Chaperone Properties of Bacterial Elongation Factor EF-G and Initiation Factor IF2*

Elongation factor G(EF-G) and initiation factor 2 (IF2) are involved in the translocation of ribosomes on mRNA and in the binding of initiator tRNA to the 30 S ribosomal subunit, respectively. Here we report that the Escherichia coli EF-G and IF2 interact with unfolded and denatured proteins, as do molecular chaperones that are involved in protein folding and protein renaturation after stress. EF-G and IF2 promote the functional folding of citrate synthase and α-glucosidase after urea denaturation. They prevent the aggregation of citrate synthase under heat shock conditions, and they form stable complexes with unfolded proteins such as reduced carboxymethyl α-lactalbumin. Furthermore, the EF-G and IF2-dependent renaturations of citrate synthase are stimulated by GTP, and the GTPase activity of EF-G and IF2 is stimulated by the permanently unfolded protein, reduced carboxymethyl α-lactalbumin. The concentrations at which these chaperone-like functions occur are lower than the cellular concentrations of EF-G and IF2. These results suggest that EF-G and IF2, in addition to their role in translation, might be implicated in protein folding and protection from stress.

The elongation phase of protein synthesis is promoted by two G proteins, elongation factor EF-Tu, 1 which delivers aminoacyl tRNAs to the ribosome, and EF-G, which catalyzes the translocation step, during which the A-and P-site tRNAs move to the P and E sites of the elongating ribosome, respectively, and mRNA is advanced by one codon (1)(2)(3). EF-G binds to the ribosome in its GTP form, hydrolyzes GTP to drive tRNA movement on the ribosome (4), and is released in its GDP form. The functional cycle is completed upon GDP release and reactivation of the empty factor by binding of a GTP molecule (1)(2)(3).
IF2 is the only G protein among the three translation initiation factors in Escherichia coli. It promotes the binding of fMet-tRNA f Met to the 30 S ribosomal subunit leading to the formation of the 30 S initiation complex (for review see Refs. 5 and 6). IF2, together with IF3, helps the 30 S subunit to select the initiator tRNA over other tRNAs (7). Assembly of the 70 S initiation complex is accompanied by IF2-dependent hydrolysis of GTP and release of IF2 from the 70 S complex, a step that is essential for the first peptide bond formation (8).
In addition to their role in translation, several ribosomal proteins and translational factors are involved in other mechanisms, including replication, transcription, RNA processing, DNA repair, regulation of translation, malignant transformation, and regulation of development (reviewed in Ref. 9). IF2 is suspected to participate in the activation of transcription of the rrnB operon (10) and displays DNA binding activity (11). EF-Tu interacts with the Q␤ replicative complex, the transcriptional apparatus, and membranes (reviewed in Ref. 12); EF-1␣ (the eukaryotic counterpart of EF-Tu) and EF-2 (the eukaryotic counterpart of EF-G) bind to actin filaments and influence the assembly of cytoskeletal polymers (13)(14)(15). Both EF-1␣ and EF-Tu apparently participate in the degradation of N-terminally blocked proteins by the 26 S proteasome (16), and EF-Tu possesses chaperone-like properties in protein folding (17,18) and displays protein disulfide isomerase activity (19). The chaperone properties of elongation factor EF-Tu (17,18) and of ribosomes (20) prompted us to check whether EF-G and IF2 also possess chaperone properties.
Molecular chaperones form a class of polypeptide-binding proteins that are implicated in protein folding, protein targeting to membranes, protein renaturation or degradation after stress, and the control of protein-protein interactions. They can distinguish native proteins from their non-native forms, owing to the specificity of their peptide binding site, and they catalyze protein folding and renaturation in vitro (reviewed in Refs. [21][22][23]. The major classes of bacterial chaperones comprise DnaK⅐hsp70 (and its assistants DnaJ and GrpE), GroEL⅐hsp60 (and its assistant GroES), HtpG⅐hsp90, and the small heat shock proteins (21)(22)(23). In the present study, we show that EF-G and IF2, in a manner similar to that of molecular chaperones, increase the refolding of unfolded proteins, protect proteins against thermal denaturation, and form complexes with unfolded proteins. We propose that, in addition to their role in translation, EF-G and IF2 might assist in protein folding and renaturation in the cytoplasm.
Purification of EF-G, IF2, and DnaK-EF-G was purified by covalent chromatography on thiol Sepharose (24). Crude extracts from the E. coli K12 strain C600 (leuB6 thi-1thr-1 supE44) were prepared by a lysozyme/EDTA method (25). EF-G was purified by DEAE-Sephacel chromatography with column buffer (20 mM Tris, pH 8.0, 0.2 mM EDTA, 1 mM dithiothreitol, 10% glycerol) and elution with a linear 0 -0.35 M NaCl gradient in the same buffer, followed by covalent chromatography on thiol Sepharose with column buffer (20 mM potassium phosphate, 1 mM EDTA, pH 7.0) and elution with a gradient of 0 -10 mM cysteine in the same buffer. The purified protein was dialyzed for 3 h against 20 mM Tris, pH 7.4, 100 mM KCl, 5 mM MgCl 2 , 1 mM dithiothreitol and concentrated by ultrafiltration. EF-G was more than 98% pure as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Initiation factor 2 was purified according to Luchin et al. (8). The purified protein was dialyzed against 20 mM Tris, pH 7.4, 100 mM KCl, 5 mM MgCl 2 , 1 mM dithiothreitol. DnaK was prepared as described previously (26) from an overproducing strain of E. coli bearing plasmid pLNA2 derived from plasmid pDM38 (27)  Refolding of Citrate Synthase and ␣-Glucosidase-Denaturation and renaturation reactions were carried out at 25°C. For both proteins, renaturation was initiated by pouring the renaturation solvent onto the unfolded protein, under vortex agitation, in Eppendorf polyethylene tubes. Citrate synthase was denatured at a concentration of 10 M in 8 M urea, 50 mM Tris-HCl, 2 mM EDTA, 20 mM dithiothreitol, pH 8.0 for 50 min. Renaturation was initiated by a 100-fold dilution in 40 mM Hepes, 50 mM KCl, 10 mM (NH 4 ) 2 SO 4 , 2 mM potassium acetate, pH 8.0. The enzymatic activity of citrate synthase was measured as described (28). ␣-Glucosidase was denatured at a concentration of 3 M in 8 M urea, 0.1 M potassium phosphate, 2 mM EDTA, 20 mM dithiothreitol, pH 7.0 for 15 min. Renaturation was initiated by a 30-fold dilution in 40 mM Hepes-KOH, pH 7.8 at 20°C. The enzymatic activity of ␣-glucosidase was measured as described (28). Citrate synthase renaturation in the presence of nucleotides was done as described above (unless otherwise indicated) in the presence of 150 M GDP, GTP, or GTP␥S and 200 M MgCl 2 . The effect of nucleotides alone on citrate synthase renaturation was low (GTP and GTP␥S) or negligible (GDP) and was subtracted.
Thermal Aggregation of Citrate Synthase-The native enzyme (80 M) was diluted 100-fold in 40 mM Hepes, 50 mM KCl, 10 mM (NH 4 ) 2 SO 4 , 2 mM potassium acetate, pH 8.0 at 43°C in the absence of added proteins or in the presence of DnaK, EF-G, or IF2. Citrate synthase aggregation was monitored by measuring the absorbance at 650 nm as described in Ref. 28.
Size Exclusion Chromatography-For binding assays of R-CMLA and unfolded BPTI to EF-G, IF2, and DnaK, gel permeation columns (Bio-Gel P-200 from Bio-Rad for studies with R-CMLA or Sephadex G-75 from Amersham Pharmacia Biotech for studies with BPTI (300-l bed volume of each)) were equilibrated with column buffer containing 50 mM Tris-HCl (pH 8.2 for studies with R-CMLA and pH 7.4 for studies with BPTI), 50 mM KCl, 1 mM dithiothreitol, 100 g/ml bovine serum albumin. Reaction mixtures containing EF-G, IF2, or DnaK and radiolabeled native BPTI, unfolded BPTI, or R-CMLA at indicated concentrations were incubated for 20 min at 23°C in column buffer without serum albumin and applied to the column at room temperature. Fractions were collected at a flow rate of 1 drop/fraction/30 s and counted for radioactivity. DnaK was incubated for 3 h at 37°C before use. Unfolded BPTI was prepared as described previously from native BPTI (29). Unfolded BPTI, native BPTI, and R-CMLA were 3 H-labeled by reductive methylation (30).  Fig. 7D (8). The reactions were terminated by applying 3 l of each sample to a polyethyleneimine cellulose thin layer chromatography plate (29). 3 H-labeled GTP was from Amersham Pharmacia Biotech and was used at 1 Ci/mmol. The specific activities of EF-G and of IF2 were similar to those reported in Refs. 8 and 31.

EF-G Increases the Amount of Correctly Folded Citrate
Synthase and ␣-Glucosidase-We first investigated whether EF-G acts as a molecular chaperone in the folding of proteins. Citrate synthase and ␣-glucosidase, whose refolding is facilitated by several chaperones such as GroEL, DnaK, hsp90, and small hsps (28, 32-34), were chosen as substrates for this reaction. They were unfolded in the presence of 8 M urea, and allowed to refold upon dilution of the denaturant, in the absence or in the presence of EF-G (protein folding in the presence of DnaK was studied in parallel). Under our experimental conditions, the refolding yield of 0.1 M citrate synthase was increased from 7% in the absence of added proteins to 21% in the presence of 5 M EF-G and 31% in the presence of 3 M DnaK (Fig. 1A). The dependence of citrate synthase reactivation on the concentration of EF-G is shown in Fig. 1B. The maximal recovery of citrate synthase activity reaches 24% in the presence of 10 M EF-G, and half-maximal reactivation occurs at 2 M EF-G, a concentration slightly higher than that of DnaK (around 1 M) required for half-maximal reactivation of citrate synthase in similar conditions (not shown and Ref. 34). The EF-G concentration required for half-maximal reactivation of citrate synthase is somewhat higher than the concentration of citrate synthase but is lower than the concentrations of EF-G in the cytoplasm (around 20 M) (3,35). As reported previously (34), other proteins such as ovalbumin and lysozyme were unable to stimulate citrate renaturation, whereas serum albumin could stimulate it to some extent (not shown). In similar experiments, the refolding of 0.1 M ␣-glucosidase was increased from 6% in the absence of added protein to 18% in the presence of 2 M EF-G and 26% in the presence of 2 M DnaK (not shown). These results suggest that, like molecular chaperones, EF-G interacts with unfolded proteins and increases their productive folding.
Effect of Nucleotides on the EF-G-dependent Refolding of Citrate Synthase-Elongation factor G belongs to the GTPase superfamily of proteins whose functional cycle includes at least four major conformational states, the nucleotide-free state, the GDP complex, the GTP complex, and the GTPase state. The stimulation of citrate synthase renaturation by EF-G in the absence of nucleotide or in the presence of GDP or GTP␥S was 2.6-, 2.4-, and 1.6-fold, respectively ( Fig. 2A). Thus, the nucleotide-free form and the GDP form of EF-G are more active in citrate synthase renaturation than its GTP␥S form. We obtained similar results for citrate synthase renaturation by elongation factor Tu (18). In the presence of GTP the level of renaturation of citrate synthase is intermediate between those observed in the presence of GDP and GTP␥S ( Fig. 2A). However, if the concentration of EF-G is reduced to 0.1 M during the renaturation of 0.1 M citrate synthase, GTP produces a significant stimulation of citrate synthase renaturation, which rises from 1.3-fold in the presence of EF-G alone (1.4-fold in the presence of EF-G and GDP) to 1.9-fold in the presence of EF-G and GTP (Fig. 2B). Under these conditions, GTP alone did not stimulate citrate synthase renaturation. This suggests that the GTPase activity of EF-G can stimulate its chaperone activity and that this stimulation is apparent when EF-G is added in limiting amounts. A stimulation by GTP of the EF-Tu/EF-Tsdependent rhodanese renaturation has been observed by others (17). In addition to the stimulation by GTP of the EF-G-dependent citrate synthase renaturation, the GTPase activity of EF-G can be stimulated by the permanently unfolded protein R-CMLA.
Stimulation of the GTPase Activity of EF-G by an Unfolded Protein-In the absence of ribosomes, EF-G promotes very little GTP hydrolysis (k cat lower than 0.002 min Ϫ1 at 30°C). The EF-G GTPase is efficiently stimulated by ribosomal particles (k cat ϭ 20 min Ϫ1 ) and, to a lesser extent, by several aliphatic alcohols such as 2-propanol (k cat ϭ 0.03 min Ϫ1 in the presence of 20% 2-propanol) (31). R-CMLA is a permanently unfolded protein that maintains an extended conformation without any stable secondary structure in the absence of denaturant (36). It interacts with chaperones, and it stimulates their ATPase activity (23, 34, 36). R-CMLA stimulates 10-fold the GTPase activity of EF-G (Fig. 3), whose k cat at 25°C (calculated from the results shown in Fig. 3) rises from 0.002 min Ϫ1 in the absence of added protein, to 0.02 min Ϫ1 in the presence of 15 M R-CMLA. The K a of the stimulation of the EF-G GTPase by R-CMLA is around 5 M. The stimulation of the EF-G GTPase by an unfolded protein is reminiscent of that of the DnaK ATPase, whose k cat rises from 0.04 min Ϫ1 in the absence of a peptide substrate to 0.16 min Ϫ1 in its presence (37).
EF-G Protects Citrate Synthase from Irreversible Aggregation during Thermal Stress-We investigated the function of EF-G under heat shock conditions. As reported previously (28,32,34), citrate synthase loses its native conformation and undergoes aggregation during incubation at 43°C. The addition of EF-G (3 M) or DnaK (2 M) partially reduces citrate synthase (0.8 M) aggregation, whereas both 5 M DnaK and 8 M EF-G suppress citrate synthase aggregation (Fig. 4). In contrast, the addition of up to 35 M bovine serum albumin (Fig. 4), ovalbumin, or lysozyme (not shown and Ref. 34) does not protect citrate synthase from thermal aggregation. Thus, EF-G is nearly as efficient as DnaK and other chaperones (28,34) in protecting citrate synthase from thermal denaturation and is much more efficient than other proteins such as bovine serum albumin, ovalbumin, or lysozyme. Furthermore, the concentration of EF-G required for an efficient thermal protection of citrate synthase (around 5 M) is severalfold lower than its cellular concentration (around 20 M).
Interaction between EF-G and Unfolded Proteins-One characteristic of molecular chaperones is their preferential interaction with unfolded proteins (21-23). R-CMLA, a permanently unfolded protein that maintains an extended conformation without any stable secondary structure in the absence of denaturant, strongly interacts with several chaperones, including DnaK (34,36). Complex formation between R-CMLA (20,000 Da) and EF-G (82,000 Da) was analyzed by gel filtration on a Bio-Gel P-200 column. When R-CMLA (2 M) is filtered in the presence of EF-G (10 M), 14% of R-CMLA fractionates as a higher molecular weight complex than R-CMLA alone (Fig. 5). The interaction between EF-G and R-CMLA is not significantly different from that observed between R-CMLA and DnaK; 10 M DnaK binds 33% of R-CMLA (not shown). Thus, EF-G seems to interact strongly with R-CMLA in a manner similar to that of DnaK. Unfolded BPTI is known to interact with chaperones, including DnaK (34,38). Complex formation between 1 M unfolded BPTI (6,000 Da) and 10 M EF-G (82,000 Da) was studied by gel filtration on a Sephadex G-75 column; a significant percentage (39%) of unfolded BPTI fractionates as higher molecular weight material than unfolded BPTI alone (Fig. 6). In similar conditions 4 M DnaK retained 27% of unfolded BPTI (Fig. 6). In contrast, when 1 M native BPTI and 10 M EF-G were loaded on the gel permeation column, native BPTI did not elute as a high molecular weight complex (Fig. 6). When similar experiments were carried out with bovine serum albumin (30 M) or ovalbumin (30 M), unfolded BPTI did not elute as a high molecular weight complex (not shown). Thus, EF-G, like molecular chaperones, interacts preferentially with unfolded proteins.
Chaperone Properties of IF2-Citrate synthase was unfolded in the presence of 8 M urea and allowed to refold upon dilution of the denaturant in the absence or in the presence of IF2 (under conditions similar to those described for the experiment represented by Fig. 1 for EF-G). The refolding yield of 0.1 M citrate synthase was increased from 6% in the absence of IF2 to 20% in the presence of 1 M IF2 (Fig. 7A). Half-maximal reactivation of citrate synthase occurs at 0.15 M IF2 (Fig. 7A). Thus, IF2 stimulates the refolding of citrate synthase with a similar efficiency as EF-G and is efficient at lower concentra-tions. In similar experiments, the refolding of 0.1 M ␣-glucosidase was increased from 7% in the absence of added protein to 15% in the presence of 1 M IF2 (not shown).
We investigated the ability of IF2 to protect citrate synthase from irreversible aggregation during thermal stress under con- ditions similar to those described for the experiment represented by Fig. 4 for EF-G. IF2 (3 M) reduces citrate synthase (0.8 M) aggregation at 43°C with an efficiency that is similar to that of 3 M EF-G (compare Figs. 7B and 4).
We also studied the interaction of IF2 with unfolded proteins. The interaction of IF2 with unfolded BPTI is shown in Fig. 7C. Complex formation between 1 M unfolded BPTI (6,000 Da) and 5 M IF2 (97,000 Da) was studied by gel filtration on a Sephadex G-75 column, under conditions similar to those described for the experiment represented by Fig. 6. A significant percentage (32%) of unfolded BPTI fractionates as higher molecular weight material than unfolded BPTI alone (Fig. 7C). In contrast, when 1 M native BPTI and 5 M IF2 were loaded on the gel permeation column, native BPTI did not elute as a high molecular weight complex (not shown). Similar experiments with IF2 and the permanently unfolded protein R-CMLA showed a strong interaction between the initiation factor and R-CMLA (not shown). Thus, IF2, like EF-G and molecular chaperones, interacts preferentially with unfolded proteins.
We have shown that R-CMLA can stimulate the GTPase activity of EF-G. IF2 possesses a GTPase activity that has been reported to be totally dependent on the presence of the ribosome (5,8). We measured the GTPase activity of IF2 in the presence of several concentrations of R-CMLA (in the absence of ribosomes). The IF2 GTPase activity, which is undetectable in the absence of ribosomes, as reported by others, is stimulated by R-CMLA with a K a around 15 M and a k cat of 0.4 min Ϫ 1 at 25°C (calculated from the results shown in Fig. 7D). This activity is similar to that of IF2 in the presence of ribosomes (k cat ϭ 0.8 min Ϫ 1 at 37°C). In accordance with the stimulation of the IF2 GTPase by an unfolded protein, the renaturation of 0.1 M unfolded citrate synthase by 0.1 M IF2 was stimulated 1.7-fold by 1 mM GTP (not shown). DISCUSSION We present biochemical evidence suggesting that EF-G and IF2 play a chaperone-like function in protein folding, protection against thermal denaturation, and interaction with unfolded proteins. EF-G and IF2 increase approximately 3-fold the yield of citrate synthase and ␣-glucosidase renaturation, as do molecular chaperones. The stimulation factors of protein renaturation (more than 3-fold) and EF-G or IF2 concentrations required for half-maximal protein renaturation (2 M and 0.15 M, respectively) are not significantly different from those obtained with DnaK, hsp90, or small hsps (this study and Refs. 28,[32][33][34]. Notably, the concentrations of EF-G and IF2 used in this study (in the micromolar range) are lower than their estimated concentration in the bacterial cytoplasm (around 20 and 4 M, respectively) (3,6). At micromolar concentrations, EF-G and IF2 protect citrate synthase from thermal denaturation. These concentrations are similar to those of DnaK and of small hsp (expressed as monomers) required for a similar protection (28,34). They are in the same range as the cytoplasmic concentrations of EF-G and IF2. Furthermore, other proteins tested (bovine serum albumin, ovalbumin, and lysozyme) do not protect citrate synthase efficiently (34). Like DnaK and other chaperones (32)(33)(34), EF-G and IF2 form stable complexes with unfolded proteins. EF-G and IF2 interact with unfolded BPTI but not with native BPTI and thus appear to discriminate between unfolded and native proteins.
EF-G⅐GDP appears more active in binding unfolded proteins than EF-G⅐GTP␥S. Interestingly, the conversion of EF-G⅐GTP to EF-G⅐GDP is accompanied by an increase in binding of the hydrophobic fluorescent probe anilino-naphtalenesulfonate (39), suggesting that EF-G⅐GDP exposes a hydrophobic surface that might interact with unfolded proteins (hydrophobic interactions occur between chaperones and unfolded proteins) (21)(22)(23). GTP stimulates the EF-G-and IF2-dependent citrate synthase renaturations, under conditions where these translation factors are limiting, suggesting that they use their GTPase activity to catalyze protein folding. Similarly, the renaturation of rhodanese by EF-Tu, in the presence of EF-Ts, is stimulated by GTP (17). The stimulation of the GTPase activity of EF-G and of IF2 by the permanently unfolded protein R-CMLA is reminiscent of that of the ATPase activity of the DnaK and GroEL chaperones by unfolded proteins (21)(22)(23).
Our results do not indicate whether EF-G or IF2 interact with nascent chains during translation, and neither EF-G nor IF2 has been identified as one of the nascent chains-binding proteins (40 -43). Interestingly, EF-G is a component of the 4.5 S RNA complex, which acts as a secretion-specific chaperone for several nascent proteins (44). Hardesty and colleagues (17) have shown that EF-G is not able to catalyze rhodanese renaturation but can stimulate, in the presence of GTP, the ribosomedependent rhodanese renaturation (20). Our results suggest that EF-G itself can display chaperone properties, because it interacts specifically with the unfolded proteins citrate synthase, ␣-glucosidase, and R-CMLA and unfolded BPTI, similar to classical chaperones. The recent results concerning the chaperone properties of ribosomes and elongation factors support the hypothesis of a cotranslational folding of nascent proteins (45), in which folding into the tertiary structure is facilitated by interactions with ribosomes and elongation factors (17)(18)(19)(20).
The efficient protection of citrate synthase from thermal denaturation afforded by EF-G and IF2 suggests that during heat shock these translation factors might contribute to a reservoir of chaperones and chaperone-like molecules that serve as a chaperone-buffer in preventing the aggregation of nonnative proteins until permissive renaturation conditions are restored. Finally, the chaperone properties of EF-Tu (17,18), EF-G, and IF2 (this study) and the protein disulfide isomerase activity of EF-Tu (19) suggest that translation factors are ancestral protein folding factors that appeared before dedicated chaperones and protein disulfide isomerases.