 119911, gallein, NSC 119893, NSC 119913 and NSC 119910) that inhibited firefly luciferase production (Fig. 7A and B). Hydroxyl group substitutions at the 2' and 7' position yielded compounds (NSC 2608 and NSC 9037) that were inactive as inhibitors of protein synthesis. Although far from an exhaustive study, these results indicate that a 4',5'-hydroxyl substituted xanthenyl ring structure is important for activity; most compounds lacking these hydroxyl groups are inactive. However, exceptions such as NSC 119889 and NSC 378139 indicate that a second structural component of the small molecule (contributed by the aromatic ring) is also important for activity.
Figure 7. Structure–activity relationships of NSC 119889 analogs. Translations were performed as indicated in Materials and Methods. (A) Effect of 50 μM compound on translation in Krebs extracts programmed with (CAG)33/FF/HCV/Ren mRNA. Luciferase values were normalized to the activity obtained in the control translations (which were set to 1). Experiments were performed three times with each compound and the error of the mean is indicated. The chemical structures are shown below the bar graph. (B) A representative experiment demonstrating the relative translation efficiencies of firefly and renilla production from (CAG)33/FF/HCV/Ren mRNA in the presence of the different NSC analogs. Following in vitro translation in Krebs extracts in the presence of methionine, 50 μM compound, and 5 ng/μl mRNA, samples were treated in SDS sample buffer and electrophoresed into 10% SDS–polyacrylamide gels. Gels were treated with EN3Hance, dried and exposed to X-Omat (Kodak) film. The identity of the compound used in the translation reaction is denoted above the panel. The position of migration of firefly and renilla luciferase proteins is denoted. The position of molecular mass markers (NEB) is indicated to the left.
DISCUSSION
Monitoring translation of a bicistronic mRNA reporter in a eukaryotic translation extract was utilized to identify novel protein synthesis inhibitors. This system was adapted to a miniaturized format to save time, cost and reagents. The initial screen was set up to identify general inhibitors of eukaryotic protein synthesis, as well as small molecule ligands that could bind to the CAG repeats engineered upstream of the firefly ORF, or that could bind to the HCV IRES. The detection of known protein synthesis inhibitors from chemical libraries by our screen provided an internal validation (Table 1). Given the importance of RNA helices to ribosome function and the presence of double-stranded regions on mRNA templates, it is not surprising that we identified a large number of intercalators by our assay (Table 1). Indeed, several previous reports have documented the ability of nucleic acid-binding agents to inhibit protein synthesis. Ethidium bromide (32) and acriflavine (33) have been reported to inhibit the aminoacylation of tRNAs in vitro, presumably by intercalating into tRNA and preventing its recognition by aminoacyl transferases. The intercalators 3-nitrobenzothiazolo{3,2-a}quinolinium and fagaronine have been shown to inhibit protein synthesis in cell-free extracts, possibly by affecting elongation (34). Detailed characterization of the mode of action of nucleic acid-binding ligands on translation will be presented elsewhere (A. Malina and J. Pelletier, manuscript in preparation).
Whereas general inhibitors of translation affecting elongation and initiation were identified, our screen did not identify ligands that could bind to the CAG repeats and inhibit translation. This may indicate the need to screen a larger, more diverse collection of compounds or alternatively may indicate that ligands binding to the CAG repeat were unable to provide a kinetic barrier to inhibit initiation. If the latter case was true, then an approach incorporating the use of inducers of dimerization to recruit protein surfaces to the RNA target may improve the inhibitory potential of RNA-binding ligands (17). This would necessitate re-formating our screen to identify such compounds. Our screen did, however, identify two compounds, ethidium bromide and acriflavine, which showed preferential inhibition of HCV-driven renilla expression when present at 10 μM in the translation reaction (Fig. 3B). Detailed characterization of these compounds indicates that at 10 μM they block ribosome–HCV IRES interaction (A. Malina and J. Pelletier, manuscript in preparation).
Our screen identified four protein synthesis inhibitors that did not bind double-stranded DNA when tested in vitro (Fig. 3A). Three of these compounds, suramine, NSC 111041 and NSC 115183, do not inhibit initiation complex formation (Fig. 5). We have tested the potential of these compounds to cause misreading and did not obtain any evidence that these compounds affected ribosome translation fidelity (A.-S. Guenier and J. Pelletier, data not shown). Suramine is a polysulfonated napthylurea to which has been associated a large number of biological activities, ranging from trypanocidal to anticancer activity (for a review see 35). The large number of negative charges associated with this molecule may suggest that it inhibits translation in a non-specific fashion by binding factors with cationic charges, although we have no direct evidence for this mode of action. Inhibition of protein synthesis by suramine may be the basis for some of the toxic side-effects associated with the use of this compound in the clinic (35). The synthesis of NSC 111041 has been reported (36) but we know of no studies documenting an effect on protein synthesis by this compound. Unfortunately, no structure for NSC 115183 is currently available from the NCI-DTP. Our results suggest that these compounds inhibit either elongation or termination of translation.
Suramine, NSC 111041 and NSC 115183 had no effect when tested in a prokaryotic translation system, whereas NSC 119889 significantly reduced protein synthesis when present at 50 μM (Fig. 4D). These results indicate a strong discrimination on the part of suramine, NSC 111041 and NSC 115183 for the eukaryotic translation apparatus. This may be due to the absence of the compound target in prokaryotes or reflect structural differences in the ligand-binding site between eukaryotes and prokaryotes. Alternatively, there is a marked difference in the divalent cation optima between prokaryotic and eukaryotic extracts. Prokaryotic systems have a higher in vitro Mg2+ optimum (10–15 mM) (37), whereas the eukaryotic systems we used have a lower Mg2+ optimum (2.0–3.0 mM for wheatgerm and rabbit reticulocyte) (38,39), which in turn could influence binding of ligand to their targets or the conformation of the binding site. Of course, we cannot exclude non-specific effects such as the possibility that macromolecules in the prokaryotic extract bind to and titrate out the ligand, hence preventing inhibition by the compound. Differences in ion concentration or protein target concentration between Krebs and reticulocyte lysates may also be responsible for differences in IC50 values noted between these extracts for suramine, NSC 111041, NSC 115183 and NSC 119889 (Fig. 4B and C).
NSC 119889 inhibits initiation of protein synthesis, as it prevented the formation of 48S and 80S initiation complexes (Fig. 5). The ability of the HCV IRES to direct protein synthesis in the presence of NSC 119889 indicates that this compound does not exert additional effects on elongation or termination (Fig. 6). NSC 119889 is not acting through the CAG repeats present within the 5'-UTR of (CAG)33/FF/HCV/Ren since it was also active on reporter constructs lacking these repeats (Fig. 6). Additionally, NSC 119889 inhibited expression of cistrons whose initiation was dependent on the EMC or poliovirus IRES, but had no effect on translation initiation mediated by the HCV IRES (Fig. 6). These results provide insight into the mode of action of NSC 119889 as it has been documented that the canonical factor requirement for ribosome binding to the EMC and poliovirus IRESs differ from those of the HCV IRES (for a review see 40). The HCV IRES can directly recruit the 43S particle and makes intimate contact with the 40S ribosomal subunit and eIF3 (41). Aside from eIF2/GTP/Met-tRNAi and eIF3, no other initiation factors appear necessary for high affinity binding of the HCV IRES to 43S ribosomes (40,41). In contrast, internal initiation on the picornaviral EMC and poliovirus IRESs requires IRES transacting factors (ITAFs), as well as a larger complement of canonical initiation factors, including eIF4F or eIF4A and the central third of eIF4G to which eIF4F binds (reviewed in 40). The observed inhibition on EMCV and poliovirus IRES-mediated translation, but not on HCV IRES-mediated initiation, suggests that NSC 119889 may be targeting an ITAF or an initiation factor other than eIF3 or eIF2.
NSC 119889 was previously identified in a screen for antagonists of the interaction between HIV nucleocapsid protein NC-p7 and nucleic acids and exerted significant anti-HIV activity in vivo when present in the culture medium of HIV-infected cells (42). Characterization of NSC 119889 in this assay revealed that it bound to NC-p7 with a Kd of 2.5 x 10–7 M. It is thought that NSC 119889 interferes with NC-p7 function during the reverse transcription process (42). Consistent with our results indicating that NSC 119889 does not interact with DNA (Fig. 3A), BIAcore studies failed to elucidate an interaction between a related molecule, tetrachlorogallein, and DNA (42).
Additionally, fluorescein analogs are active against a variety of enzymes that use ATP or NADH as cofactors (43–47). It is thought that the fluorescein analogs can reversibly compete for ATP- or NADH-binding sites. Specifically, eosin can inhibit the mitochondrial ADP/ATP carrier (43) and the erythrocyte calcium pump (44) and erythrosine inhibits lactate dehydrogenase (45), aspartate transcarbamylase (46) and creatine kinase (47). These observations suggest a mode of action for NSC 119889 (and structural analogs) in translation, possibly by interfering with the activity of a factor that utilizes ATP (or GTP). Clearly, NSC 119889 is not a promiscuous inhibitor of GTP-binding proteins since elongation of protein synthesis remains unaffected when initiation is directed by the HCV IRES (Fig. 3).
In sum, we have established a high throughput screen that has allowed the identification of a large number of protein synthesis inhibitors. We found that a large class of these were nucleic acid-binding ligands. The characterization of four non-intercalating compounds indicated that we isolated inhibitors of both initiation and elongation. The characterization of NSC 119889 as having a differential effect on initiation driven from picornaviral versus HCV IRESs suggests that it might be used to better understand mechanisms of internal initiation.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the technical assistance of Mr Virabhadrachari Vipparti and the kind help of Abba Malina in screening the NCI diversity set in the initial stages of this work. We are immensely grateful to the NIH/NCI Developmental Therapeutics Program for their generous supply of the discrete chemical compound library as well as for re-supply of positive hits for confirmation. J.P. is a Canadian Institutes of Health Research (CIHR) Senior Investigator. This work was supported by grants from the Hereditary Disease Foundation (CHDI) to J.P., the National Cancer Institute of Canada (nos 011040 and 014313) to J.P., funding from the Jean-Louis Lévesque Foundation and a VRQ (Valorisation Recherche Quebec) grant (2201-147).
REFERENCES
Pestka,S. (1977) Inhibitors of Protein Synthesis. Academic Press, New York, NY.
Vazquez,D. (1979) Inhibitors of protein biosynthesis. Mol. Biol. Biochem. Biophys., 30, 1–312.
Watkins,S.J. and Norbury,C.J. (2002) Translation initiation and its deregulation during tumorigenesis. Br. J. Cancer, 86, 1023–1027.
Graff,J.R. and Zimmer,S.G. (2003) Translational control and metastatic progression: enhanced activity of the mRNA cap-binding protein eIF-4E selectively enhances translation of metastasis-related mRNAs. Clin. Exp. Metastasis, 20, 265–273.
Hershey,J.W. and Miyamoto,S. (2000) Translation Control and Cancer. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Ruggero,D. and Pandolfi,P.P. (2003) Does the ribosome translate cancer? Nature Rev. Cancer, 3, 179–192.
Meric,F. and Hunt,K.K. (2002) Translation initiation in cancer: a novel target for therapy. Mol. Cancer Ther., 1, 971–979.
Brodersen,D.E., Clemons,W.M.,Jr, Carter,A.P., Morgan-Warren,R.J., Wimberly,B.T. and Ramakrishnan,V. (2000) The structural basis for the action of the antibiotics tetracycline, pactamycin and hygromycin B on the 30S ribosomal subunit. Cell, 103, 1143–1154.
Carter,A.P., Clemons,W.M., Brodersen,D.E., Morgan-Warren,R.J., Wimberly,B.T. and Ramakrishnan,V. (2000) Functional insights from the structure of the 30S ribosomal subunit and its interactions with antibiotics. Nature, 407, 340–348.
Pioletti,M., Schlunzen,F., Harms,J., Zarivach,R., Gluhmann,M., Avila,H., Bashan,A., Bartels,H., Auerbach,T., Jacobi,C. et al. (2001) Crystal structures of complexes of the small ribosomal subunit with tetracycline, edeine and IF3. EMBO J., 20, 1829–1839.
Hansen,J.L., Ippolito,J.A., Ban,N., Nissen,P., Moore,P.B. and Steitz,T.A (2002) The structures of four macrolide antibiotics bound to the large ribosomal subunit. Mol. Cell, 10, 117–128.
Mandal,M., Boese,B., Barrick,J.E., Winkler,W.C. and Breaker,R.R. (2003) Riboswitches control fundamental biochemical pathways in Bacillus subtilis and other bacteria. Cell, 113, 577–586.
Mironov,A.S., Gusarov,I., Rafikov,R., Lopez,L.E., Shatalin,K., Kreneva,R.A., Perumov,D.A. and Nudler,E. (2002) Sensing small molecules by nascent RNA: a mechanism to control transcription in bacteria. Cell, 111, 747–756.
Nahvi,A., Sudarsan,N., Ebert,M.S., Zou,X., Brown,K.L. and Breaker,R.R. (2002) Genetic control by a metabolite binding mRNA. Chem. Biol., 9, 1043–1049.
Werstuck,G. and Green,M.R. (1998) Controlling gene expression in living cells through small molecule–RNA interactions. Science, 282, 296–298.
Harvey,I., Garneau,P. and Pelletier,J. (2002) Inhibition of translation by RNA–small molecule interactions. RNA, 8, 452–463.
Harvey,I., Garneau,P. and Pelletier,J. (2002) Forced engagement of a RNA/protein complex by a chemical inducer of dimerization to modulate gene expression. Proc. Natl Acad. Sci. USA, 99, 1882–1887.
Gold,L., Polisky,B., Uhlenbeck,O. and Yarus,M. (1995) Diversity of oligonucleotide functions. Annu. Rev. Biochem., 64, 763–797.
Sambrook,J. and Russell,D.W. (2001) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Svitkin,Y.V. and Agol,V.I. (1978) Complete translation of encephalomyocarditis virus RNA and faithful cleavage of virus-specific proteins in a cell-free system from Krebs-2 cells. FEBS Lett., 87, 7–11.
Dyer,B.W., Ferrer,F.A., Klinedinst,D.K. and Rodriguez,R. (2000) A noncommercial dual luciferase enzyme assay system for reporter gene analysis. Anal. Biochem., 282, 158–161.
Tsukiyama-Kohara,K., Iizuka,N., Kohara,M. and Nomoto,A. (1992) Internal ribosome entry site within hepatitis C virus RNA. J. Virol., 66, 1476–1483.
Pestova,T.V., Shatsky,I.N., Fletcher,S.P., Jackson,R.J. and Hellen,C.U. (1998) A prokaryotic-like mode of cytoplasmic eukaryotic ribosome binding to the initiation codon during internal translation initiation of hepatitis C and classical swine fever virus RNAs. Genes Dev., 12, 67–83.
Svitkin,Y.V., Ovchinnikov,L.P., Dreyfuss,G. and Sonenberg,N. (1996) General RNA binding proteins render translation cap dependent. EMBO J., 15, 7147–7155.
Zhang,J.-H., Chung,T.D.Y. et al. (1999) A simple statistical parameter for use in evaluation and validation of high throughput screening assays. J. Biomol. Screening, 4, 67–73.
Carrasco,L. and Vazquez,D. (1972) Survey of inhibitors in different steps of protein synthesis by mammalian ribosomes. J. Antibiot., 25, 732–737.
Fresno,M., Gonzales,A., Vazquez,D. and Jimenez,A. (1978) Bruceantin, a novel inhibitor of peptide bond formation. Biochim. Biophys. Acta, 518, 104–112.
SirDeshpande,B.V. and Toogood,P.L. (1995) Mechanism of protein synthesis inhibition by didemnin B in vitro. Biochemistry, 34, 9177–9184.
Ahuja,D., Vera,M.D., SirDeshpande,B.V., Morimoto,H., Williams,P.G., Joullie,M.M. and Toogood,P.L. (2000) Inhibition of protein synthesis by didemnin B: how EF-1alpha mediates inhibition of translocation. Biochemistry, 39, 4339–4346.
Zalacain,M., Zaera,E., Vazquez,D. and Jimenez,A. (1982) The mode of action of the antitumor drug bouvardin, an inhibitor of protein synthesis in eukaryotic cells. FEBS Lett., 148, 95–97.
Cannon,M., Jimenez,A. and Vazquez,D. (1976) Competition between trichodermin and several other sesquiterpene antibiotics for binding to their receptor site(s) on eukaryotic ribosomes. Biochem. J., 160, 137–145.
Landez,J.H., Roskoski,R.J.,Jr and Coppoc,G.L. (1969) Ethidium bromide and chloroquine inhibition of rat liver cell-free aminoacylation. Biochim. Biophys. Acta, 195, 276–279.
Birkmayer,G.D. and Balda,B.R. (1970) Proflavine inhibition of protein synthesis in malignant hamster melanom. FEBS Lett., 11, 221–223.
Torres,C.A.C. and Baez,A. (1986) Effects of the antitumor drugs 3-nitrobenzothiazolo{3,2-a}quinolinium and fagaronine on nucleic acid and protein synthesis. Biochem. Pharmacol., 35, 679–685.
Kaur,M., Reed,E., Sartor,O., Dahut,W. and Figg,W.D. (2002) Suramin’s development: what did we learn? Invest. New Drugs, 20, 209–219.
Pettit,G., Fleming,W. and Paull,K.D. (1968) Synthesis of the 6- and 7-hydrozy-5,8-dioxocarbostrils. J. Org. Chem., 33, 1089–1092.
Chen,H.-Z. and Zubay,G. (1983) Prokaryotic coupled transcription–translation. Methods Enzymol., 101, 674–690.
Anderson,C.W., Straus,J.W. and Dudock,B.S. (1983) Preparation of a cell-free protein-synthesizing system from wheat germ. Methods Enzymol., 101, 635–644.
Merrick,W.C. (1983) Translation of exogenous mRNAs in reticulocyte lysates. Methods Enzymol., 101, 606–615.
Pestova,T.V., Kolupaeva,V.G., Lomakin,I.B., Pilipenko,E.V., Shatsky,I.N., Agol,V.I. and Hellen,C.U. (2001) Molecular mechanisms of translation in eukaryotes. Proc. Natl Acad. Sci. USA, 98, 7029–7036.
Kieft,J.S., Zhou,K., Jubin,R. and Doudna,J.A. (2001) Mechanism of ribosome recruitment by hepatitis C IRES RNA. RNA, 7, 194–206.
Stephen,A.G., Worthy,K.M., Towler,E., Mikovits,J.A., Sei,S., Roberts,P., Yang,Q.E., Akee,R.K., Klausmeyer,P., McCloud,T.G. et al. (2002) Identification of HIV-1 nucleocapsid protein: nucleic acid antagonists with cellular anti-HIV activity. Biochem. Biophys. Res. Commun., 296, 1228–1237.
Majima,E., Yamaguchi,N., Chuman,H., Shinohara,Y., Ishida,M., Goto,S. and Terada,H. (1998) Binding of the fluorescein derivative eosin Y to the mitochondrial ADP/ATP carrier: characterization of the adenine nucleotide binding site. Biochemistry, 37, 424–432.
Gatto,C. and Milanick,M.A. (1993) Inhibition of the red blood cell calcium pump by eosin and other fluorescein analogues. Am. J. Physiol., 264, C1577–C1586.
Wassarman,P.M. and Lentz,P.J.,Jr (1971) The interaction of tetra-iodofluorescein with dogfish muscle lactate dehydrogenase: a chemical and X-ray crystallographic study. J. Mol. Biol., 60, 509–522.
Jacobsverg,L.B., Kantrowitz,E.R. and Lipscomb,W.N. (1975) Interaction of tetraiodofluorescein with aspartate transcarbamylase and its isolated catalytic and regulatory subunits. J. Biol. Chem., 24, 9238–9249.
Somerville,L.L. and Quiocho,F.A. (1977) The interaction of tetraiodofluorescein with creatine kinase. Biochim. Biophys. Acta, 481, 493–499.
Zuker,M. (2003) Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res., 31, 3406–3415.(Olivia Novac1,2, Anne-Sophie Guenier1,2 )
*To whom correspondence should be addressed at: McIntyre Medical Sciences Building, Room 810, 3655 Promenade Sir William Osler, McGill University, Montreal, Quebec H3G 1Y6, Canada. Tel: +1 514 398 2323; Fax: +1 514 398 7384; Email: jerry.pelletier@mcgill.ca
ABSTRACT
The use of small molecule inhibitors of cellular processes is a powerful approach to understanding gene function that complements the genetic approach. We have designed a high throughput screen to identify new inhibitors of eukaryotic protein synthesis. We used a bicistronic mRNA reporter to multiplex our assay and simultaneously screen for inhibitors of cap-dependent initiation, internal initiation and translation elongation/termination. Functional screening of >90 000 compounds in an in vitro translation reaction identified 36 inhibitors, 14 of which are known inhibitors of translation and 18 of which are nucleic acid-binding ligands. Our results indicate that intercalators constitute a large class of protein synthesis inhibitors. Four non-intercalating compounds were identified, three of which block elongation and one of which inhibits initiation. The novel inhibitor of initiation affects 5' end-mediated initiation, as well as translation initiated from picornaviral IRESs, but does not significantly affect internal initiation from the hepatitis C virus 5'-untranslated region. This compound should be useful for delineating differences in mechanism of initiation among IRESs.
INTRODUCTION
Inhibition of prokaryotic protein synthesis by small molecule ligands has provided modern medicine with an important arsenal of anti-infective compounds and has helped dissect the pathway of protein synthesis. The cumulative knowledge gained from studying inhibitors of translation demonstrates the power of applying a chemical biology approach to study this process. Pharmacological inhibitors of protein synthesis have allowed the characterization of events leading to the assembly of active polysomes, trapping intermediates of the initiation and elongation cycle and providing insight into the molecular functions of protein factors (1,2).
Surprisingly, no report documents the systematic search for new inhibitors of eukaryotic protein synthesis, with the majority of compounds currently used having been identified several decades ago (1,2). The impetus for establishing new chemical screens stems not only from the usefulness of protein synthesis inhibitors as tools, but from the growing number of studies implicating deregulated translation as a contributor to cancer initiation and progression (3,4). For example, overexpression of some initiation factors can lead to malignant transformation, whereas down-regulation of these same factors can suppress the transformed phenotype (reviewed in 5). Components of the translation apparatus are overexpressed or mutated in cancers (5). The tumor suppressor gene product (pRB) directly impacts on the translation process by affecting the levels of ribosomes (for a review see 6). The PI3K/Akt signaling pathway is deregulated in a large number of cancers and modulates activity of factors regulating initiation of protein synthesis (7). As such, translation represents an interesting, under-explored target for chemotherapeutic intervention.
The majority of known protein synthesis inhibitors bind to ribosomes, with some of these making key contacts with the RNA component (8–11). These findings validate the idea of structured RNA as a biological target for small molecule interdiction. Additionally, recent studies have demonstrated that highly structured domains within prokaryotic mRNAs can serve as metabolite-responsive genetic switches (12). In the cases studied, the metabolite binds to an RNA motif and induces an allosteric change that alters transcription termination (13) or translation initiation (14). In eukaryotes, small molecule–RNA interactions (15,16) or small molecules acting as chemical inducers of dimerization (17) can inhibit protein synthesis by interfering with translation initiation. The potential for regulating gene expression with chemical ligands that modulate mRNA function has remained largely unexplored. Many features of RNA make it an attractive target: it is central to many functions of the cell, contains complex secondary and tertiary structural folds and lacks a cellular repair mechanism. The ability of small molecules to interact with RNA is well recognized (18) and these interactions have the potential to prevent or enhance gene expression, achieve allele-specific modulation of gene expression (when allelic sequence differences result in altered RNA conformations), achieve isoform-specific modulation of gene expression and exhibit allosteric effects.
We have undertaken a high throughput screen designed to identify general inhibitors of eukaryotic protein synthesis, as well as small molecule ligands that inhibit gene-specific expression at the level of mRNA–ribosome interaction. The use of a bicistronic mRNA reporter in an in vitro translation extract allowed us to multiplex the assay, thus reducing our reagent cost and consumption of small molecule ligand collection. This report demonstrates the validation and execution of a functional eukaryotic translation screen of >90 000 distinct compounds. Herein, we have identified several novel protein synthesis inhibitors and characterized their activity profiles.
MATERIALS AND METHODS
Materials and general methods
Restriction endonucleases, calf intestinal alkaline phosphatase, the Klenow fragment of DNA polymerase I, T4 DNA ligase and SP6 RNA polymerase were purchased from New England Biolabs (Beverly, MA). cytidine triphosphate (20.5 Ci/mmol) and methionine (>1000 Ci/mmol) were obtained from Perkin Elmer Life Sciences (Boston, MA). Preparation of plasmid DNA, restriction enzyme digestion, agarose gel electrophoresis of DNA and RNA, DNA ligation and bacterial transformations were carried out using standard methods (19).
Plasmid construction and cell-free extract preparation
The bicistronic vector, pSP/(CAG)33/FF/HCV/Ren·pA51, was generated as follows. A sequence encoding a poly(dA) tail was inserted immediately downstream of the renilla luciferase gene by inserting an oligonucleotide containing 51 dA residues d(CTAGGTTCAAGCTTCGGTGGA51GGGG) into the XbaI and BamHI sites of phRL-null (Promega). This plasmid was called phRL-null·pA51. The HCV 5'-untranslated region (5'-UTR) (nucleotides 14–383) was amplified by PCR using oligonucleotides A and B from an HCV internal ribosome entry site (IRES) bicistronic plasmid (kindly provided by Dr N. Sonenberg, McGill University), digested with EcoRI and XbaI and cloned into phRL-null·pA51 which had been linearized with EcoRI and NheI. This places the HCV ATG initiation codon (nucleotide 342) in-frame with the renilla initiation codon. The resulting vector was called HCV/RL·pA51. Plasmid pGL3-Basic (Promega) was digested with XbaI, Klenow fragment repaired and the firefly luciferase gene excised with BglII and ligated into HCV/RL·pA51 which had been digested with EcoRI (blunted with Klenow fragment) and BglII. The resultant vector was called pFF/HCV/RL·pA51. Thirty-three CAG repeats were cloned into the EcoRI site of pSP64 using an oligonucleotide containing the sequence d(GAATTC15CAATT15CAATT3CAATTC), to generate pSP/(CAG)33. The FF/HCV/RL·pA51 fragment was excised from pFF/HCV/RL·pA51 with BamHI and BglII and inserted into pSP/(CAG)33 which had been digested with BamHI. All cloned PCR products and oligonucleotide inserts were sequenced to ensure the absence of undesirable mutations.
The construction of pKS/FF/EMC/Ren was undertaken by first digesting E·CAT (which contains the 5'-UTR of EMC) with MscI and ligating NcoI linkers d(CAGCCATGGCTG). The EcoRI–NcoI fragment was then transferred using the same sites in phRL-null·pA. The resultant plasmid was digested with EcoRI, blunted with the Klenow fragment of DNA polymerase I and digested with BglII. The EMC/renilla-containing fragment was inserted into the XbaI (blunted) and BglII sites of pGL-3. The BglII–BamHI fragment containing FF/EMC/Ren was then transferred into the BamHI site of pKSII+. To generate plasmid pKS/FF/Ren, phRL-null·pA was digested with EcoRI, blunted with the Klenow fragment of DNA polymerase I and digested with BglII. The released fragment was cloned into the XbaI (blunted) and BglII sites of pGL-3. The BglII–BamHI fragment containing FF/Ren was then transferred into the BamHI site of pKSII+. Plasmid pcDNA3/Ren/P2/FF was a kind gift of Dr N.Sonenberg (McGill University).
In vitro Krebs translation extracts were prepared as follows. Krebs ascites cells were passaged in mice for a period of 8–10 days, after which time they were harvested, washed with cold HNG buffer (35 mM HEPES pH 7.3, 10 mM glucose, 145 mM NaCl) and resuspended in 2 packed cell volumes of hypotonic buffer (25 mM HEPES pH 7.3, 50 mM KCl, 1.5 mM MgCl2, 1 mM DTT). The cells were left to swell on ice for 20 min, transferred to a 40 ml type B Dounce homogenizer (Bellco) and lysed (20–40 strokes) (lysis was monitored by staining cells with Trypan blue and visual inspection by microscopy). Approximately 60–80% lysis was routinely achieved. A one-ninth volume of 10x concentrate (250 mM HEPES pH 7.3, 1 M KOAc, 29 mM MgCl2, 30 mM DTT) was added to the lysate, mixed gently and centrifuged at 17 000 g for 20 min. The supernatant was collected, aliquoted and stored at –80°C. A typical yield of translation extract from 30 mice was 100 ml.
Discrete compound collection, liquid handling and data capture
The discrete compound collection used in the high throughput screen consisted of 40 000 synthetic compounds from ChemBridge (San Diego, CA), 1680 synthetic and natural products from MicroSource Discovery (Gaylordsville, CT), 42 000 synthetic compounds from NCI-DTP, 2056 compounds from TimTec (Neward, DE), 8669 InterBioScreen (IBS) compounds (Moscow, Russia) and 800 purified natural products collected by one of us (J.P.). Each 96-well plate had 80 discrete compounds, four negative controls (DMSO controls), two positive controls (10 μM anisomycin) and two wells containing firefly luciferase protein. Compounds re-ordered from the Drug Synthesis and Chemistry Branch of the National Cancer Institute, IBS, or TimTec were dissolved in DMSO and equivalent amounts of solvent were included in all in vitro control reactions.
All liquid handling steps were performed on a Biomek FXII pipetting station (Beckman). Liquid handling protocols were optimized: (i) for the distribution of programmed translation mix and compound to minimize formation of air bubbles, (ii) to ensure proper reagent mixing; (iii) to eliminate cross-well and nuclease contamination. All data were captured and analyzed using ActivityBaseTM (IDBS, Boston, MA). To determine the inhibitory effect of the compounds tested, data for each compound were internally normalized to the control mean values (translations in the absence of compound) on each plate and expressed as relative inhibition.
In vitro translation assays and secondary assay
In vitro transcriptions were performed essentially as described previously (17). In vitro translations in rabbit reticulocyte lysate and Escherichia coli S30 extracts were performed according to the manufacturer’s recommendations (Promega). In vitro translations in Krebs extracts were prepared as described previously (20), except that they were formatted in 96-well conical bottom plates (Costar 3897). Ten microliter translation reactions were set up containing 4 ng/μl in vitro transcribed mRNA. The addition of compound to a final concentration of 10 μM was performed with a solid pin tool (V&P Scientific FP3 pins). The pin tool was washed in water, ethanol and DMSO, then blotted onto a cleaning pad (V&P Scientific 540D) after each compound transfer to avoid carryover and contamination. Translation plates were sealed and incubated at 30°C for 1 h, after which time 25 μl of luciferase stabilization buffer (70 mM HEPES pH 7.7, 7 mM MgSO4, 3 mM DTT, 1% BSA) was added and the plates frozen at –20°C until processed for luciferase measurements.
For in vitro translations using methionine and exogenously supplied (CAG)33/FF/HCV/Ren·pA51 mRNA, Krebs extracts were first treated with micrococcal nuclease to digest endogenous mRNA, then programmed with mRNA at a final concentration of 15 ng/μl. For gel analysis of translation products, aliquots from translation mixtures were mixed with SDS/sample buffer, boiled for 2 min and applied to a 10% SDS–polyacrylamide gel. Gels were fixed in 10% methanol, 7.5% acetic acid, treated with EN3HANCE (NEN) and exposed to X-OMAT (Kodak) film at –70°C with an intensifying screen. For in vitro translations using methionine incorporation into endogenous protein, translations were set up as above except that the micrococcal nuclease treatment step was omitted. methionine incorporation was assessed by spotting an aliquot of the lysate onto Whatman 3MM paper (that had been pre-blocked with 0.1% methionine), dried and placed in cold 10% TCA for 20 min. Filters were transferred to 5% TCA, boiled for 15 min, washed once with 5% TCA and once with 95% ethanol and dried. Radioactivity was determined by scintillation counting.
A counterscreen was developed to identify compounds that inhibited firefly or renilla luciferase activity non-specifically. Large scale translations (500 μl) in Krebs extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA were performed at 30°C for 1 h. Aliquots of 9 μl were placed in white flat-bottomed flash plates (Costar 3912) and 1 μl of appropriately diluted compound added. Plates were then processed for luciferase measurements.
Dual-luciferase reporter assay
Firefly and renilla luciferase activities were measured using a plate reading luminometer equipped with two reagent injectors (Lmax; Molecular Devices). Dual activity measurements were performed by injecting 50 μl of firefly luciferase reagent (21), waiting for 2 s, followed by a 10 s measurement period. An aliquot of 50 μl of renilla luciferase reagent was then injected, followed by a 2 s pre-measurement delay and a 10 s measurement period.
Ribosome-binding assays
Ribosome-binding assays were performed as described previously (17). Essentially, 32P-labeled CAT mRNA was incubated in 25 μl of extract at 20°C for 10 min in the presence of 0.6 mM cycloheximide or 1 mM GMP-PNP and 50 μM small molecule ligand. The final KOAc concentration was adjusted to 150 mM. Initiation complexes formed were analyzed by sedimentation through 10–30% sucrose gradients. Centrifugation was for 3.5 h at 39 000 r.p.m. at 4°C in an SW40 rotor. Fractions of 500 μl were collected and radioactivity was determined by scintillation counting.
RESULTS
Rationale of the assay
We have devised a multiplex screen based on in vitro translation of a specific mRNA reporter that enables us to simultaneously identify (i) inhibitors of cap-dependent translation (Fig. 1A, step 1), (ii) small molecules that bind to a particular RNA motif in the 5'-UTR of the first cistron and inhibit ribosome scanning (step 2), (iii) inhibitors of elongation or termination (steps 3 and 4) and (iv) inhibitors of IRES-mediated initiation (step 5). The basis of the screen lies in the ability to quantitatively assess the products of the firefly (FF) and renilla (Ren) cistrons, following in vitro translation of the test transcript. Any ligand that inhibits 5'-mediated (cap-dependent) initiation or that binds to a particular RNA motif within the 5'-UTR and inhibits 40S ribosome scanning will decrease expression of firefly luciferase, but should not affect renilla luciferase activity. We engineered 33 CAG repeats within the FF 5'-UTR to screen for compounds that could bind to the imperfect stem–loop structure predicted to be formed by this trinucleotide repeat (Fig. 1B) and subsequently inhibit ribosome scanning. Small molecule ligands interacting with RNA structural motifs within mRNA 5'-UTRs have been previously shown to inhibit translation (15,16) at the level of initiation (16). The (CAG)33 repeat was chosen because of its implication in the pathogenesis of Huntington’s disease through expansion of the repeat within the huntingtin coding region, the goal being to identify a small molecule ligand that could preferentially inhibit translation of the mRNA encoded by the allele carrying the expanded CAG repeat, due to increased target copy number.
Figure 1. Outline of the strategy used to multiplex the high throughput search for inhibitors of protein synthesis. (A) In the configuration of the assay, a bicistronic mRNA is used in which initiation can occur at the 5' end via a cap-dependent mechanism, as well as internally via an internal ribosome entry site (IRES). The firefly (FF) luciferase gene is the first shaded box and the renilla (Ren) luciferase is the second black box. The hepatitis C virus (HCV) IRES is denoted by a thickened line. The plasmid is prepared for in vitro transcription by linearizing with BamHI. Translation initiation of renilla luciferase is driven by the HCV IRES. The position of 33 CAG nucleotide triplets upstream of the firefly luciferase initiation codon is denoted. Inhibition of translation in a cell-free translation extract by a small molecule ligand can occur at one of five sites: by interfering with cap-dependent initiation (Step 1), by interfering with migration of the 40S subunit along the 5'-UTR of the mRNA (Step 2), by inhibiting elongation (Step 3), by inhibiting termination (Step 4) or by inhibiting binding of the 40S subunit to the IRES (Step 5). (B) Predicted secondary structure (G = –38 kcal/mol) formed by the CAG tract within the (CAG)33/FF/HCV/Ren·pA51 5'-UTR using Mfold (48). The bracketed region indicates that this section of the stem is repeated 13 times. (C) Schematic representation of the strategy undertaken to identify novel inhibitors of eukaryotic protein synthesis. See text for details.
Translation initiation on HCV occurs by a cap-independent mechanism and is directed by a highly structured 5'-UTR, consisting of four stem–loops and a pseudoknot structure. It contains a unique cis-acting element known as an IRES that mediates cap-independent internal initiation (22). The hepatitis C virus IRES is significantly different from other IRESs since it does not have a strict requirement for the canonical initiation factors (eIFs), other than eIF2 and eIF3 (23). Hence, a decrease in renilla luciferase activity, but not of firefly luciferase production, should occur if a small molecule ligand inhibits HCV IRES-mediated initiation. A general inhibitor of translation (initiation, elongation or termination) would be expected to reduce both firefly and renilla luciferase production.
The strategy we have established for the multiplexed screen is outlined in Figure 1C. Following the primary in vitro translation screen, positive hits are re-tested in duplicate to ensure that the initial result was a true positive. A counterscreen is then performed to eliminate non-specific inhibitors of luciferase enzyme activity. Candidates are evaluated in a secondary assay in which in vitro translation reactions in micrococcal nuclease-treated Krebs extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA are analyzed by polyacrylamide gel electrophoresis. At this point, compounds that inhibited in the secondary assay are considered candidate inhibitors.
Assay validation
To maximize the sensitivity of the in vitro translation system for detection of protein synthesis inhibitors using the bicistronic luciferase reporter several conditions first had to be met. The choice of the cell-free extract was important to allow both cap-dependent initiation and HCV IRES-mediated internal initiation. For eukaryotic in vitro translation screens, we had the option of using several extracts, derived from wheatgerm, rabbit reticulocyte lysates, Krebs cells or HeLa cells. Wheatgerm extracts were not chosen since they do not support translation initiation from the HCV IRES (J. Pelletier, unpublished data). Lysates prepared from rabbit reticulocytes do not faithfully execute cap-dependent translation and hence could not be used in our particular assay (24). Both HeLa and Kreb cell extracts can initiate in a cap-dependent fashion, as well as internally on the HCV IRES. In our small molecule screen we used extracts prepared from Krebs cells as these could be easily prepared in large quantities.
In vitro transcription from plasmid DNA template using SP6 RNA polymerase was used to generate (CAG)33/FF/HCV/Ren·pA51 mRNA for in vitro translation experiments. The quality of the RNA template was assessed by fractionation on a denaturing formaldehyde/agarose gel and indicated that the mRNA was of the expected 4 kb size (Fig. 2A). To establish the range of mRNA concentration for translation in Krebs extracts, we varied the (CAG)33/FF/HCV/Ren·pA51 mRNA concentration in Krebs extracts from 0.5 to 50 ng/μl (Fig. 2B). The results indicate that for both firefly and renilla luciferase expression, this concentration range was within the linear range. We chose to perform in vitro translations at a final mRNA concentration of 4 ng/μl in order to maintain a good signal-to-noise ratio while minimizing consumption of the mRNA template. To assess the sensitivity of Krebs extracts to DMSO, we titrated DMSO in extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA (Fig. 2C). The results indicate that the extracts can tolerate DMSO concentrations 1%, with a minimal effect of protein synthesis activity. Finally, to ensure that the Krebs extracts were expressing firefly luciferase in a cap-dependent fashion, we performed translations in extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA and containing either 500 μM GDP (Fig. 2D, lane 2) or 500 μM m7GDP (Fig. 2D, lane 3). Since addition of m7GDP, but not GDP, to the reaction significantly altered the firefly to renilla expression ratio (compare lane 3 with lanes 1 and 2), these results indicate that translation of firefly luciferase in Krebs extracts is cap dependent.
Figure 2. Evaluation of the assay format for screening inhibitors of eukaryotic protein synthesis. (A) Formaldehyde/agarose gel analysis of mRNA generated by in vitro transcription of pSP/(CAG)33/FF/HCV/Ren·pA51 linearized with BamHI. In vitro transcriptions were performed as described previously (17). RNA was fractionated on a 1.2% agarose/formaldehyde gel and visualized by staining with SYBR gold (Molecular Probes). The RNA markers are from Gibco BRL and the arrow denotes the position of migration of full-length RNA. (B) Titration of (CAG)33/FF/HCV/Ren·pA51 mRNA into Krebs translation extracts. Both axes are drawn to log scale. Experiments were performed in duplicate and the error of the mean is too small to be visualized. (C) Sensitivity of Krebs extract to DMSO utilizing (CAG)33/FF/HCV/Ren·pA51 mRNA (5 ng/ml). Experiments were performed in duplicate and the error of the mean is shown. (D) Translation in Krebs extracts is cap dependent. Translations were performed in Krebs extracts with (CAG)33/FF/HCV/Ren·pA51 mRNA (5 ng/ml) alone (lane 1), in the presence of 500 μM GDP (lane 2) or with 500 μM m7GDP (lane 3). The ratio of renilla to firefly luciferase is plotted on the ordinate. Experiments were performed in duplicate and the error of the mean is shown. The absolute RLU obtained in the control experiments (lane 1) was 249 022 and 451 008 (for luciferase) and 1 259 448 and 2 630 247 (for renilla).
Using the above assay conditions, in vitro translations were performed in control 96-well plates (with no compounds). The data indicated a well-to-well variation of 4.4% CV, a signal-to-background ratio of 50:1 and a Z' factor of 0.8 (data not shown) (25). These parameters suggest the suitability of this assay for use in a miniaturized format. To validate that this screen could identify translation inhibitors, we chose a set of known inhibitors representing examples of different activity spectra (Supplementary Material, Fig. S1). The behavior of many of these compounds in Krebs extracts has not been previously documented. Anisomycin and bruceantin are inhibitors of peptidyl transferase (26,27), didemnin B binds to ribosome/eEF-1 complexes and blocks eEF-2 binding to pre-translocative ribosomes (28,29), bouvardin inhibits eEF1-dependent binding of aminoacyl-tRNA and eEF2-dependent translocation (30) and verrucarin A inhibits elongation, but is only effective on reinitiating ribosomes (31). The mechanism of action of baccharinol has not been reported, but it is a trichothecin epoxide that likely inhibits elongation. Compounds were titrated into 96-well plates, reformatted, added to translation reactions and tested at assay concentrations ranging from 0.65 nM to 10 μM. The IC50 values of the compounds ranged from 4 (bouvardin) to 300 nM (verrucarin) and demonstrated that different protein synthesis inhibitors with various potencies are readily detected in our assay.
Discrete compound screen
To assess the behavior of our assay in a high throughput screen format, we screened 50 96-well plates of our discrete small molecule collection (Supplementary Material, Fig. S2). The distribution of the firefly and renilla values obtained from the tested compound population indicated a good performance of this assay in the high throughput screen format. Using a cut-off value of 3-fold inhibition of translation (RLU < 0.33), 0.75% of compounds were positive for inhibition of firefly activity and 1.1% of compounds showed inhibition of renilla activity. Of these, five (0.125%) inhibited both firefly and renilla activity.
Based on these results, our discrete compound collection was screened once in 96-well format. The set of small molecule ligands still positive after the selection criteria denoted in Figure 1C are listed in Table 1. Of the 36 compounds identified in our screen (0.05% hit rate), 14 of these were known or structurally related to known protein synthesis inhibitors. These compounds provide an internal validation of the ability of our screen to identify inhibitors of protein synthesis. Since the mechanisms of action of these compounds are well characterized, they were not further studied. As many of the compounds we identified were nucleic acid-binding ligands or intercalators (e.g. acriflavine, ethidium bromide, dequalinium, Hoechst 33258, etc.), we established an assay to quickly determine if a given compound was a nucleic acid-binding ligand. This assay consists of incubating supercoiled DNA with 50 μM compound and monitoring alterations in electrophoretic mobility on agarose gels (Fig. 3A). DNA binding results in loss of negative supercoils producing a decrease in mobility when analyzed by agarose gel electrophoresis (Fig. 3A, compare lane 3 with lane 2). Based on this assay, we conclude that NSC 305831 (lane 4), NSC 22907 (lane 5), NSC 130813 (lane 6), NSC 254681 (lane 7), NSC 101327 (lane 10) and NSC 85701 (lane 12) bind to DNA. In this assay, NSC 119889 (lane 8), suramine (lane 9), NSC 111041 (lane 11) and NSC 115183 (lane 13) did not affect the mobility of supercoiled DNA and for the purposes of this study are not classified as general DNA-binding compounds (Table 1).
Table 1. Small molecule ligands identified in the high throughput screen
Figure 3. (A) Intercalation assay assessing the ability of compounds to alter the mobility of supercoiled plasmid DNA. Supercoiled plasmid DNA was incubated with 50 μM ligand at room temperature for 20 min and electrophoresed into a 0.8% agarose gel. The gel was stained with ethidium bromide (0.5 μg/ml) and visualized using short wavelength UV light. The identity of the compounds incubated with supercoiled DNA is indicated above the panel. The position of migration of supercoiled DNA is indicated by an arrow. (B–D) Relative production of firefly and renilla by translation of (CAG)33/FF/HCV/Ren·pA51 mRNA in the presence of small molecule ligands. Following in vitro translations, methionine-labeled samples were treated in SDS sample buffer and electrophoresed into a 10% SDS–polyacrylamide gel. The gels were treated with EN3Hance, dried and exposed to X-Omat (Kodak) film. The concentration of compound used in the translation mix is indicated above the panel and the identity of the compound indicated below the panel. The position of migration of firefly and Renilla luciferase protein are indicated.
Since the initial high throughput translation assay was set up to identify compounds that could inhibit one of several steps, we undertook a series of experiments to deconvolute their mode of action. Addition of compound to Krebs extracts programmed with (CAG)33/FF/HCV/Ren·pA51 mRNA, followed by SDS–PAGE analysis of methionine products, indicated that the small molecule ligand could be classified into one of three categories, depending on which ORF they preferentially inhibited. (i) At 50 μM, acriflavine (Fig. 3B) and ethidium bromide (A. Malina and J. Pelletier, data not shown) inhibited expression of both firefly and renilla luciferase, however, at 10 μM, translation of renilla luciferase was significantly inhibited (>10-fold) whereas that of firefly luciferase was affected 2-fold (Fig. 3B). (ii) At 50 μM, NSC 119889 completely inhibited firefly luciferase translation, but had little effect on renilla expression (Fig. 3C). (iii) The remaining set of compounds were general inhibitors of translation that equally inhibited synthesis of firefly and renilla luciferase (Fig. 3D). These included NSC 22907, NSC 305831, NSC 130813, NSC 254681, NSC 101327, NSC 85701, NSC 111041 and NSC 115183 (Fig. 3D), as well as quinacrine, Hoechst 33258, mitoxanthrone, chartreusin, nogalamycin, quinolinium and ellipticine derivatives (data not shown). Given the large number of intercalators identified in this study, detailed characterization of their modes of action will be presented elsewhere (A. Malina and J. Pelletier, manuscript in preparation) and for the current study we will focus on the four non-intercalating compounds, NSC 119889, NSC 111041, suramine and NSC 115183 (Fig. 4A).
Figure 4. (A) Structures of NSC 119889, NSC 111041 and suramine. The structure of NSC 115183 is currently not available. (B) Titration of compounds in Krebs extracts. Translations were performed in the presence of the indicated amounts of compound and at final mRNA and K+ concentrations of 5 μg/ml and 100 mM, respectively. Control translation reactions contained 0.5% DMSO, which is equivalent to what was present in reactions containing compound. Luciferase values are normalized to the activity obtained in the control translations (which were set to 1). The relative firefly luciferase values are represented by white bars and the relative renilla values are represented by black bars. Translations were performed three times and the average values are presented along with the error of the mean. (C) Titration of compounds in rabbit reticulocyte lysate. The relative firefly luciferase values are represented by white bars and the relative renilla values are represented by black bars. Translations were performed twice and the average values are presented along with the error of the mean. (D) Effect of compounds (50 μM) on translation in E.coli S30 extracts. Protein synthesis was monitored by assessing the amount of methionine incorporation into TCA-precipitable material by endogenous mRNA. Translations were performed twice and the average values are presented along with the error of the mean.
Titration of these compounds into Krebs translation extracts programmed with (CAG)33/FF/HCV/Ren·pA51 indicated IC50 values of 10, 1 and 0.4 μM for inhibition of both firefly and renilla translation by suramine, NSC 111041 and NSC 115183, respectively. The IC50 values for inhibition of firefly and renilla translation were 10 and 50 μM, respectively, for NSC 119889, indicating a differential effect on expression of firefly and renilla cistrons by this compound (Fig. 4B). The inhibitory activity of these compounds was also assessed in rabbit reticulocyte lysates programmed with (CAG)33/FF/HCV/Ren·pA51 (Fig. 4C) and showed the same relative potency as in Krebs extracts, with suramine being the least active (IC50 50 μM) and NSC 115183 being the most potent compound (IC50 0.2 μM) (Fig. 4C). NSC 119889 maintained the same differential effect on firefly and renilla expression as observed in Krebs extracts (Fig. 4C). To assess if any of these compounds exerted an effect on prokaryotic protein synthesis, they were tested in an E.coli S30 extract. Suramine, NSC 111041 and NSC 115183 did not significantly affect translation when tested to a final concentration of 50 μM, whereas NSC 119889 inhibited translation when present at a final concentration of 50 μM (Fig. 4D).
A compound that inhibits initiation of translation will prevent formation of 80S ribosome–mRNA complexes, whereas inhibitors of elongation or termination will not affect loading of ribosomes onto mRNA templates. Ribosome binding experiments were performed in the presence of cycloheximide (an inhibitor of elongation) and 50 μM compound. Ribosome–mRNA complexes were fractionated on sucrose gradients and the radioactivity in the individual fractions determined by scintillation counting (Fig. 5). Suramine, NSC 111041 and NSC 115183 did not affect the efficiency of 80S ribosome–mRNA complex formation (Fig. 5A and B). These experiments indicate that suramine, NSC 111041 and NSC 115183 exert their effects by inhibiting downstream of initiation. On the other hand, NSC 119889 prevented mRNA–ribosome interaction, indicating that this compound inhibits initiation of protein synthesis (Fig. 5C). A non-hydrolyzable analog of GTP, GMP-PNP, can be used to trap 43S initiation complexes at initiation codons of mRNA templates since it inhibits the release of assembled initiation factors from the small subunit, preventing 60S subunit joining (Fig. 5D). When NSC 119889 was present in the binding reactions with GMP-PNP, 43S complex binding was inhibited, indicating that this compound inhibits ribosome binding to mRNA templates (Fig. 5D).
Figure 5. Effect of compounds on initiation complex assembly in rabbit reticulocyte lysates. 32P-labeled CAT mRNA was incubated in rabbit reticulocyte lysates and initiation complexes resolved by centrifugation on sucrose gradients. Fractions from each sucrose gradient were collected using a Brandel Tube Piercer connected to an ISCO fraction collection and were individually counted. (A) Complex formation in the presence of 600 μM cycloheximide, 600 μM cycloheximide and 50 μM NSC 111041 or 600 μM cycloheximide and 50 μM suramine. The total counts recovered and the percentage mRNA bound to 80S complexes were: CAT mRNA/cycloheximide (320 856 c.p.m., 6% binding); CAT mRNA/cycloheximide + NSC 111041 (304 285c.p.m., 6% binding); CAT mRNA/cycloheximide + suramine (315 307 c.p.m., 3% binding). (B) Complex formation in the presence of 600 μM cycloheximide and 50 μM NSC 115183. The total counts recovered and the percentage mRNA bound to 80S complexes were: CAT mRNA/cycloheximide (109 065 c.p.m., 10% binding); CAT mRNA/cycloheximide + NSC 115183 (112 299 c.p.m., 9% binding). (C) Complex formation in the presence of 600 μM cycloheximide and 50 μM NSC 119889. The total counts recovered and the percentage mRNA bound to 80S complexes were: CAT mRNA/cycloheximide (182 569 c.p.m., 7% binding); CAT mRNA/cycloheximide + NSC 119889 (286 708 c.p.m., 0% binding). (D) Complex formation in the presence of 1 mM 5'-guanylylimidodiphosphate (GMP-PNP) or 1 mM GMP-PNP and 50 μM NSC 119889. Total counts recovered from each gradient and the percentage mRNA bound to 43S complexes were: CAT mRNA/GMP-PNP (98 359 c.p.m., 7% binding); CAT mRNA/GMP-PNP + NSC 119889 (95 225 c.p.m., 0% binding).
To better gauge the inhibitory spectrum of NSC 119889, its effects on expression of several IRESs was assessed (Fig. 6). The IRESs tested were derived from EMC, poliovirus and HCV (Fig. 6A). As previously indicated, translation initiation from the HCV IRES was particularly resistant to inhibition by NSC 119889 , whereas cap-dependent translation was completely inhibited when NSC 119889 was present in the translation reaction at a final concentration of 50 μM. The effect of NSC 119889 on expression of firefly and renilla ORFs from Krebs extracts programmed with FF/EMC/Ren or Ren/P2/FF mRNAs was similar in that expression of both ORFs appeared to be equally inhibited from the two templates. These results indicate that unlike with the HCV IRES, initiation from both EMC and poliovirus IRESs is sensitive to inhibition by NSC 119889 (Fig. 6B and C). Translation of both firefly and renilla ORFs from FF/Ren mRNA was equally affected by NSC 119889, consistent with expression of both ORFs from this mRNA template being cap dependent (Fig. 6B and C). In this particular case, expression of renilla luciferase is expected to arise as a consequence of reinitiation by ribosomes that have completed translation of the firefly ORF. These results are consistent with different initiation mechanisms for ribosome binding between HCV and picornaviral IRESs.
Figure 6. (A) Schematic representation of constructs used to assess the inhibitory potential of NSC 119889 on different IRESs. The names of the plasmids harboring the inserts is provided to the left and the restriction sites used to linearize the plasmids denoted to the right. Note that mRNA derived from pKS/FF/Ren, pSP(CAG)33/FF/HCV/Ren·pA51 and pKS/FF/EMC/Ren will contain a poly(A) tail of 51 adenylate residues. (B) Relative translational efficiencies obtained in Krebs extracts programmed with FF/Ren, (CAG)33/FF/HCV/Ren, FF/EMC/Ren and Ren/P2/FF mRNA in the presence of NSC 119889. Control translation reactions contained 0.5% DMSO. Luciferase values were normalized to the activity obtained in the control translations (which were set at one). The relative firefly luciferase values are represented by white bars and the relative renilla values are represented by black bars. The bar represents the average of two experiments and the error of the mean is denoted. (C) A representative experiment demonstrating the relative translation efficiencies of firefly and renilla production from the different test mRNAs in the presence of NSC 119889. Following in vitro translations in Krebs extracts in the presence of methionine, the indicated amounts of NSC 119889 and 5 ng/ul mRNA, samples were treated in SDS sample buffer and electrophoresed into 10% SDS–polyacrylamide gels. Gels were treated with EN3Hance, dried and exposed to X-Omat (Kodak) film. The concentration of NSC 119889 in the translation reaction is indicated above the panel and the identity of the mRNA used to program the translation reaction is indicated below the panel. The position of migration of firefly and renilla luciferase proteins is denoted. The position of molecular mass markers (NEB) is indicated to the left.
NSC 119889 contains a xanthenyl ring structure relating it to fluorescein or gallein. A small number of structurally related compounds were obtained to identify the functionally important groups of NSC 119889 for translation inhibition. The structure–activity study was performed at a final concentration of 50 μM in Krebs extracts programmed with (CAG)33/FF/HCV/Ren (Fig. 7A). Fluorescein and related compounds (NSC 119892, NSC 122391, NSC 119894, NSC 157411, NSC 122390, NSC 119887 and NSC 119888) were ineffective at inhibiting firefly luciferase (Fig. 7A and B), whereas NSC 119889 and NSC 378139 were active. With the exception of NSC 622475, the addition of two hydroxyl groups to the xanthenyl ring system at positions 4' and 5' yielded compounds (NSC 119915, NSC 119911, gallein, NSC 119893, NSC 119913 and NSC 119910) that inhibited firefly luciferase production (Fig. 7A and B). Hydroxyl group substitutions at the 2' and 7' position yielded compounds (NSC 2608 and NSC 9037) that were inactive as inhibitors of protein synthesis. Although far from an exhaustive study, these results indicate that a 4',5'-hydroxyl substituted xanthenyl ring structure is important for activity; most compounds lacking these hydroxyl groups are inactive. However, exceptions such as NSC 119889 and NSC 378139 indicate that a second structural component of the small molecule (contributed by the aromatic ring) is also important for activity.
Figure 7. Structure–activity relationships of NSC 119889 analogs. Translations were performed as indicated in Materials and Methods. (A) Effect of 50 μM compound on translation in Krebs extracts programmed with (CAG)33/FF/HCV/Ren mRNA. Luciferase values were normalized to the activity obtained in the control translations (which were set to 1). Experiments were performed three times with each compound and the error of the mean is indicated. The chemical structures are shown below the bar graph. (B) A representative experiment demonstrating the relative translation efficiencies of firefly and renilla production from (CAG)33/FF/HCV/Ren mRNA in the presence of the different NSC analogs. Following in vitro translation in Krebs extracts in the presence of methionine, 50 μM compound, and 5 ng/μl mRNA, samples were treated in SDS sample buffer and electrophoresed into 10% SDS–polyacrylamide gels. Gels were treated with EN3Hance, dried and exposed to X-Omat (Kodak) film. The identity of the compound used in the translation reaction is denoted above the panel. The position of migration of firefly and renilla luciferase proteins is denoted. The position of molecular mass markers (NEB) is indicated to the left.
DISCUSSION
Monitoring translation of a bicistronic mRNA reporter in a eukaryotic translation extract was utilized to identify novel protein synthesis inhibitors. This system was adapted to a miniaturized format to save time, cost and reagents. The initial screen was set up to identify general inhibitors of eukaryotic protein synthesis, as well as small molecule ligands that could bind to the CAG repeats engineered upstream of the firefly ORF, or that could bind to the HCV IRES. The detection of known protein synthesis inhibitors from chemical libraries by our screen provided an internal validation (Table 1). Given the importance of RNA helices to ribosome function and the presence of double-stranded regions on mRNA templates, it is not surprising that we identified a large number of intercalators by our assay (Table 1). Indeed, several previous reports have documented the ability of nucleic acid-binding agents to inhibit protein synthesis. Ethidium bromide (32) and acriflavine (33) have been reported to inhibit the aminoacylation of tRNAs in vitro, presumably by intercalating into tRNA and preventing its recognition by aminoacyl transferases. The intercalators 3-nitrobenzothiazolo{3,2-a}quinolinium and fagaronine have been shown to inhibit protein synthesis in cell-free extracts, possibly by affecting elongation (34). Detailed characterization of the mode of action of nucleic acid-binding ligands on translation will be presented elsewhere (A. Malina and J. Pelletier, manuscript in preparation).
Whereas general inhibitors of translation affecting elongation and initiation were identified, our screen did not identify ligands that could bind to the CAG repeats and inhibit translation. This may indicate the need to screen a larger, more diverse collection of compounds or alternatively may indicate that ligands binding to the CAG repeat were unable to provide a kinetic barrier to inhibit initiation. If the latter case was true, then an approach incorporating the use of inducers of dimerization to recruit protein surfaces to the RNA target may improve the inhibitory potential of RNA-binding ligands (17). This would necessitate re-formating our screen to identify such compounds. Our screen did, however, identify two compounds, ethidium bromide and acriflavine, which showed preferential inhibition of HCV-driven renilla expression when present at 10 μM in the translation reaction (Fig. 3B). Detailed characterization of these compounds indicates that at 10 μM they block ribosome–HCV IRES interaction (A. Malina and J. Pelletier, manuscript in preparation).
Our screen identified four protein synthesis inhibitors that did not bind double-stranded DNA when tested in vitro (Fig. 3A). Three of these compounds, suramine, NSC 111041 and NSC 115183, do not inhibit initiation complex formation (Fig. 5). We have tested the potential of these compounds to cause misreading and did not obtain any evidence that these compounds affected ribosome translation fidelity (A.-S. Guenier and J. Pelletier, data not shown). Suramine is a polysulfonated napthylurea to which has been associated a large number of biological activities, ranging from trypanocidal to anticancer activity (for a review see 35). The large number of negative charges associated with this molecule may suggest that it inhibits translation in a non-specific fashion by binding factors with cationic charges, although we have no direct evidence for this mode of action. Inhibition of protein synthesis by suramine may be the basis for some of the toxic side-effects associated with the use of this compound in the clinic (35). The synthesis of NSC 111041 has been reported (36) but we know of no studies documenting an effect on protein synthesis by this compound. Unfortunately, no structure for NSC 115183 is currently available from the NCI-DTP. Our results suggest that these compounds inhibit either elongation or termination of translation.
Suramine, NSC 111041 and NSC 115183 had no effect when tested in a prokaryotic translation system, whereas NSC 119889 significantly reduced protein synthesis when present at 50 μM (Fig. 4D). These results indicate a strong discrimination on the part of suramine, NSC 111041 and NSC 115183 for the eukaryotic translation apparatus. This may be due to the absence of the compound target in prokaryotes or reflect structural differences in the ligand-binding site between eukaryotes and prokaryotes. Alternatively, there is a marked difference in the divalent cation optima between prokaryotic and eukaryotic extracts. Prokaryotic systems have a higher in vitro Mg2+ optimum (10–15 mM) (37), whereas the eukaryotic systems we used have a lower Mg2+ optimum (2.0–3.0 mM for wheatgerm and rabbit reticulocyte) (38,39), which in turn could influence binding of ligand to their targets or the conformation of the binding site. Of course, we cannot exclude non-specific effects such as the possibility that macromolecules in the prokaryotic extract bind to and titrate out the ligand, hence preventing inhibition by the compound. Differences in ion concentration or protein target concentration between Krebs and reticulocyte lysates may also be responsible for differences in IC50 values noted between these extracts for suramine, NSC 111041, NSC 115183 and NSC 119889 (Fig. 4B and C).
NSC 119889 inhibits initiation of protein synthesis, as it prevented the formation of 48S and 80S initiation complexes (Fig. 5). The ability of the HCV IRES to direct protein synthesis in the presence of NSC 119889 indicates that this compound does not exert additional effects on elongation or termination (Fig. 6). NSC 119889 is not acting through the CAG repeats present within the 5'-UTR of (CAG)33/FF/HCV/Ren since it was also active on reporter constructs lacking these repeats (Fig. 6). Additionally, NSC 119889 inhibited expression of cistrons whose initiation was dependent on the EMC or poliovirus IRES, but had no effect on translation initiation mediated by the HCV IRES (Fig. 6). These results provide insight into the mode of action of NSC 119889 as it has been documented that the canonical factor requirement for ribosome binding to the EMC and poliovirus IRESs differ from those of the HCV IRES (for a review see 40). The HCV IRES can directly recruit the 43S particle and makes intimate contact with the 40S ribosomal subunit and eIF3 (41). Aside from eIF2/GTP/Met-tRNAi and eIF3, no other initiation factors appear necessary for high affinity binding of the HCV IRES to 43S ribosomes (40,41). In contrast, internal initiation on the picornaviral EMC and poliovirus IRESs requires IRES transacting factors (ITAFs), as well as a larger complement of canonical initiation factors, including eIF4F or eIF4A and the central third of eIF4G to which eIF4F binds (reviewed in 40). The observed inhibition on EMCV and poliovirus IRES-mediated translation, but not on HCV IRES-mediated initiation, suggests that NSC 119889 may be targeting an ITAF or an initiation factor other than eIF3 or eIF2.
NSC 119889 was previously identified in a screen for antagonists of the interaction between HIV nucleocapsid protein NC-p7 and nucleic acids and exerted significant anti-HIV activity in vivo when present in the culture medium of HIV-infected cells (42). Characterization of NSC 119889 in this assay revealed that it bound to NC-p7 with a Kd of 2.5 x 10–7 M. It is thought that NSC 119889 interferes with NC-p7 function during the reverse transcription process (42). Consistent with our results indicating that NSC 119889 does not interact with DNA (Fig. 3A), BIAcore studies failed to elucidate an interaction between a related molecule, tetrachlorogallein, and DNA (42).
Additionally, fluorescein analogs are active against a variety of enzymes that use ATP or NADH as cofactors (43–47). It is thought that the fluorescein analogs can reversibly compete for ATP- or NADH-binding sites. Specifically, eosin can inhibit the mitochondrial ADP/ATP carrier (43) and the erythrocyte calcium pump (44) and erythrosine inhibits lactate dehydrogenase (45), aspartate transcarbamylase (46) and creatine kinase (47). These observations suggest a mode of action for NSC 119889 (and structural analogs) in translation, possibly by interfering with the activity of a factor that utilizes ATP (or GTP). Clearly, NSC 119889 is not a promiscuous inhibitor of GTP-binding proteins since elongation of protein synthesis remains unaffected when initiation is directed by the HCV IRES (Fig. 3).
In sum, we have established a high throughput screen that has allowed the identification of a large number of protein synthesis inhibitors. We found that a large class of these were nucleic acid-binding ligands. The characterization of four non-intercalating compounds indicated that we isolated inhibitors of both initiation and elongation. The characterization of NSC 119889 as having a differential effect on initiation driven from picornaviral versus HCV IRESs suggests that it might be used to better understand mechanisms of internal initiation.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the technical assistance of Mr Virabhadrachari Vipparti and the kind help of Abba Malina in screening the NCI diversity set in the initial stages of this work. We are immensely grateful to the NIH/NCI Developmental Therapeutics Program for their generous supply of the discrete chemical compound library as well as for re-supply of positive hits for confirmation. J.P. is a Canadian Institutes of Health Research (CIHR) Senior Investigator. This work was supported by grants from the Hereditary Disease Foundation (CHDI) to J.P., the National Cancer Institute of Canada (nos 011040 and 014313) to J.P., funding from the Jean-Louis Lévesque Foundation and a VRQ (Valorisation Recherche Quebec) grant (2201-147).
REFERENCES
Pestka,S. (1977) Inhibitors of Protein Synthesis. Academic Press, New York, NY.
Vazquez,D. (1979) Inhibitors of protein biosynthesis. Mol. Biol. Biochem. Biophys., 30, 1–312.
Watkins,S.J. and Norbury,C.J. (2002) Translation initiation and its deregulation during tumorigenesis. Br. J. Cancer, 86, 1023–1027.
Graff,J.R. and Zimmer,S.G. (2003) Translational control and metastatic progression: enhanced activity of the mRNA cap-binding protein eIF-4E selectively enhances translation of metastasis-related mRNAs. Clin. Exp. Metastasis, 20, 265–273.
Hershey,J.W. and Miyamoto,S. (2000) Translation Control and Cancer. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Ruggero,D. and Pandolfi,P.P. (2003) Does the ribosome translate cancer? Nature Rev. Cancer, 3, 179–192.
Meric,F. and Hunt,K.K. (2002) Translation initiation in cancer: a novel target for therapy. Mol. Cancer Ther., 1, 971–979.
Brodersen,D.E., Clemons,W.M.,Jr, Carter,A.P., Morgan-Warren,R.J., Wimberly,B.T. and Ramakrishnan,V. (2000) The structural basis for the action of the antibiotics tetracycline, pactamycin and hygromycin B on the 30S ribosomal subunit. Cell, 103, 1143–1154.
Carter,A.P., Clemons,W.M., Brodersen,D.E., Morgan-Warren,R.J., Wimberly,B.T. and Ramakrishnan,V. (2000) Functional insights from the structure of the 30S ribosomal subunit and its interactions with antibiotics. Nature, 407, 340–348.
Pioletti,M., Schlunzen,F., Harms,J., Zarivach,R., Gluhmann,M., Avila,H., Bashan,A., Bartels,H., Auerbach,T., Jacobi,C. et al. (2001) Crystal structures of complexes of the small ribosomal subunit with tetracycline, edeine and IF3. EMBO J., 20, 1829–1839.
Hansen,J.L., Ippolito,J.A., Ban,N., Nissen,P., Moore,P.B. and Steitz,T.A (2002) The structures of four macrolide antibiotics bound to the large ribosomal subunit. Mol. Cell, 10, 117–128.
Mandal,M., Boese,B., Barrick,J.E., Winkler,W.C. and Breaker,R.R. (2003) Riboswitches control fundamental biochemical pathways in Bacillus subtilis and other bacteria. Cell, 113, 577–586.
Mironov,A.S., Gusarov,I., Rafikov,R., Lopez,L.E., Shatalin,K., Kreneva,R.A., Perumov,D.A. and Nudler,E. (2002) Sensing small molecules by nascent RNA: a mechanism to control transcription in bacteria. Cell, 111, 747–756.
Nahvi,A., Sudarsan,N., Ebert,M.S., Zou,X., Brown,K.L. and Breaker,R.R. (2002) Genetic control by a metabolite binding mRNA. Chem. Biol., 9, 1043–1049.
Werstuck,G. and Green,M.R. (1998) Controlling gene expression in living cells through small molecule–RNA interactions. Science, 282, 296–298.
Harvey,I., Garneau,P. and Pelletier,J. (2002) Inhibition of translation by RNA–small molecule interactions. RNA, 8, 452–463.
Harvey,I., Garneau,P. and Pelletier,J. (2002) Forced engagement of a RNA/protein complex by a chemical inducer of dimerization to modulate gene expression. Proc. Natl Acad. Sci. USA, 99, 1882–1887.
Gold,L., Polisky,B., Uhlenbeck,O. and Yarus,M. (1995) Diversity of oligonucleotide functions. Annu. Rev. Biochem., 64, 763–797.
Sambrook,J. and Russell,D.W. (2001) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Svitkin,Y.V. and Agol,V.I. (1978) Complete translation of encephalomyocarditis virus RNA and faithful cleavage of virus-specific proteins in a cell-free system from Krebs-2 cells. FEBS Lett., 87, 7–11.
Dyer,B.W., Ferrer,F.A., Klinedinst,D.K. and Rodriguez,R. (2000) A noncommercial dual luciferase enzyme assay system for reporter gene analysis. Anal. Biochem., 282, 158–161.
Tsukiyama-Kohara,K., Iizuka,N., Kohara,M. and Nomoto,A. (1992) Internal ribosome entry site within hepatitis C virus RNA. J. Virol., 66, 1476–1483.
Pestova,T.V., Shatsky,I.N., Fletcher,S.P., Jackson,R.J. and Hellen,C.U. (1998) A prokaryotic-like mode of cytoplasmic eukaryotic ribosome binding to the initiation codon during internal translation initiation of hepatitis C and classical swine fever virus RNAs. Genes Dev., 12, 67–83.
Svitkin,Y.V., Ovchinnikov,L.P., Dreyfuss,G. and Sonenberg,N. (1996) General RNA binding proteins render translation cap dependent. EMBO J., 15, 7147–7155.
Zhang,J.-H., Chung,T.D.Y. et al. (1999) A simple statistical parameter for use in evaluation and validation of high throughput screening assays. J. Biomol. Screening, 4, 67–73.
Carrasco,L. and Vazquez,D. (1972) Survey of inhibitors in different steps of protein synthesis by mammalian ribosomes. J. Antibiot., 25, 732–737.
Fresno,M., Gonzales,A., Vazquez,D. and Jimenez,A. (1978) Bruceantin, a novel inhibitor of peptide bond formation. Biochim. Biophys. Acta, 518, 104–112.
SirDeshpande,B.V. and Toogood,P.L. (1995) Mechanism of protein synthesis inhibition by didemnin B in vitro. Biochemistry, 34, 9177–9184.
Ahuja,D., Vera,M.D., SirDeshpande,B.V., Morimoto,H., Williams,P.G., Joullie,M.M. and Toogood,P.L. (2000) Inhibition of protein synthesis by didemnin B: how EF-1alpha mediates inhibition of translocation. Biochemistry, 39, 4339–4346.
Zalacain,M., Zaera,E., Vazquez,D. and Jimenez,A. (1982) The mode of action of the antitumor drug bouvardin, an inhibitor of protein synthesis in eukaryotic cells. FEBS Lett., 148, 95–97.
Cannon,M., Jimenez,A. and Vazquez,D. (1976) Competition between trichodermin and several other sesquiterpene antibiotics for binding to their receptor site(s) on eukaryotic ribosomes. Biochem. J., 160, 137–145.
Landez,J.H., Roskoski,R.J.,Jr and Coppoc,G.L. (1969) Ethidium bromide and chloroquine inhibition of rat liver cell-free aminoacylation. Biochim. Biophys. Acta, 195, 276–279.
Birkmayer,G.D. and Balda,B.R. (1970) Proflavine inhibition of protein synthesis in malignant hamster melanom. FEBS Lett., 11, 221–223.
Torres,C.A.C. and Baez,A. (1986) Effects of the antitumor drugs 3-nitrobenzothiazolo{3,2-a}quinolinium and fagaronine on nucleic acid and protein synthesis. Biochem. Pharmacol., 35, 679–685.
Kaur,M., Reed,E., Sartor,O., Dahut,W. and Figg,W.D. (2002) Suramin’s development: what did we learn? Invest. New Drugs, 20, 209–219.
Pettit,G., Fleming,W. and Paull,K.D. (1968) Synthesis of the 6- and 7-hydrozy-5,8-dioxocarbostrils. J. Org. Chem., 33, 1089–1092.
Chen,H.-Z. and Zubay,G. (1983) Prokaryotic coupled transcription–translation. Methods Enzymol., 101, 674–690.
Anderson,C.W., Straus,J.W. and Dudock,B.S. (1983) Preparation of a cell-free protein-synthesizing system from wheat germ. Methods Enzymol., 101, 635–644.
Merrick,W.C. (1983) Translation of exogenous mRNAs in reticulocyte lysates. Methods Enzymol., 101, 606–615.
Pestova,T.V., Kolupaeva,V.G., Lomakin,I.B., Pilipenko,E.V., Shatsky,I.N., Agol,V.I. and Hellen,C.U. (2001) Molecular mechanisms of translation in eukaryotes. Proc. Natl Acad. Sci. USA, 98, 7029–7036.
Kieft,J.S., Zhou,K., Jubin,R. and Doudna,J.A. (2001) Mechanism of ribosome recruitment by hepatitis C IRES RNA. RNA, 7, 194–206.
Stephen,A.G., Worthy,K.M., Towler,E., Mikovits,J.A., Sei,S., Roberts,P., Yang,Q.E., Akee,R.K., Klausmeyer,P., McCloud,T.G. et al. (2002) Identification of HIV-1 nucleocapsid protein: nucleic acid antagonists with cellular anti-HIV activity. Biochem. Biophys. Res. Commun., 296, 1228–1237.
Majima,E., Yamaguchi,N., Chuman,H., Shinohara,Y., Ishida,M., Goto,S. and Terada,H. (1998) Binding of the fluorescein derivative eosin Y to the mitochondrial ADP/ATP carrier: characterization of the adenine nucleotide binding site. Biochemistry, 37, 424–432.
Gatto,C. and Milanick,M.A. (1993) Inhibition of the red blood cell calcium pump by eosin and other fluorescein analogues. Am. J. Physiol., 264, C1577–C1586.
Wassarman,P.M. and Lentz,P.J.,Jr (1971) The interaction of tetra-iodofluorescein with dogfish muscle lactate dehydrogenase: a chemical and X-ray crystallographic study. J. Mol. Biol., 60, 509–522.
Jacobsverg,L.B., Kantrowitz,E.R. and Lipscomb,W.N. (1975) Interaction of tetraiodofluorescein with aspartate transcarbamylase and its isolated catalytic and regulatory subunits. J. Biol. Chem., 24, 9238–9249.
Somerville,L.L. and Quiocho,F.A. (1977) The interaction of tetraiodofluorescein with creatine kinase. Biochim. Biophys. Acta, 481, 493–499.
Zuker,M. (2003) Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res., 31, 3406–3415.(Olivia Novac1,2, Anne-Sophie Guenier1,2 )