Ribosomal P-protein Stalk Function Is Targeted by Sordarin Antifungals*

Sordarin derivatives are remarkably selective inhibitors of fungal protein synthesis. Available evidence points to a binding site for these inhibitors on elongation factor 2, but high affinity binding requires the presence of ribosomes. The gene mutated in one of the two isolated complementation groups ofSaccharomyces cerevisiae mutants resistant to the sordarin derivative GM193663 has now been identified. It is RPP0, encoding the essential protein of the large ribosomal subunit stalk rpP0. Resistant mutants are found to retain most of the binding capacity for the drug, indicating that mutations in rpP0 endow the ribosome with the capacity to perform translation elongation in the presence of the inhibitor. Other proteins of the ribosomal stalk influence the expression of resistance, pointing to a wealth of interactions between stalk components and elongation factors. The involvement of multiple elements of the translation machinery in the mode of action of sordarin antifungals may explain the large selectivity of these compounds, even though the individual target components are highly conserved proteins.

Sordarin derivatives are remarkably selective inhibitors of fungal protein synthesis. Available evidence points to a binding site for these inhibitors on elongation factor 2, but high affinity binding requires the presence of ribosomes. The gene mutated in one of the two isolated complementation groups of Saccharomyces cerevisiae mutants resistant to the sordarin derivative GM193663 has now been identified. It is RPP0, encoding the essential protein of the large ribosomal subunit stalk rpP0. Resistant mutants are found to retain most of the binding capacity for the drug, indicating that mutations in rpP0 endow the ribosome with the capacity to perform translation elongation in the presence of the inhibitor. Other proteins of the ribosomal stalk influence the expression of resistance, pointing to a wealth of interactions between stalk components and elongation factors. The involvement of multiple elements of the translation machinery in the mode of action of sordarin antifungals may explain the large selectivity of these compounds, even though the individual target components are highly conserved proteins.
The ribosomal stalk is a defined morphological feature most evident in prokaryotic ribosomes, where it is composed of the L7/L12 tetramer, L10 and L11 proteins, and a highly conserved domain of the 23 S rRNA termed the "GTPase center." The stalk plays a prominent role in the elongation phase of ribosomal protein synthesis (see Refs. 1 and 2 for a review).
Equivalent structures have been visualized on eukaryotic ribosomes (3) and functional homologies with the bacterial components have been described (see Ref. 4 for a review). The recent nomenclature proposal for Saccharomyces cerevisiae ribosomal proteins (5) will be followed throughout this manuscript. In this organism, proteins rpP1␣, rpP1␤, rpP2␣, and rpP2␤ take the place of the bacterial L7/L12 tetramer. rpP0 corresponds functionally to bacterial L10 and rpL12 to bacterial L11 (6 -10). rpP0 is thought to play a central role in stalk architecture, interacting with rpL12 and 26 S rRNA at its amino-terminal half and with the other four P-proteins at its carboxyl-terminal end (4,11). None of the P1/P2 proteins are absolutely essential for growth, although in the absence of at least one member of each class (P1 and P2), vegetative growth is strongly retarded (12).
The molecular detail of the interplay between stalk components and elongation factors is unknown. Recent morphological evidence (13) indicates that elongation factors make multiple contacts with stalk components, and a detailed footprinting analysis of EF-G on rRNA place the factor at the base of the stalk (14). Resolution is however not high enough to ascribe roles to individual molecules.
Semisynthetic derivatives of the natural product sordarin (15,16) have proven to be highly potent inhibitors of eukaryotic protein synthesis, with remarkable selectivity for the fungal translation machinery (17,18). Sordarin inhibitors seem to bind to elongation factor 2 1,2 (18), and point mutations on the factor make cells resistant to the inhibitors 1 (18), but high affinity binding requires the presence of ribosomes, 2 and resistant mutants were detected, which did not map on EF2. 1 This work describes the gene mutated in a second complementation group of S. cerevisiae mutants resistant to GM193663, a potent sordarin derivative. The gene encodes the ribosomal stalk protein rpP0. It is further shown that other stalk components are also involved in the translation elongation function inhibited by GM193663, substantiating the ribosomal requirement for binding of the inhibitor to EF2 and pointing at interactions between the factor and specific stalk components.
In Vitro Activity and Binding Assays-Growth inhibitory activity of GM193663 was determined on liquid medium by the antibiotic 2-fold serial dilution technique; 100 l of YPD were inoculated with 10 5 colonyforming units/well. IC 50 was defined as the minimal concentration of compound that inhibited 50% of the control growth after 24 h of incuba-* 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. tion at 30°C. For strains living with heterologous rpP0 proteins, A 600 was measured after 48 h of incubation because of their slower growth. Cycloheximide (Sigma) and GM193663 stocks were prepared in Me 2 SO.
Cell-free S50 extracts were prepared from cells grown in YPD medium to an A 600 ϭ 1.5. Washed cells were broken in lysis buffer (30 mM Hepes/KOH, 100 mM potassium acetate, 12.5 mM magnesium acetate, 2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 8.5% D-mannitol, pH ϭ 7.4) by vortexing with glass beads for 10 1-min periods, with intervening 1-min cooling in ice. Unbroken cells and debris were pelleted at 4,000 ϫ g for 10 min, and the S50 supernatant was obtained by centrifugation at 50,000 ϫ g for 30 min.
[ 3 H]Sordarin was labeled at 180 GBq/mmol in the aldehyde group by the Isotope Chemistry Group at Glaxo Wellcome (Stevenage, United Kingdom). Binding assays were performed with 0.5 mg of protein from S50 extracts and 0.1 g of [ 3 H]sordarin (36 kBq, 20 nmol), in 80 mM Hepes/KOH, 1 mM dithiothreitol, 10 mM magnesium acetate, in a final volume of 500 l. This amount of sordarin was enough to have at least 90% of the input compound free at equilibrium. After 30 min of incubation at room temperature, 400 l from each sample were applied to PD-10 Sephadex-G 25 M columns (Amersham Pharmacia Biotech), presoaked in lysis buffer without D-mannitol. [ 3 H]Sordarin bound to excluded macromolecules was measured by scintillation counting. All incubations were done in duplicate tubes and the results averaged. Measurements from paired tubes never differed by more than 20%. Counts obtained in parallel incubations containing a 100-fold excess of cold sordarin were subtracted from all data points. This nonspecific binding was always less than 10% of the total binding detected.
Molecular Mapping of rpP0 Mutations-Localization of mutations that conferred resistance to GM193663 within the rpP0 open reading frame was carried out by the single strand conformation polymorphism technique (22). Briefly, standard PCRs were performed using cloned rpP0 from all mutant and wild type strains as template. Three overlapping fragments covering the entire open reading frame were amplified and electrophoresed under denaturing conditions. PCR fragments showing an altered mobility compared with their corresponding wild types were sequenced using an Applied Biosystems PRISM 310 genetic analyzer.

Mutations in Ribosomal Protein P0 Confer
Resistance against Sordarin Antifungals-It has been communicated previously that genetic analysis of spontaneous mutants resistant to the sordarin antifungal GM193663, an inhibitor of translation elongation, identified two different genetic complementation groups, denominated FPR1 and FPR2. 1 The major one, FPR1, grouped mutations in EFT2, encoding translation elongation factor 2. FPR2 was composed of just four mutants and, contrary to expectations, did not arise from EFT1, the second EF2-coding gene in S. cerevisiae. Growth and resistance characteristics of the mutants in rich medium are presented in Table I. All mutants displayed cross-resistance on plates to other members of the sordarin family, but not to the protein synthesis inhibitors cycloheximide, hygromycin, or verrucarin A, showing that sordarins act by a mechanism different from the one used by those other inhibitors.
Protein synthesis in a cell-free extract from an FPR2 mutant also displayed a reduced sensitivity toward GM193663 inhibition, 1 thereby ruling out a transport mechanism of resistance.
To investigate the identity of the gene mutated in the FPR2 group, an educated guess was made considering that the most likely possibilities were genes encoding macromolecules that functionally interacted with EF2. Candidates included proteins of the large ribosomal subunit stalk, namely rpP0, the acidic P-proteins, and the rpL12 protein (formerly rpL15) located on the GTPase center (6 -8, 10, 23). rRNA was an unlikely candidate given the semidominant nature of the mutations and the large number of rRNA genes present in the cell. Strains genetically tagged at the loci encoding stalk proteins rpP0, rpP1␣, rpP1␤, rpP2␣, and rpP2␤ were obtained from the laboratory of J. P. G. Ballesta (12) and crossed to GM193663 R mutant FPR2-15, which displayed the most dominant resistance phenotype. Tetrad analysis did not detect a genetic linkage between the resistance locus and the P-proteins, except in the case of rpP0, where a tight genetic linkage between the resistance mutation and RPP0, the rpP0 locus, was found. All 36 tetrads dissected displayed the parental ditype phenotype. This result strongly suggested that the resistance and RPP0 loci were very close to each other or were the same one. RPP0 was amplified and cloned from the four mutants in group 2 as well as from their corresponding parental strains. The cloned genes were transformed into a sensitive haploid strain, and the transformants were tested for their sensitivity to GM193663. Strains transformed with an RPP0 gene from a resistant strain became resistant themselves in all cases, confirming that the gene mutated in the second complementation group was RPP0. Thus, mutations in protein rpP0, an essential protein of the large ribosomal subunit, can confer a level of resistance to sordarins higher than the one afforded by mutations in EF2, the proposed binding protein for sordarins (Table  II). This could be the first such occurrence described, and it strongly suggests that rpP0 and EF2 interact functionally.
Resistance Mutations Cluster Together in the rpP0 Sequence-Chromosomal RPP0 genes from the four mutants were cloned by gap repair and analyzed by single-stranded conformational polymorphism (see "Experimental Procedures"). Mobility changes were found in the central one-third of all alleles. The region was sequenced on both strands, and amino acid changes were detected in all mutants (Table I, Fig.  1). To confirm that the observed changes were not introduced by the PCR, genomic fragments from the four mutants were recovered by the plasmid gap repair technique (24). A wild type RPP0 gene was gapped by removing all the open reading frame  except the last 86 amino acids and then transformed into the resistant mutants. Plasmids were recovered from stable transformants, sequenced, and re-transformed into a wild type strain. All recovered plasmids had the expected mutation and conferred resistance to GM193663, establishing that the observed amino acid changes (Table I) were responsible for the resistance to the inhibitor. Unexpectedly, spontaneous mutant FPR2-17 had four nucleotides changed, leading to two consecutive substitutions in the amino acid sequence of the protein.
Comparison of the sequence of the rpP0 region, where the GM193663 R mutations were found across different phyla (Fig.  1), showed that this domain is a highly conserved one and hence presumably important for the function of rpP0. It is thus somewhat surprising that the mutations do not impede the normal function of the protein, as indicated by the nearly wild type growth rate of the mutant strains in the absence of the inhibitor (Table I). In fact, there seems to be a lot of flexibility in the primary sequence requirements for rpP0 function. The group of J. P. G. Ballesta has shown that the yeast ribosome can function with P0 proteins from different species. 4 Using their strains, we have found that GM193663 has a higher IC 50 for growth inhibition in S. cerevisiae cells carrying rpP0 proteins from Dictyostelium or rat than in cells expressing wild type rpP0 (Table III). This confirms that rpP0 plays a role in the mode of action of sordarin compounds and shows that it is at least partly responsible for the selectivity of these inhibitors.
Mutations in rpP0 Do Not Abolish Binding of the Inhibitor to Macromolecules-Sordarin is a close analogue of GM193663 (Fig.  2) and both compounds display cross-resistance and compete for binding to macromolecules. 2 Fig. 2 shows [ 3 H]sordarin binding to macromolecules in the rpP0 mutants. Substantial binding was found in all mutant strains, showing that growth, and therefore protein synthesis, can take place in the presence of nearly wild type levels of inhibitor binding in some strains. Data from an EF2 mutant have been included for comparison, and it can be seen that in this case resistance is accompanied by the total loss of sordarin binding capacity, as described previously 1 (Fig. 2). The measured binding is to the ribosomal fraction of the extracts plus associated cytoplasmic factors, because the post-ribosomal supernatant showed negligible binding (data not shown). Therefore, the mechanism of resistance mediated by the rpP0 protein is different from the simple loss of affinity for the drug, and it may be related to the molecular details of the interaction between EF2 and the ribosomal stalk.
Other Proteins of the Ribosomal Stalk Modulate Resistance Conferred by rpP0 -Given the known or inferred interactions between the acidic P-proteins and rpP0 (see Ref. 4 for a review), it was of interest to know if lack of the former affected resistance due to rpP0, which would substantiate the relevance of the interactions between P-proteins in translation elongation.
All four nonessential P-proteins were interrupted singly and in pairs in the FPR2-15 mutant background. It was found that deleting either rpP1␣ or rpP2␤ reduced resistance levels more than 10-fold. Deletion of either rpP1␤ or rpP2␣ did not have a significant effect. Deleting both rpP1␣ and rpP2␤ had an additive effect, whereas deleting the other pair did not affect resistance (Table II). The effect of the protein composition of the ribosomal stalk on the sensitivity toward GM193663 is observed in the mutant RPP0 background only. The same deletions do not affect the sensitivity of a wild type strain (data not shown). These results confirm that the four P-proteins are not functionally equivalent, with rpP1␣ and rpP2␤ establishing a  H]sordarin to macromolecules in cell-free extracts from strains with GM193663 R mutations in rpP0. Numbering refers to the different mutants in complementation group 2 as detailed in Table I. Binding data from each mutant or group of mutants (shaded bars) are arranged following their corresponding parental wild type strain (white bars). A mutant from the first complementation group (EF2 mutants) was included for comparison. No counts above background were detected in this case, and a horizontal line has been added to symbolize this fact. Results from three independent experiments are averaged. Error bars correspond to the S.E. of the mean. See "Experimental Procedures" for a description of the binding assay. type of interaction with rpP0, and probably EF2, different from the one maintained by rpP1␤ or rpP2␣.
The influence of the stalk on the sensitivity toward sordarin antifungals seems to be specific for this type of translational inhibitors, because deleting the acidic P-proteins, either on a wild type strain or in the FPR2-15 mutant, did not change their sensitivity toward a different elongation inhibitor, cycloheximide. However, GM193663 resistance mediated by a mutant EF2 was also diminished by deleting rpP1␣ and to a lesser extent by deleting rpP2␤ (Table II). This is in agreement with the findings presented above and further indicates that the soluble factor establishes functional interactions with the ribosomal stalk, and not only with rpP0, but also with rpP1␣ and perhaps with rpP2␤ as well. DISCUSSION Sordarins are a new family of antifungal compounds and, together with tenuazonic acid (25), the only protein synthesis inhibitors reported to show selectivity between different eukaryotes (17,18). They have been described to target and bind to elongation factor 2 1,2 (18), but it was also immediately apparent that the inhibitor's mode of action involved additional cellular components, because resistant mutants were obtained outside EF2 1 , and high affinity binding of the drug to macromolecules required the presence of ribosomes, 2 suggesting that the binding site on EF2 is fully formed only after the factor interacts with a ribosome. In this work, experiments are described showing that rpP0, an essential component of the ribosomal large subunit stalk, plays a large role in determining the sensitivity to sordarin antifungal GM193663, as single point mutations in rpP0, and substitution of heterologous rpP0 proteins for the endogenous one, reduce measurably the sensitivity of yeast toward these compounds (Tables II and III). The results also indicate that there is a functional interaction between rpP0 and EF2, because mutations in any one of them render cells insensitive to the same translational inhibitor. This fits well with evidence indicating that elongation factors are probably recruited to the ribosome through interactions with the stalk (7,9,13,26).
The mechanism of resistance afforded by mutations in rpP0 is intriguing, and its elucidation may shed light into the molecular details of ribosomal translocation. The binding capacity for the inhibitor has been retained to a large extent by the resistant mutants (Fig. 2). If the hypothesis that sordarin antifungals impede conformational changes in EF2 is correct 1 (18), then changes in rpP0 could somehow force or facilitate the appropriate bending of the EF2 molecule. According to the data presented here, in order for rpP0 to be able to "help" EF2 when blocked by GM193663, the protein must be part of a ribosomal stalk in which rpP1␣ and rpP2␤ are present. Stalk structures deficient in either of those proteins substantially reduce the capacity of mutant rpP0 to overcome the presence of the inhibitor. Interestingly, rpP1␣ and rpP2␤ are also the P-proteins whose deletion has the largest effect on growth rate (12). rpP1␣ may be able to establish some interaction with EF2 by itself, because lack of this ribosomal protein also reduces significantly resistance levels afforded by mutations in EF2 (Table II). There seems to be also an effect of the rpP2␤ deletion on EF2-mediated resistance, although the reduction in the IC 50 of the inhibitor is smaller in this case. All this hints at a wealth of functional interactions between elongation factors, or at least EF2, and components of the ribosomal P-protein stalk. Pairwise physical interactions between all these proteins are currently being analyzed by genetic and biochemical means.
The screen for GM193663 R mutants is by no means satu-rated. Ongoing work at several laboratories will possibly detect additional gene products involved in the mode of action of sordarin antifungals, which may in turn uncover new functional interactions, or confirm suspected ones, between components of the translation apparatus. In yeast there are described ribosomal protein mutations involved in resistance to cycloheximide, cryptopleurine, trichodermin, and now sordarin analogues (27)(28)(29)(30)(31). Sordarin antifungals may however be unique in targeting the functional interaction between a soluble translation factor and ribosomal proteins, in such a manner as to allow both types of proteins to influence the cell's sensitivity to the inhibitor. The contribution of multiple elements to sensitivity may explain the large selectivity of these compounds, despite the high sequence conservation between the individual components of the eukaryotic translation apparatus. Sordarin antifungals are very selective drugs with a good therapeutic potential, and they are also turning out to be useful tools for the analysis of the elongation cycle during protein translation.