Elongation factor 2 as a novel target for selective inhibition of fungal protein synthesis.

Elongation factor 2 (EF2) is an essential protein catalyzing ribosomal translocation during protein synthesis and is highly conserved in all eukaryotes. It is largely interchangeable in translation systems reconstituted from such divergent organisms as human, wheat, and fungi. We have identified the sordarins as selective inhibitors of fungal protein synthesis acting via a specific interaction with EF2 despite the high degree of amino acid sequence homology exhibited by EF2s from various eukaryotes. In vitro reconstitution assays using purified components from human, yeast, and plant cells demonstrate that sordarin sensitivity is dependent on fungal EF2. Genetic analysis of sordarin-resistant mutants of Saccharomyces cerevisiae shows that resistance to the inhibitor is linked to the genes EFT1 and EFT2 that encode EF2. Sordarin blocks ribosomal translocation by stabilizing the fungal EF2-ribosome complex in a manner similar to that of fusidic acid. The fungal specificity of the sordarins, along with a detailed understanding of its mechanism of action, make EF2 an attractive antifungal target. These findings are of particular significance due to the need for new antifungal agents.

The elongation phase of translation in fungi requires the soluble elongation factors EF1␣, EF2, and EF3. EF1␣ and EF2 are members of the GTPase superfamily of proteins and are characterized by common structural motifs and their ability to alternate between conformational states in response to binding GDP or GTP. These proteins are required for translation in all eukaryotes, while EF3 is unique to fungi and essential for fungal protein synthesis (1). EF2 catalyzes the translocation of the ribosome along messenger RNA, presumably by stimulating a gross rearrangement of the ribosome that results in peptidyl-tRNA transfer and the movement of mRNA by one codon. The protein sequence of EF2 has been highly conserved throughout evolution, with Saccharomyces cerevisiae EF2 sharing 66% identity and 85% homology to human EF2. Despite this high degree of similarity, a class of tetracyclic diterpene glycoside natural products, the sordarins, has now been identified as selective inhibitors of EF2 function in fungal protein synthesis. Sordarin, produced by species of the fungal genus Sordaria, was described as an antifungal agent in 1970 (2,3), but the mode of action of this family has not been examined until now. In this report, we show that sordarins specifically bind to the S. cerevisiae EF2-ribosome complex and block protein synthesis by inhibiting the release of EF2 from the posttranslocational ribosome. Our observations show that it is possible to inhibit fungal EF2 specifically, which may provide an opportunity for developing antifungal agents with a unique and selective mechanism of action.
EXPERIMENTAL PROCEDURES Sordarin was isolated essentially as described for Sordaria arenosa (2). Reticulocyte and wheat germ lysates were obtained from Promega.
Assays-IC 50 values were determined from growth inhibition assays in which cells were inoculated at 2 ϫ 10 5 cells/ml in YPAD medium (4) containing sordarin serially diluted 2-fold from 0.2-100 g/ml, followed by incubation at 29°C for approximately 16 h. S30 extracts and translation factors EF1, EF2, and EF3 were prepared from S. cerevisiae cells harvested in mid-to-late logarithmic phase (A 600 ϳ2). Translation assays using S30 extracts were performed at 22°C essentially as described in Ref. 5 with poly(U) (Sigma) at 160 g/ml as message and [ 3 H]phenylalanine as precursor during the linear period of incorporation. Translation factors were purified essentially as described by Skogerson (6) except for the substitution of a Mono Q cartridge (Pharmacia Biotech Inc.) for DEAE Sephadex. Yeast EF2 was nearly homogeneous by SDS-gel electrophoresis and silver staining and had a final specific activity in the diphtheria catalyzed ADP-ribosylation assay (6) of 2.5 pmol/g. Rat liver EF1 and EF2 were prepared as described in (7) and the wheat germ proteins as in Ref. 8. Ribosomes were prepared from all three sources by sedimentation three times through 0.5 M KCl, 20% sucrose, 10 mM MgCl 2 cushions and concentrations estimated using the figure of 18.6 pmol/A 260 unit. In vitro translation assays with purified ribosomes and translation factors were performed as described for S. cerevisiae (7) except that [ 3 H]Phe-tRNA was prepared as precursor at a specific activity of 2000 dpm/pmol. After 20-min incubation at room temperature the whole volume was applied to a minicolumn (Isolab GS-QS quick-sep) packed with 1.6 ml (settled volume) of Sephadex G-75 (Type 120 from Sigma) presoaked in Buffer B (Buffer A plus 150 mM NaCl). As soon as the sample entered the gel bed, 0.7 ml of Buffer B was added and eluate collected in minivials for counting.
Molecular Mapping of EF2 Mutations-Plasmid-dependent sordarinresistant mutants were spontaneously generated from the eft1⌬/eft2⌬ deletion strain YEFD12h harboring an episomal copy of EFT1 * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
(pEFT1URA) or EFT2 (YCpEFT2). 1 ϫ 10 7 and 1 ϫ 10 8 yeast cells were plated on SC medium containing 2% agar and sordarin at 5 g/ml and incubated at 29°C until colonies appeared (3-7 days). Plasmid DNA was recovered and transformed into the Escherichia coli strain DH5␣ (Life Technologies, Inc.) by established methods (4,12). Plasmid DNA was prepared from E. coli using Qiaprep spin columns (Qiagen) and transformed into the yeast strains YEFD12h/pEFT1URA or YEFD12h/ YCpEFT2 to demonstrate plasmid-dependent resistance to sordarin. Plasmids that conferred resistance were sequenced with an ABI Prism 373 DNA sequencer according to the manufacturer's recommendations (Applied Biosystems). Sequences were analyzed using Sequencher DNA analysis software (Gene Codes Corp.). The mutations in the sR1 and sR2 strains were identified by sequencing the PCR products. Based on the locations of the plasmid mutations, specific regions of approximately 0.4 kilobase pairs were amplified from genomic DNA using Klentaq (CLONTECH) according to the manufacturer's instructions. Following initial identification of a mutation, five additional independent PCR reactions were performed and the products sequenced to rule out the possibility of errors due to misincorporation.

RESULTS AND DISCUSSION
Sordarin as a Selective Inhibitor of Fungal Protein Synthesis-Exposure of intact S. cerevisiae to low concentrations of sordarin rapidly inhibits the incorporation of [ 3 H]threonine into proteins while having no effect on incorporation of uracil into RNA or adenine into DNA. 2 The IC 50 for this effect (6 g/ml) corresponds closely to the IC 50 values for growth inhibition of wild-type strains in YPAD medium. In contrast, sordarin has no effect on HeLa protein synthesis (leucine incorporation) or cell proliferation at concentrations of Ն100 g/ml. 2 In vitro translation experiments using S30 extracts of S. cerevisiae and a poly(U) message revealed that sordarin inhibits phenylalanine incorporation (IC 50 Ϸ30 ng/ml) while having no effect on similar systems from rabbit reticulocyte or wheat germ at levels of up to 100 g/ml. 2 Based on these findings it appears that sordarin is a specific inhibitor of protein synthesis in fungi while having no effect on other eukaryotic systems, a specificity within eukaryotes that is without precedent for known translation inhibitors.
Sordarin Binding Assay-Hydrolytic cleavage of sordarin to sordarose and sordaricin abolishes activity both for inhibition of whole cell growth and for in vitro translation. However activity is restored by replacement of the sugar with a short chain alkyl ether. L-793,422 (Fig. 1A), a sordaricin derivative with an isobutyl ether side chain, is 1000-fold more active in growth inhibition assays than sordarin (IC 50 of 6 ng/ml) with the wild-type strain. A tritiated analog (2,3ditrito-2-methyl propyl ether) of specific radioactivity high enough for a sensitive binding assay was prepared. With the addition of both Mg 2ϩ and GTP, or the nonhydrolyzable analog GTP␥S or GDP, to desalted S30 extracts, binding of Resolution of the S30 extract into a soluble S100 fraction and salt-washed ribosomes showed that both ribosomes and a soluble component are necessary for the reconstitution of binding. The S100 protein was concentrated by precipitation with 70% ammonium sulfate, desalted, and subjected to anion exchange chromatography. Binding activity, reconstituted with ribosomes and GTP, was recovered in the large protein peak eluted at 0.2-0.3 M KCl. Assay of EF2 by ADP-ribosylation and of EF3 by ribosome-dependent GTPase activity showed that they overlap within this peak. When EF2 and EF3 are resolved by CM Sepharose chromatography (6), the soluble binding component comigrates with EF2.
Fungal Specificity Conferred by EF2-The ability of EF2 and ribosomes purified from rat liver and from wheat germ to bind [ 3 H]-L793,422 in the presence of GTP was examined. In contrast to S. cerevisiae EF2, neither rat liver nor wheat germ EF2 shows any binding when tested with ribosomes from rat liver, wheat germ, or yeast. Nor does binding occur using S. cerevisiae ribosomes and either higher eukaryotic EF2. However, substantial binding is conferred by S. cerevisiae EF2 upon addition of ribosomes from either rat liver or wheat germ (Fig.  1A). Thus EF2 is the determinant of the observed fungal specificity of sordarin. Furthermore, fungal specificity in polymerization can be demonstrated in reconstituted translation assays dependent on added purified EF2. Polyphenylalanine formation from [ 3 H]Phe-tRNA as precursor by rat liver ribosomes plus rat liver EF1 is insensitive to sordarin when rat liver EF2 is used, but sensitive with yeast EF2. Incorporation by yeast ribosomes plus yeast EF1 and EF3 is sensitive with yeast EF2, reconstituting the sensitivity of the unfractionated system, but becomes insensitive when rat liver EF2 is used (Fig. 1B).
Stabilization of the EF2-Ribosome Complex-During the translocation cycle, GTP is bound by the EF2-ribosome complex, followed by an extremely rapid hydrolysis to EF2-GDP, and a conformational change that releases EF2 for the next round of translocation (13). Fusidic acid is a universal EFG and EF2 inhibitor that inhibits translocation by stabilizing the EF2-GDP-ribosome complex (14).  ). B, polymerization with purified components. Solid circles, rat liver ribosomes plus rat liver EF1 and rat liver EF2; open squares, rat liver ribosomes plus rat liver EF1 and yeast EF2; open circles, yeast ribosomes plus yeast EF1, yeast EF2, and yeast EF3; filled squares, yeast ribosomes plus yeast EF1, rat liver EF2, and yeast EF3. When present, as noted for each combination, components were used in the following quantities in 50-l reactions with 8 pmol of [ 3 H]Phe-tRNA: rat or yeast salt-washed ribosomes (0.05 A 260 units); rat EF1 (1 g of partially purified, free of EF2) and yeast EF1 (0.4 g, free of both EF2 and EF3); rat EF2 (0.2 pmol) or yeast EF2 (0.15 pmol); yeast EF3 (0.2 g). Incorporation for 5 min (22°C) was still linearly time-dependent and was approximately 2 pmol in each uninhibited system. Results are expressed as percentage of incorporation in the absence of sordarin for each combination.

EF2 and Inhibition of Fungal Protein Synthesis 3149
to EF2 plus ribosomes. In similar experiments sordarin increases the half-life of the GDP-EF2-ribosome complex (Fig. 2), from less than 0.5 min to approximately 6 min (t1 ⁄2 for fusidic acid ϳ10 min). Sordarin analogs that show little or no effect in vivo fail to cause stabilization in vitro. Fusidic acid levels up to 5 mM fail to compete with the binding of [ 3 H]L-793,422 to the S. cerevisiae EF2-ribosome complex. Thus the universal binding site for fusidic acid appears not to overlap the fungal specific sordarin binding site. Genetic Linkage of Sordarin Resistance-Genetic confirmation that fungal EF2 confers specificity to sordarin was demonstrated by analysis of mutant yeast strains resistant to sordarin. Spontaneous mutants resistant to sordarin were selected on SC medium containing 5 g/ml sordarin, and these appeared at a frequency of approximately 1 ϫ 10 Ϫ8 . Progeny from crosses of the sordarin-resistant strains sR1 and sR2 with strain YEFT2Dh (eft2⌬:TRP1) that is deleted for the genomic copy of the EFT2 gene were analyzed. Dissection of 20 tetrads from each cross showed that sordarin resistance was never associated with tryptophan prototrophy, confirming genetic linkage, and that a mutant allele of eft2 is responsible for sordarin resistance in both sR1 and sR2 strains.
Identification of Mutations That Confer Sordarin Resistance-To facilitate determining which substitutions can result in sordarin resistance, mutants conferring plasmid-dependent resistance were spontaneously generated by plating cells on solid medium containing sordarin, using eft1⌬/eft2⌬ deletion strains that harbor an episomal copy of EFT1 or EFT2. Fifteen mutant strains were chosen that range in resistance from having an IC 50 Х 10 g/ml sordarin, to an IC 50 Ն 100 g/ml (Table  I). Plasmids conferring sordarin resistance were recovered, clonally purified, and sequenced. A single base change causing an amino acid substitution was identified in each clone, with the exception of pSR7, which has a three base pair deletion, resulting in the loss of G790 (Table I). Based on the locations of the mutations in the episomal copies of the EFT1 and EFT2 genes, we used PCR to amplify specific regions of EFT2 from genomic DNA prepared from the original sR1 and sR2 mutants. Single base changes that result in amino acid substitutions in each mutant were identified. The amino acid changes conferring resistance to sordarin are clustered into three regions of the EF2 protein (Fig. 3). Amino acid sequence alignment of EF2 with its prokaryotic counterpart EFG demonstrates that the EF2 substitutions are located in regions with homology to domains I, III, and IV in EFG. These domains of EFG are thought to interact with one another during GTP hydrolysis and translocation (15)(16)(17). The sequence alignment also revealed that mutations in EF2 conferring resistance to sordarin are located in proximity to mutations in EFG that give rise to fusidic acid resistance (18,19) (Fig. 3). Since fusidic acid does not permeate yeast cells, translational sensitivity to fusidic acid was determined in S30 extracts prepared from each sordarin-resistant mutant. Many mutations that confer resistance to sordarin also confer resistance to fusidic acid (Table I), although levels of resistance to the two drugs are not closely correlated. The partial cross resistance of mutations and the proximity of mutations that confer sordarin and fusidic acid resistance in the alignment of EF2 with EFG support the biochemical evidence of mechanistic similarity of the two drugs. One or more of the EF2 residues altered in the resistant mutants may be located in the sordarin binding site, but there is not sufficient evidence, without cross-linking or other biochemical approaches, and a three-dimensional structure for EF2, to identify the binding site.
Sordarin belongs to a class of natural products with multiple    3. The amino acid sequences of EF2 and S. typhimurium EFG were aligned using the GAP alignment tool from the Genetics Computer Group, Version 7, April 1991, 575 Science Drive, Madison, WI. Regions of homology between the EF2 and EFG proteins are illustrated by solid bars; gaps in homology are depicted by breaks in the solid bars. EFG domains are indicated (16,17). Substitutions that result in fusidic acid resistance in EFG, and sordarin resistance in EF2, are indicated by arrows. The scale at the bottom of the figure is relative only to the aligned seqences of EFG and EF2.
EF2 and Inhibition of Fungal Protein Synthesis 3150 functional groups readily accessible to chemical modification. The addition of this class to the small number of known EF2 inhibitors (fusidic acid family, ADP-ribosylators, both universal for eukaryotic translocation) provides a powerful new tool for dissecting the EF2 and ribosomal sites involved in translocation. Furthermore, the ability to inhibit protein synthesis specifically in fungi with sordarin demonstrates the potential for exploiting EF2 as a target for developing novel antifungal agents at a time when resistance to current agents is increasing. These findings show that homology between essential proteins and processes in the host and pathogen does not necessarily exclude them as potential targets for selective chemotherapy.