The N Terminus of Eukaryotic Translation Elongation Factor 3 Interacts with 18 S rRNA and 80 S Ribosomes*

Elongation factor-3 (EF-3) is an essential fungal-specific translation factor which exhibits a strong ribosome-dependent ATPase activity and has sequence homologies that may predict domains critical for its role in protein synthesis, including a domain at the N terminus, which exhibits sequence homology with Escherichia coli ribosomal protein S5. A portion of the N terminus of Saccharomyces cerevisiaeEF-3 (spanning the S5 homology region) has been cloned, expressed, and purified from E. coli. UV cross-linking experiments revealed that the N-terminal EF-3 protein (N-term EF-3) can be specifically cross-linked to 18 S rRNA. Filter-binding assays confirmed these data, and also established that the interaction has aK d ∼238 nm. Additional evidence shows that N-term EF-3 is able to associate with yeast ribosomes and inhibit the ribosome-dependent ATPase activity of native EF-3. These data taken together suggest that at least one of the ribosome-binding sites of EF-3 is located at the N terminus.

Fungi are unique among eukaryotes in their requirement for a third soluble translation elongation factor, called elongation factor 3 (EF-3) 1 (1). EF-3 was first identified as a 125-kDa single polypeptide in Saccharomyces cerevisiae (2,3) and has subsequently been identified in other fungal species (3)(4)(5). Although the exact role of EF-3 remains to be elucidated, it reportedly stimulates aminoacyl-tRNA binding to the ribosomal A site (6) and may be involved in release of deacylated tRNA from the E site (7).
EF-3 possesses strong ribosome-dependent ATPase and GTPase activities (2,3). Examination of the domain structure of EF-3 reveals that it has several regions that bear homology to members of the ATP-binding cassette family of nucleotide binding proteins. It has been shown recently that point mutations within two of these ATP-binding motifs in EF-3 severely impair the ribosome-dependent ATPase activity (8). EF-3 also exhibits other sequence homologies that may predict structural domains critical for its role in protein synthesis (9). These include a C-terminal region with homology to a conserved sequence element in aminoacyl-tRNA synthetases, a basic Cterminal tail, and an extended domain at the N terminus that shares sequence homology with Escherichia coli ribosomal protein S5.
Ribosomal protein S5 interacts with 16 S rRNA (10 -12), therefore we wished to explore whether the S5 homology domain of EF-3 could interact with 18 S rRNA. This N-terminal region of EF-3 (N-term EF-3, amino acids 98 -388) was cloned with a His 6 tag, expressed, and purified from E. coli. UV crosslinking and filter-binding results indicate that the interaction between N-term EF-3 and 18 S rRNA is specific. N-term EF-3 associates with yeast ribosomes and inhibits the ribosome-dependent ATPase activity of native EF-3. These data suggest that at least one of the ribosome-binding sites of EF-3 may be at the N terminus, and also raises the possibility that the association of EF-3 with the fungal ribosome may be dependent upon RNA/protein interactions.

EXPERIMENTAL PROCEDURES
Cloning and Plasmid Constructions-A plasmid containing the S. cerevisiae rRNA operon was digested with XbaI and EcoRI to yield a 1414-bp fragment corresponding to nt 160 -1573 of S. cerevisiae 18 S rRNA. This fragment was cloned downstream of a T7 promoter in XbaI/EcoRI-digested pKSIIϩ (Stratagene) to yield pKS-18 S. A fragment containing nt 542-577 of E. coli 16 S rRNA was constructed by annealing the following two oligos: (sense) 5Ј-AATTCTGTAATACGA-CTCACTATAGTGCAAGCGTTAATCGGAATTACTGGGCGTAAAGC-GGTAC-3Ј; (antisense) 5Ј-CGCTTTACGCCCAGTAATTCCGATTAAC-GCTTGCACTATAGTGAGTCGTATTACAG-3Ј. The annealed DNA, containing EcoRI and KpnI sites at the 5Ј-and 3Ј-ends respectively, was cloned into pUC18 to yield pUC-16Sf. A plasmid encoding a subfragment of 18 S rRNA (nt 588 -625) downstream of a T7 promoter (pUC-18 S f ) was a gift from Ian Jeffrey (St. George's Hospital, London), and pECM1RNA, containing the E. coli RNase P (M1) RNA sequence downstream of a T7 promoter, was a generous gift from Lisa Hegg (SB Pharmaceuticals).
Cloning and Expression of N-term EF-3-DNA corresponding to nucleotides 292-1168, encoding the S5 homology domain of EF-3 (amino acids 98 -388), was amplified directly from yeast genomic DNA by a polymerase chain reaction using the following primers: (sense) 5Ј-CGG-GATCCAACGCAGGTAACAAGGAC-3Ј; (antisense) 5Ј-GGGGTACCG-GTGATGTGCCTGAACCA-3Ј. The polymerase chain reaction fragment was cloned into the BamHI/KpnI sites of the 6 ϫ His-expression vector pQE-30 (Qiagen). The resulting plasmid, pQE-NtermEF3, was transformed into competent E. coli M15 (pRep4) cells according to the directions of the QIAexpress Kit (Qiagen). 1 liter of bacterial cells containing the plasmid was induced with isopropyl-1-thio-␤-D-galactopyranoside, and the overexpressed protein was purified under denaturing conditions and then subsequently renatured following the manufacturer's instructions.
Purification of Native EF-3-EF-3 was purified from S. cerevisiae cells (strain ABYS1) essentially as described previously (13). Cells were grown in YEPD media overnight, harvested, and lysed by passing through a French Press (SLM Instruments, Inc.) five times at 1100 p.s.i. The supernatant was separated from cell debris by centrifuging at 15,000 rpm for 15 min. The supernatant was loaded onto a heparin-Sepharose column (Pharmacia Biotech Inc.). The peak fractions were pooled and loaded onto an ATP-agarose column (Sigma). The column was washed sequentially with Buffer C (20 mM HEPES, pH 7.0, 25 mM KCl 2 mM Mg(OAc) 2 , 1 mM (NH 4 ) 2 SO 4 , 0.8 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine), buffer C ϩ 2 mM NAD, and buffer C (100 mM KCl) ϩ 2 mM NAD. A gradient of buffer C * 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 4 Cl, 100 mM ␤-mercaptoethanol). 100 l of a 10 mg/ml solution of lyticase were added to 30 g of cell paste and left on ice for 3.5 h. 40 ml of buffer A were then added, and the cells were lysed by passage through a French press. The lysate was then centrifuged at 17,000 rpm for 40 min in an SS-34 rotor to remove the cell debris. The resulting supernatant was then layered onto a sucrose cushion containing buffer B (20 mM HEPES, pH 7.5, 10.5 mM Mg(OAc) 2 , 0.5 mM EDTA, 0.5 M NH 4 Cl, 10 mM ␤-mercaptoethanol, and 1.1 M sucrose) and centrifuged at 40,000 rpm in a Beckman 60Ti rotor at 4°C for 16 h. The ribosome pellet was rinsed with 2ϫ buffer B, and then resuspended in 1ϫ buffer B containing 50% glycerol. Ribosomes were stored at Ϫ20°C.
Synthesis of Labeled RNA-Radiolabeled RNAs were synthesized by in vitro transcription of pKS-18 S rRNA linearized with HindIII using T7 RNA polymerase and [␣-32 P]UTP. Nonspecific control RNAs were synthesized from pECM1RNA linearized with KpnI and pUC18-18 S f linearized with HindIII. Unlabeled RNAs were synthesized using ME-GAshortscript kit (Ambion Inc., Austin, TX).
UV Cross-linking and Filter-binding Assays-In vitro binding reactions were carried out using 200 ng of purified N-term EF-3 in a reaction volume of 30 l containing 50 mM Tris-HCl, pH 7.6, 2 mM MgCl 2 , and 100 mM KCl. Reaction mixtures containing 5 fmol of radiolabeled RNA (with or without unlabeled competitor RNAs) were incubated at 30°C for 20 min, then transferred to Parafilm on ice and exposed to UV light (254 nm) at a distance of 4 cm for 10 min. Following treatment with RNase A (final concentration 1 mg/ml) at 37°C for 15 min, the samples were resolved by SDS-PAGE on 10% gels. For filterbinding experiments, binding reactions were incubated for the indicated times at 30°C and then filtered immediately on 0.45 M nitrocellulose filters (Whatman), washed with 1.0 ml of binding buffer, and then counted in a scintillation counter.
ATPase Activity Assays-A colorimetric assay adapted from Chan et al. (14) was used to assess ribosome-dependent EF-3 ATPase activity with the following modifications. Each 40 l of assay contained 2.4 pmol of pure native EF-3, 4.8 pmol of 80 S ribosomes, 12 nmol of ATP, and different amounts of N-term EF-3 in 1ϫ assay buffer (50 mM Tris-Cl, pH 8.5, 100 mM KCl, 10 mM MgCl 2 ). After a 30-min incubation at 30°C, the reaction was terminated by the addition of 75 l of malachite green mixed reagent, and incubation was continued at room temperature for 5 min. The plates were then read at 655 nm in an enzyme-linked immunosorbent assay plate reader.
Binding of N-term EF-3 to Yeast 80 S Ribosomes-5 pmol of yeast 80 S ribosomes were incubated with native or N-term EF-3 in buffer B (25 mM Tris-Cl, pH 7.5, 10 mM Mg(OAc) 2 , 50 mM NH 4 Cl, 1 mM dithiothreitol, 3% glycerol) in a total volume of 30 l for 10 min at 30°C. To detect proteins that bound to ribosomes, reaction mixtures were layered onto 1-ml cushions of buffer D (same as buffer B, containing 10% glycerol) and centrifuged for 1 h at 45,000 rpm in a Beckman TLA 100.3 rotor at 4°C. Pelleted material was washed once with 1.0 ml of buffer D, resuspended in 30 l of SDS-PAGE sample buffer, and subjected to SDS-PAGE on a 10% gel. Separated proteins were transferred to polyvinylidene difluoride membrane for Western blot analysis using the ECL reagent kit (Amersham Corp.). ␣-EF-3 antibody was kindly provided by Dave Colthurst (Canterbury, UK), and RGS His antibody (against the His 6 tag) was purchased from Qiagen.

RESULTS
Analysis of the amino acid sequence of EF-3 has indicated several evolutionarily conserved functional domains that may predict important features critical for protein synthesis (9), including a domain at the N terminus that exhibits some primary sequence homology to the ribosomal protein S5 from E. coli and Bacillus stearothermophilus. An alignment of the N terminus of EF-3 with B. stearothermophilus S5 revealed 26% identity and 55% similarity between the two proteins (9), including a region of S5 predicted to interact with rRNA (15). Based upon these sequence similarities, as well as studies on the interaction between S5 and 16 S rRNA (10 -12), we were interested to determine if this region within the N-terminal domain of EF-3 could interact with 18 S rRNA in an analogous manner. We cloned the N terminus of EF-3 (amino acids 98 -388) as a His-tagged fusion protein, and expressed the resultant protein in E. coli. Fig. 1 shows the purification of the protein (referred to here as N-term EF-3) under denaturing conditions. By SDS-PAGE and Coomassie blue staining, we judge the protein preparation to be highly enriched for N-term EF-3, with only minor smaller molecular weight contaminants.
To examine whether N-term EF-3 was capable of interacting with 18 S rRNA, a UV cross-linking assay was used (Fig. 2). This analysis revealed an RNA/protein cross-link which migrated at ϳ35 kDa on the gel, in accordance with the size of the recombinant protein (Fig. 3, lane 1). Addition of 5-, 10-, and 250-fold molar excess of unlabeled nonspecific competitor RNAs indicated that fragments of E. coli 16 S rRNA (nt 542-577, lanes 8 -10) and S. cerevisiae 18 S rRNA (nt 588 -625, lanes 5-7) were unable to compete for the cross-link, while addition of unlabeled specific competitor 18 S rRNA (lanes 2-4) did effectively compete. These results indicate that N-term EF-3 is able to form a cross-link with 18 S rRNA, and that this interaction is specific.
In an attempt to further characterize the interaction between N-term EF-3 and 18 S rRNA, a filter-binding assay was established. Binding reactions were performed as described under "Materials and Methods." These results, shown in Fig. 3, demonstrate that the addition of 125-fold molar excess of unlabeled specific competitor 18 S rRNA results in a 66% decrease in binding of N-term EF-3 to labeled 18 S rRNA, while the addition of other nonspecific RNAs does not appreciably compete binding. These results are in good general agreement with the UV cross-linking data shown in Fig. 2, and further supports the notion that the interaction between N-term EF-3 and 18 S rRNA is specific. An additional filter-binding experiment was done using a range of purified N-term EF-3 in molar excess of labeled 18 S rRNA (Fig. 4). The results indicate the apparent K d to be approximately 238 nM.
The results described above suggest that the N terminus of EF-3 interacts specifically with 18 S rRNA, suggesting that native EF-3 may interact with ribosomal RNA via its N-terminal domain. To directly address the question whether N-term EF-3 could interact with ribosomes, purified native and N-term EF-3 were incubated with yeast 80 S ribosomes. Factors associated with ribosomes were sedimented through a glycerol cushion, and resolved via SDS-PAGE. An immunoblot analysis is shown in Fig. 5, and reveals that native EF-3 (Fig. 5A, lane  2) and N-term EF-3 (Fig. 5B, lane 5) do not pellet through the glycerol cushion in the absence of 80 S ribosomes. As expected, native EF-3 associates with yeast 80 S ribosomes (Fig. 5A, lane   4). A similar experiment was done using increasing amounts of N-term EF-3 bound to ribosomes. Fig. 5B shows that N-term EF-3 is able to specifically associate with yeast 80 S ribosomes (lanes 3 and 4), further supporting the hypothesis that the N terminus of EF-3 can interact with ribosomes.
Because the ATPase activity of EF-3 is enhanced by two orders of magnitude in the presence of yeast ribosomes (3,6), it is likely that this activity requires a direct contact between EF-3 and ribosomal proteins and/or 18 S rRNA. We next asked whether the ribosome-dependent ATPase activity of native EF-3 could be inhibited by the addition of excess N-term EF-3. When increasing amounts of N-term EF-3 were added to the reaction, a concomitant decrease in the ribosome-dependent ATPase activity of native EF-3 was observed (Fig. 6); For example, at a 25-fold molar excess of N-term EF-3 to native EF-3, we observed almost a 50% decrease in the ATPase activity of native EF-3 stimulated by yeast ribosomes. This dramatic inhibition of ATPase activity was not seen in control experiments when excess bovine serum albumin was added to the reaction (Fig. 6). The results of these experiments suggest that N-term EF-3 is able to inhibit the ribosome-dependent ATPase activity of native EF-3, presumably by competing for the EF-3 binding site(s) on ribosomes or rRNA.  6. N-term EF-3 inhibits ribosome-dependent ATPase activity of native EF-3. Purified, native EF-3 was incubated with S. cerevisiae 80 S ribosomes with increasing amounts of N-term EF-3 (fOOf) or bovine serum albumin (OEOOOE), and ATPase activity was determined by a colorimetric assay (see "Experimental Procedures.").

DISCUSSION
Sequence comparisons between E. coli ribosomal protein S5 and S. cerevisiae EF-3 have revealed some degree of homology, particularly with the N-terminal domain of EF-3. Because S5 has been shown to interact with 16 S rRNA (10 -12), we wished to test the possibility that EF-3 might make contact with 18 S rRNA via its N terminus. A 291-amino acid region of EF-3 (N-term EF-3) encompassing the S5 homology domain was expressed as a His-tagged protein in E. coli, and two independent biochemical assays were used to show a specific interaction between 18 S rRNA and N-term EF-3. That a specific RNA/ protein cross-link could be induced by UV light is intriguing, particularly in light of the fact that this region of EF-3 lacks any of the canonical motifs common to many RNA-binding proteins (16). However, the failure to detect these motifs should not be taken to reflect an inability to interact with RNA. More work is needed to precisely identify and characterize the amino acid residues of N-term EF-3 involved in making contact with RNA.
Previously, Kovalchuke and Chakraburtty (17) reported an interaction between EF-3 and polynucleotides, with preferential binding to poly(G). While they showed that both 18 S and 26 S rRNA could effectively inhibit activation of EF-3 ATPase by yeast ribosomes, they did not examine whether EF-3 could interact directly with either of these RNAs in an isolated system. The results of our RNA-binding experiments demonstrate a physical interaction between N-term EF-3 and 18 S rRNA that is specific. While we did not examine the binding of N-term EF-3 to 26 S rRNA, it is possible that different domains of EF-3 are required for contacting different regions of the ribosome (see below).
Because the C terminus of EF-3 is very basic and contains three clusters of lysine residues, it has been speculated that EF-3 may interact with yeast ribosomes via this region. In support of this, Uritani et al. (18) reported that a monoclonal antibody specific for the C-terminal region of EF-3 inhibits its ribosome-dependent ATPase activity. Other work has reported that a GST-fusion protein containing the C-terminal domain of EF-3 remains associated with yeast ribosomes, suggesting that least one of the ribosome-binding sites of EF-3 resides within this region (19). Our observations that the N-terminal domain of EF-3 binds 18 S rRNA and also associates with ribosomes do not necessarily conflict with the results described above with the C-terminal domain. The two results can be easily reconciled by taking into account the fact that multiple contacts may exist between EF-3 and ribosomes or rRNA. It may be possible that some regions of EF-3 (like the N terminus) may interact directly with rRNA, while other domains are responsible for making protein-protein contacts. Further dissection of the putative protein and RNA-binding domains of EF-3 will help to elucidate the complex nature of the interactions described above, and may provide additional information regarding the functions of EF-3 in fungal translation.