Engineering and Characterization of the Ribosomal L10·L12 Stalk Complex

The ribosomal stalk protein L12 is essential for events dependent on the GTP-binding translation factors. It has been recently shown that ribosomes from Thermus thermophilus contain a heptameric complex L10·(L12)2·(L12)2·(L12)2, rather than the conventional pentameric complex L10·(L12)2·(L12)2. Here we describe the reconstitution of the heptameric complex from purified L10 and L12 and the characterization of its role in elongation factor G-dependent GTPase activity using a hybrid system with Escherichia coli ribosomes. The T. thermophilus heptameric complex resulted in a 2.5-fold higher activity than the E. coli pentameric complex. The structural element of the T. thermophilus complex responsible for the higher activity was investigated using a chimeric L10 protein (Ec-Tt-L10), in which the C-terminal L12-binding site in E. coli L10 was replaced with the same region from T. thermophilus, and two chimeric L12 proteins: Ec-Tt-L12, in which the E. coli N-terminal domain was fused with the T. thermophilus C-terminal domain, and Tt·Ec-L12, in which the T. thermophilus N-terminal domain was fused with the E. coli C-terminal domain. High GTPase turnover was observed with the pentameric chimeric complex formed from E. coli L10 and Ec-Tt-L12 but not with the heptameric complex formed from Ec-Tt-L10 and Tt·Ec-L12. This suggested that the C-terminal region of T. thermophilus L12, rather than the heptameric nature of the complex, was responsible for the high GTPase turnover. Further analyses with other chimeric L12 proteins identified helix α6 as the region most likely to contain the responsible element.

The stalk protein L12 (generally referred to as L7/L12) is the only multi-copy protein in bacterial ribosomes (1,2). It is involved in the interaction of the ribosome with the elongation factors EF-G 2 and EF-Tu, activation of the GTPase activity of both factors, and associated reactions during translation (3)(4)(5)(6). L12 is comprised of three regions: the N-terminal domain, which participates in the dimerization of L12 (for review see Ref. 7) and its binding to L10 (for review see Ref. 8), the globular C-terminal domain (CTD) (9,10), which interacts with the elongation factors (11)(12)(13)(14), and a hinge region, which connects the two domains (15,16). It was generally accepted that, in the bacterial ribosomal large subunit, two L12 dimers bind to L10 to form a pentameric complex, L10⅐(L12) 2 ⅐(L12) 2 (7,17). Recent studies, however, have demonstrated that some bacterial and archaeal species contain ribosomes with three L12 dimers that form a heptameric complex L10⅐(L12) 2 ⅐(L12) 2 ⅐(L12) 2 (6,18,19). This finding opened a new stage in the discussion as to the significance of multiple stalk dimers in the ribosome.
The L10⅐L12 stalk protein complex constitutes a highly flexible region in the ribosome (8, 20 -25), and its detailed structure within the ribosome has not been resolved by x-ray crystallography (26 -29). However, some structural features of isolated L12 have been determined. An earlier study of Escherichia coli L12 by x-ray crystallography revealed the structure of the CTD to be comprised of three ␣ helices (␣4, ␣5, and ␣6) and a central three-stranded ␤ sheet (10). This structure was confirmed by a recent crystallographic study of L12 from Thermotoga maritima (30) and also by NMR analyses of E. coli L12 (16). More recently, the crystal structure of the complex between the N-terminal domain (helices ␣1 and ␣2) of L12 and L10 from T. maritima was elucidated using a reconstituted sample (6). In this complex, helices ␣1 and ␣2 of the L12 N-terminal domain participate in binding to the long C-terminal helix ␣8 of L10, as well as in the formation of the homodimer. Three L12 dimers bind in the same manner, side by side, to the C-terminal helix of L10 (6). The hinge region (␣3) of each L12 monomer seems to adopt a random coil structure, which is supported by the NMR data (20,21,24).
Although it has long been known that the CTD of L12 participates in EF-G-dependent GTP hydrolysis (6,7), the detailed mechanism of the GTPase activation is not fully understood. Recent cryo-electron microscopy data showed that the L12 CTD directly interacts with the GЈ domain of EF-G (13). Sitedirected mutagenesis and kinetic studies suggested that helices * This work was supported by Grant-in-Aid for Scientific Research 14035222 (to T. U.), grants from the Uchida Energy Science Promotion Foundation (to T. U.) and the Naito Foundation (to T. U.), and a Grant-in-Aid for Global COE Program (to T. N.) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. 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 and ␣5 in the L12 CTD are involved in the binding of EF-Tu (11) and that ␣4 also participates in the binding of EF-G (12) and in the release of inorganic phosphate from the post-GTPase state of the ribosome rather than in the GTPase catalysis (12). Direct binding of the translation factors to parts of ␣4 and ␣5 in the L12 CTD was also confirmed by heteronuclear NMR spectroscopy (14). On the other hand, there is sufficient NMR evidence to show that the hinge region and the entire CTD are highly mobile within the ribosome (23,25). Cross-linking and immuno-electron microscopy data also provide evidence that the CTD interacts with various regions of the ribosome (31,32). These lines of evidence lead to a hypothesis that the movement of the L12 CTD in the ribosome increases the frequency of contact between the CTD and the translation factors, and thus, the presence of multiple copies of L12 in the ribosome might amplify the efficiency of this process. If this hypothesis is correct, ribosomes that contain three L12 dimers, which have been detected in some thermophilic bacteria and archaea, should be able to recruit the translation factors more effectively than ribosomes that contain two L12 dimers. It has been reported that the T. thermophilus ribosome contains six copies of the L12 protein. We have recently established conditions that allow the reconstitution of the T. thermophilus stalk complex and the assay of its structure/function (33). We confirmed that the T. thermophilus L12 forms a heptameric complex with L10 L10⅐(L12) 2 ⅐(L12) 2 ⅐(L12) 2 and showed that the T. thermophilus stalk complexes, even a variant complex that contains only single L12 dimer, induce higher levels of EF-Gdependent GTPase activity, but not higher EF-Tu-dependent GTPase or polyphenylalanine synthesis, than the E. coli pentameric complex L10⅐(L12) 2 ⅐(L12) 2 . Given that L12 makes a greater contribution to the activation of the EF-G GTPase than that of EF-Tu (5,34), it is likely that Tt-L12 specifically stimulates the EF-G-dependent event.
To determine the structural features of the T. thermophilus stalk complex that are responsible for the higher EF-G-dependent GTPase activity, we here constructed chimeric L10 and L12 proteins from the T. thermophilus and E. coli orthologs and reconstituted different chimeric stalk complexes that included: 1) an E. coli pentameric complex that contained L12 in which the CTD had been exchanged for that of T. thermophilus and 2) a heptameric complex that was composed of E. coli L10 whose C-terminal ␣8 region had been replaced with that of T. thermophilus and E. coli L12 whose N-terminal ␣1 and ␣2 regions had also been replaced with those of T. thermophilus (see Fig. 2, A and C). The results of functional assays demonstrated that the CTD of T. thermophilus L12, rather than the heptameric structure, was required for the high GTPase activity. Further analyses using other chimeric complexes suggested that helix ␣6 within the CTD was involved in the high GTPase turnover activity.
Protein Expression and Purification-Proteins were expressed in E. coli strain BL21 (DE3) pLysS (Novagen) by the addition of isopropyl ␤-D-thiogalactopyranoside. The cells were lysed by grinding with alumina and then resuspended in buffer A, which contained 20 mM Tris-HCl (pH 7.6), 10 mM MgCl 2 , 20 mM NH 4 Cl, and 5 mM 2-mercaptoethanol (17). All of the proteins, except Ec-Tt-L10 and Tt-L10, were obtained in the soluble (S100) fraction. The S100 fractions that contained the proteins were dialyzed against buffer B, which contained 20 mM NaOAc (pH 4.5), 7 M urea, and 5 mM 2-mercaptoethanol and then applied to a column of CM-cellulose (Whatman) that had been equilibrated with the same buffer. Ec-L12, Ec-Tt-L12, Tt⅐Ec-L12, Ec-␣3Tt-L12, Ec-␣4Tt-L12, Ec-␣5Tt-L12, and Ec-␣6Tt-L12 were all eluted in the flow-through fraction, whereas Ec-L10, Ec-L11, and Tt-L11 were eluted in a stepwise manner using buffer B containing 0.04 -0.12 M LiCl. Individual proteins that had been eluted in the flow-through fraction were dialyzed against buffer C, which consisted of 20 mM Tris-HCl (pH 8.0), 6 M urea, and 5 mM 2-mercaptoethanol and then purified using an AKTA fast protein liquid chromatography system with a Resource Q column (GE Healthcare) and a linear gradient of 0 -200 mM LiCl in buffer C. The fractions that contained Ec-L10, Ec-L11, and Tt-L11 were individually dialyzed against buffer D, which consisted of 20 mM NaOAc (pH 5.0), 6 M urea, and 5 mM 2-mercaptoethanol and purified using an AKTA fast protein liquid chromatography system with a Resource S column (GE Healthcare) and a linear gradient of 0 -200 mM LiCl in buffer D. The S100 fraction that contained Tt-L12 was heated to 70°C for 10 min to remove the E. coli proteins, and the Tt-L12 protein was then purified in a single step by DE-cellulose column chromatography (Whatman) in buffer C. The insoluble GST⅐Tt-L10 and Ec-Tt-L10 proteins were solubilized in 8 M urea and purified by the same column chromatography as Ec-L10, described above. The purity of the proteins was confirmed by SDS-polyacrylamide gel electrophoresis (see Fig. 2B).
Formation of the L10⅐L12 Stalk Complex-The purified L10 and L12 proteins were individually dialyzed against buffer E, which consisted of 20 mM Tris-HCl (pH 7.6), 300 mM KCl, and 5 mM 2-mercaptoethanol. An L10 protein (Ec-L10, GST⅐Tt-L10, or Ec-Tt-L10) was mixed with an L12 protein (Ec-L12, Tt-L12, or one of the chimeric L12 proteins) in buffer E at a molar ratio of 1:8. The complexes were formed by incubation at 37°C for 5 min. In the case of the complex that contained GST⅐Tt-L10, the GST moiety was removed by digestion with PreScission Protease (GE Healthcare) according to the manufacturer's instructions and subsequent heating at 70°C for 30 min. The complexes that formed were analyzed by 6% polyacrylamide (acrylamide/bisacrylamide ratio 39:1) native gel electrophoresis at 6.5 V/cm with a buffer system that contained 50 mM Tris-HCl (pH 8.0), 5 mM MgCl 2 , 50 mM KCl (35). Electrophoresis was performed for 6 h at constant voltage and 4°C with buffer circulation.
Ribosomal Subunits and 50 S Core Particles-E. coli ribosomes and 50 and 30 S subunits were prepared as described previously (37,38). 50 S core particles that were deficient in L10, L12, and L11 were obtained by treatment of 50 S subunits from the L11-deficient E. coli mutant strain AM68 (39) with a solution that contained 50% ethanol and 0.5 M NH 4 Cl at 0°C, as described previously (38). 30 S subunits from E. coli strain Q13 were used for the functional assays in the present study.
Analysis of Binding of the Chimeric Stalk Complexes to 50 S Core Particles-50 S core particles were mixed with L11 and various different L10⅐L12 complexes and incubated at 37°C for 5 min. The samples were subjected to electrophoresis on an acrylamide/agarose composite gel that was composed of 3% acrylamide (acrylamide/bisacrylamide ratio 19:1) and 0.5% agarose (40), as described previously (37). The gel was stained with 0.2% Azur B. Binding of the proteins to the 50 S core particles was revealed by gel mobility shift.
Quantitative Analysis of L12 Incorporated into the Ribosome-The number of individual L12 proteins that were incorporated into a single 50 S core particle was determined according to Griaznova and Traut (17). Because L12 contains no Cys residues for labeling with [ 14 C]iodoacetamide, a single Cys was introduced by site-directed mutagenesis (17): A63C in Ec-L12 or A68C in Tt-L12. The expressed and purified L12 samples were radiolabeled by incubation at 37°C for 3 h in a solution that contained 100 mM Tris-HCl (pH 7.6), 2 mM dithiothreitol, 2 mM [ 14 C]iodoacetamide (65 mCi/mmol), as described (17). The labeled proteins were dialyzed against 20 mM Tris-HCl (pH 7.6), 10 mM MgCl 2 , 60 mM NH 4 Cl, and 5 mM 2-mercaptoethanol, and their specific radioactivities were calculated from the protein concentration values, which were determined using a Micro BCA protein assay kit (Pierce), and radioactivity, which was determined using a model 1600TR Tri-Carb liquid scintillation counter (Packard). 50 S core particles (50 pmol) were incubated with a 2-fold molar excess of an L10 protein and L11, together with a 12-16-fold molar excess of a radiolabeled L12 protein, in 50 l of reconstitution buffer, which contained 20 mM Tris-HCl (pH 7.6), 10 mM MgCl 2 , 60 mM NH 4 Cl, and 5 mM 2-mercaptoethanol at 37°C for 10 min, and then layered over a 20% sucrose cushion (1 ml) in the same buffer. The reconstituted 50 S particles were recovered by centrifugation at 60,000 rpm for 6 h at 4°C in an S100AT4 rotor using an Hitachi microultracentrifuge (CS100GX). The ribosomal pellet was resuspended in 50 l of reconstitution buffer. The amount of radiolabeled L12 that was bound to each 50 S core particle was calculated from the radioactivity of the sample and the number of 50 S particles recovered, which was determined from the A 260 (38.6 pmol of 50 S subunits/A 260 unit).
Ribosomal Function Assays-E. coli EF-G was cloned into the pET3a vector, overexpressed in E. coli cells (41) and purified from the cell extract by DEAE Sephadex A-50 column chromatography, as described (42). EF-G-dependent GTPase activity was measured as described previously (37), except that the reaction mixture contained 2.5 pmol of 50 S subunits, 7.5 pmol of 30 S subunits, and 7 mM MgCl 2 .
Footprinting Assay-70 S ribosomes (20 pmol) were preincubated in 50 l of buffer that contained 7 mM MgCl 2 , 50 mM NH 4 Cl, 5 mM 2-mercaptoethanol, and 50 mM potassium cacodylate (pH 7.2) in the presence or absence of 40 pmol of EF-G, 1 mM fusidic acid, and 0.5 mM GTP. The ribosome samples were then mixed with 1 l of dimethyl sulfate (a 1:4 dilution in ethanol) and incubated at 37°C for 10 min. An aliquot of rRNA from each sample, which had been extracted with phenol, was analyzed by primer extension and gel electrophoresis as described by Moazed and Noller (43). The primer used was 5Ј-CTCGGGGCAAGTTTCGTGCT-3Ј, which binds within the 2660 domain and is complementary to nucleotides 2758 -2777 of 23 S rRNA.

RESULTS
High EF-G-dependent GTPase Turnover Activity with the Reconstituted T. thermophilus Stalk Protein Complex-Mixing of the purified wild-type Tt-L10 and Tt-L12 proteins resulted in the formation of a stable complex (33). This L10⅐L12 complex, together with Tt-L11, was assembled onto E. coli 50 S core particles that were deficient in the equivalent proteins (38). The functionality of the T. thermophilus proteins was assessed by measuring EF-G-dependent GTP hydrolysis. The activity was compared with that obtained with E. coli proteins that had been reconstituted under the same conditions (Fig. 1). The T. thermophilus L10⅐L12 complex, together with Tt-L11, yielded ribosomes with 2.5-fold higher activity than that of ribosomes that had been reconstituted with Ec-L10⅐L12 and Ec-L11. This higher activity seems to be mainly due to the T. thermophilus FIGURE 1. Higher EF-G-dependent GTPase activity is obtained with the reconstituted L10⅐L12 complex and L11 from T. thermophilus than with their counterparts from E. coli. 50 S core particles (2.5 pmol) that lacked L11, L10, and L12 were preincubated with L11 (10 pmol) and the L10⅐L12 complex (10 pmol) from T. thermophilus (Tt) or E. coli (Ec), together with 30 S subunits (7.5 pmol), at 37°C for 5 min. The resulting ribosome samples were assayed for GTPase activity in the presence of E. coli EF-G.
L10⅐L12 complex, because in the presence of Ec-L11, this complex still gave more than 2-fold higher activity than that of E. coli reconstituted ribosomes, whereas the Ec-L10⅐Ec-L12 complex in the presence of Tt-L11 did not markedly stimulate the GTPase activity. It is therefore likely that the higher activity is due to a structural feature of the Tt-L10⅐Tt-L12 complex. This might be either the presence of three Tt-L12 dimers or other structural elements that are found in each L12 molecule.
Construction of Chimeric L10 and L12 Proteins from the E. coli and T. thermophilus Orthologs-Sequence comparisons show that the C-terminal ␣8 region of Tt-L10 contains an additional stretch of amino acids that is absent in Ec-L10. It has been suggested that this extra region is involved in the binding of the third Tt-L12 dimer (6,18). We recently confirmed by biochemical approaches that the ␣8 region of Tt-L10 is involved in the binding of the three Tt-L12 dimers (33). To characterize the structure and function of the T. thermophilus heptameric complex Tt-L10⅐(Tt-L12) 2 ⅐(Tt-L12) 2 ⅐(Tt-L12) 2 , we prepared constructs that allowed the expression in E. coli of Tt-L10, Ec-L10, and a chimeric L10 protein, in which the ␣8 region of Ec-L10 was replaced with that of Tt-L10 (Ec-Tt-L10) ( Fig. 2A). We also prepared constructs for the expression of Tt-L12, Ec-L12, and two chimeric L12 proteins: Ec-Tt-L12, in which the region that corresponded to ␣1 and ␣2 (i.e. the binding site for L10) in Tt-L12 was replaced with that of Ec-L12, and Tt-Ec-L12, in which the ␣1-␣2 region of Ec-L12 was replaced with that of Tt-L12 ( Fig. 2A). The proteins were expressed in E. coli cells and purified (Fig. 2B), as described under "Experimental Procedures." Only the Tt-L10 protein was expressed and isolated as a fusion protein with GST (lane 2). It was expected that Ec-Tt-L10, as well as Tt-L10, would bind three Tt-L12 (or Tt-Ec-L12) homodimers and that Ec-L10 would bind two Ec-Tt-L12 (or Ec-L12) homodimers (Fig. 2C).
Formation of Chimeric Stalk Protein Complexes-To investigate which structural feature of the T. thermophilus stalk complex Tt-L10⅐Tt-L12 is responsible for the higher GTPase activity, we used the chimeric L10 and L12 proteins, described above ( Fig. 2A). By using Ec-Tt-L10 and Tt-Ec-L12, we would expect to form a heptameric complex (Fig. 2C), in which the functional regions for rRNA binding (the N-terminal domain of L10) and elongation factor binding (the C-terminal domain of L12) are composed of E. coli sequences, respectively. In contrast, by using Ec-L10 and Ec-Tt-L12, we would expect to form a pentameric complex, in which the factor binding site is composed of T. thermophilus L12 sequences (Fig. 2C).
Binding of the chimeric stalk protein complexes to the E. coli 50 S core was confirmed by acrylamide/agarose composite gel electrophoresis (Fig. 3), in which the band of 50 S core particles (lane 1) was clearly shifted upward by the addition of Ec-L10 with Ec-L11 (lane 3) and supershifted by the addition of Ec-L12 together with Ec-L10 and Ec-L11 (lane 4). When a chimeric L12, Ec-Tt-L12, was added instead of Ec-L12, a similar supershift was observed (lane 5). However, another chimeric L12 protein, Tt-Ec-L12, did not cause this L12-dependent supershift (lane 6). When the chimeric Ec-Tt-L10 was used, with Ec-L11, the band shift of the 50 S core was observed (lane 7). Furthermore, the L12-dependent supershift was also observed when either Tt-L12 (lane 9) or Tt-Ec-L12 (lane 11) was added  Because the exact number of copies of the L12 variants within the reconstituted stalk complexes was not clear from the gel analysis (Fig. 3), we performed a quantitative analysis using L12 samples that had been radiolabeled with [ 14 C]iodoacetamide at a single cysteine, as described under "Experimental Procedures." 50 S ribosomal particles that had been reconstituted with the chimeric L10⅐L12 complexes that contained the radiolabeled L12 were pelleted by centrifugation through a sucrose cushion. The number of copies of L12 that were recovered relative to each reconstituted 50 S particle was calculated, as shown in Table 1. The 50 S subunits that had been reconstituted with Ec-L10 contained 4.18 copies of Ec-L12 or 4.12 copies of Ec-Tt-L12, which implied that there were two homodimers/50 S particle. The 50 S subunits that had been reconstituted with Ec-Tt-L10 contained 6.25 copies of Tt-L12 or 6.27 copies of Tt-Ec-L12, which implied that there were three homodimers/50 S particle. Only very low amounts of all the L12 proteins bound to the 50 S core in the absence of L10. These results suggested that the chimeric stalk complex Ec-L10⅐Ec-Tt-L12 was a pentamer, whereas Ec-Tt-L10⅐Tt-L12 and Ec-Tt-L10⅐Tt⅐Ec-L12 were both heptameric complexes, as depicted in Fig. 2C.

Identification of the Structural Element of the T. thermophilus Stalk Complex Responsible for High EF-G-dependent GTPase
Activity-The chimeric L10⅐L12 stalk complexes that had been reconstituted into E. coli ribosomes were tested for their ability to stimulate EF-G-dependent GTPase activity (Fig. 4). Ribosomes that contained the chimeric pentameric complex Ec-L10⅐Ec-Tt-L12 showed 2.5-fold higher activity than that of ribosomes that contained the wild-type Ec-L10⅐Ec-L12 complex. This high level of GTPase activity was close to that obtained with one of the chimeric heptameric complexes, Ec-Tt-L10⅐Tt-L12 (Fig. 4), or the wild-type heptameric Tt-L10⅐Tt-L12 (Fig. 1). In contrast, another chimeric heptameric complex, Ec-Tt-L10⅐Tt⅐Ec-L12, resulted in a low level of activity, which was comparable with that of the E. coli wild-type pentameric complex Ec-L10⅐Ec-L12 (Fig. 4). The results indicated that the very high level of GTPase activity in the presence of the Tt-L10⅐Tt-L12 complex was due to a feature of the C-terminal moiety of Tt-L12, not to the presence of three L12 homodimers.
The amino acid sequence of the C-terminal moiety, which was responsible for the high GTPase activity, was compared between the T. thermophilus and E. coli proteins (Fig. 5A). There are four regions in which the amino acid sequence differs markedly between them. These divergent areas lie within four helical regions: ␣3, ␣4, ␣5, and ␣6, according to crystal structure data from isolated T. maritima L12 (30). It should be noted that ␣3 forms the hinge region. To identify the main region that is responsible for the high GTPase activity, the divergent areas in Ec-L12, which correspond to amino acids 43-47, 60 -75, 87-94, and 101-113, were replaced with the corresponding amino acids, 43-52, 65-80, 92-98, and 105-117, respectively, from Tt-L12. These chimeric proteins were designated Ec-␣3Tt-L12, Ec-␣4Tt-L12, Ec-␣5Tt-L12, and Ec-␣6Tt-L12, respectively (Fig. 5B). These proteins were overexpressed and purified (Fig. 5C), as described under "Experimental Procedures," and used for further GTPase assays. Only the ribosomes that had been reconstituted with Ec-␣6Tt-L12, not the other chimeric proteins, showed 2-fold higher GTPase activity than that of the ribosomes that contained the wild-type Ec-L12 (Fig. 6A).
Characterization of the Interaction between the Ec-Tt-L12containing Ribosome and EF-G-As shown in Fig. 4, ribosomes that contained the chimeric pentameric complex Ec-L10⅐Ec-Tt-L12 showed 2.5-fold higher GTPase activity than that of the wild-type E. coli ribosome. We further characterized the mode of interaction of the Ec-Tt-L12-containing ribosome with EF-G. The GTPase turnover of EF-G with the Ec-Tt-L12-containing ribosome was inhibited by the antibiotic fusidic acid in a manner similar to that with the wild-type E. coli ribosome

TABLE 1
Binding of radiolabeled L12 proteins to 50 S core particles that had been reconstituted with L10 proteins A single Cys was introduced by the A63C mutation in Ec-L12 or the A68C mutation in Tt-L12. Each mutant L12 was labeled with ͓ 14 C͔iodoacetamide, as described under "Experimental Procedures." (Fig. 7A). Footprinting analysis was performed to verify the binding site of EF-G. The factor, in the presence of fusidic acid and GTP, protected Gly 2659 and Ala 2660 of 23 S rRNA in the Ec-Tt-L12-containing ribosome, as well as in the wild-type ribosome, against modification with dimethyl sulfate (Fig. 7B). The results suggested that EF-G binding to the Ec-Tt-L12-containing ribosome occurred in a similar manner as with the wildtype ribosome. It is therefore likely that the GTPase activity may be stimulated by the Tt-L12 structural element because of the efficient dissociation of the EF-G⅐GDP complex from ribosomes after a rearrangement of the ribosome⅐EF-G⅐GDP conformation (44).
It should be added that Tt-L12 seems to stimulate specifically the EF-G-dependent event. There was no marked effect of the T. thermophilus proteins on EF-Tu-dependent GTPase activity or polyphenylalanine synthesis (33). Moreover, replacement of the L12 gene with the coding sequence for the Ec-Tt-L12 chimera in the E. coli genome had no effect on cell growth; ribosomes that were isolated from the Ec-Tt-L12 mutant cells showed 2.5-hold higher EF-G-dependent GTPase activity than the wild-type ribosomes (supplemental Fig. S1).

DISCUSSION
Rather than the pentameric stalk complex that is found in E. coli, Ec-L10⅐(Ec-L12) 2 ⅐(Ec-L12) 2 , T. thermophilus ribosomes contain a heptameric complex that is composed of six copies of the Tt-L12 protein with a single copy of Tt-L10, Tt-L10⅐(Tt-L12) 2 ⅐(Tt-L12) 2 ⅐(Tt-L12) 2 (18). Here we have characterized this complex by constructing chimeric L10 and L12 proteins. The chimeric Ec-Tt-L10 protein, in which the ␣8 region of Ec-L10 was replaced with the equivalent ␣8 region from Tt-L10, could bind six copies of Tt-L12. Moreover, Ec-Tt-L10 also bound six copies of Tt-Ec-L12, in which the ␣1 and ␣2 regions of Ec-L12 were replaced with the corresponding region from Tt-L12 (Table 1). The results clearly indicated that the C-terminal ␣8 region of Tt-L10 was responsible for the binding of all six copies of Tt-L12 and that the ␣1 and ␣2 regions of Tt-L12 participated in the binding to the ␣8 region of Tt-L10, as well as in the formation of the homodimer. Our biochemical data were consistent with previous crystal structure data for the L10⅐L12 complex of T. maritima from Diaconu et al. (6). Because the ␣8 region of Tt-L10 has an additional sequence that is absent in Ec-L10, it is likely that this extra sequence is the binding site for the third Tt-L12 dimer. Understanding the relationship between the number of copies of L12/ribosome and ribosome function is of interest. Because the T. thermophilus heptameric complex, which contains six copies of L12, resulted in 2.5-fold higher EF-G-dependent GTPase activity than the E. coli pentameric complex (Fig. 1), we examined whether this higher activity was due to the presence of six copies of Tt-L12 or to elements within each Tt-L12 molecule. We showed that the activity obtained with a chimeric heptameric complex, Ec-Tt-L10⅐(Tt⅐Ec-L12) 2 ⅐(Tt⅐Ec-L12) 2 ⅐(Tt⅐Ec-L12) 2 , was at a level comparable with that obtained with the E. coli pentameric complex, whereas the activity obtained with another chimeric pentameric complex, Ec-L10⅐ (Ec-Tt-L12) 2 ⅐(Ec-Tt-L12) 2 , was comparable with that obtained with the T. thermophilus heptameric complex. Our results imply that the higher EF-G-dependent GTPase activity that is observed with the T. thermophilus stalk complex is not principally because of the presence of six copies of Tt-L12, but rather a structural element within the CTD of Tt-L12 is responsible for the higher activity. It is therefore likely that the contribution of the third stalk dimer to the activity is only small, at least at 37°C. The results are consistent with our previous functional study on the heptameric stalk complex from Pyrococcus horikoshii, i.e. loss of the third stalk dimer from the heptameric complex has only a slight effect on ribosome function (19). Therefore, the question of the functional difference between the pentameric and heptameric stalk complexes in the ribo- coli L12 were replaced with the homologous sequences from T. thermophilus L12. A, sequence comparison of the CTD of L12 between E. coli (Ec, residues 37-120) and T. thermophilus (Tt, residues 37-124). B, the chimeric L12 constructs used were Ec-␣3Tt-L12, Ec-␣4Tt-L12, Ec-␣5Tt-L12, and Ec-␣6Tt-L12, which consisted of Ec-L12 in which residues 37-50 in ␣3, residues 63-77 in ␣4, residues 80 -94 in ␣5, and residues 100 -113 in ␣6, respectively, were replaced with the equivalent amino acids from Tt-L12. The portions replaced with T. thermophilus sequences are shown in gray. C, SDS-PAGE of the purified chimeric proteins. some remains unanswered. The third stalk dimer may play its role only at very high temperatures. It is also conceivable that the complex with three stalk dimers is more stable than the complex with two stalk dimers, and this may make the assembly of the complex at extremely high temperatures more efficient. However, it should be mentioned that the heptameric stalk complex might not be found in ribosomes from all thermophilic species (45).
It has been reported that the EF-G⅐GTP complex has the ability to bind to ribosomes in the absence of L12, although the binding is further enhanced by the presence of L12 (5,46,47). This implies that L12 is not essential for the binding of EF-G to the ribosome but enhances its binding. However, L12 may be essential for the correct recruitment of EF-G into the GTPaseactivating site of the ribosome or alternatively for the GTPase activation itself. The present study, using the chimeric proteins derived from the E. coli and T. thermophilus orthologs, clearly indicated that the CTD of Tt-L12 contained structural features that are required for the higher GTPase activity. Although the L12 orthologs are well conserved between E. coli and T. thermophilus, there are four divergent areas within the ␣3, ␣4, ␣5, and ␣6 regions of the CTD (Fig. 5A). The present study, using chimeric L12 proteins, also demonstrated that the unique amino acid sequence in the ␣6 region of Tt-L12 was required for the higher GTPase activity (Fig. 6A). Although the ␣4 and ␣5 regions are known to provide the binding sites for the elongation factors (12), the Tt-L12 sequences from the ␣4 and ␣5 regions gave only a slight stimulation of the GTPase activity (Fig. 6A). Therefore, it seems to be unlikely that the enhancement of GTPase activity by Tt-L12 is due to an increased binding of EF-G. The detailed role of ␣6 in EF-G-dependent GTPase turnover is not known at present. It is interesting that the amino acid residues in the ␣6 region of Tt-L12 that differ from the corresponding ones in Ec-L12 (Gln 105 , Glu 106 , Glu 109 , Ile 110 , Lys 113 , Ala 116 , and Val 117 ; Fig. 5A) are found on the same side of the helix (Fig. 6B) and that this side faces toward ␣4, to which the translation factors presumably bind (12). We infer that ␣6 can interact with ␣4 within L12, at least during certain stages of factor binding, and this interaction between ␣4 and ␣6 may be relevant for the enhancement of GTPase turnover by Tt-L12. The interaction may be involved in the rearrangement of the ribosome⅐EF-G⅐GDP complex that causes dissociation of EF-G from the ribosome. Furthermore, we should not rule out the possibility of a direct or indirect role for ␣6 in the activation of FIGURE 6. Identification of the structural element in the C-terminal domain of L12 that is responsible for the high EF-G-dependent GTPase turnover. A, each L12 protein (80 pmol of Ec-L12, Ec-␣3Tt-L12, Ec-␣4Tt-L12, Ec-␣5Tt-L12, or Ec-␣6Tt-L12) was preincubated with Ec-L10 (10 pmol) and E. coli 50 S core particles (2.5 pmol), together with Ec-L11 (10 pmol), and then assayed for EF-G-dependent GTPase activity in the presence of 30 S subunits (7.5 pmol). B, location of helices ␣4, ␣5, and ␣6 in the E. coli L12 CTD. The amino acid residues within ␣4 and ␣6 that differ between Ec-L12 and Tt-L12 are shown in black. The figure was created using PyMOL (Protein Data Bank code 1CTF).

FIGURE 7. EF-G binding features of the Ec-Tt-L12-containing ribosome.
A, inhibitory effect of fusidic acid on EF-G-dependent GTPase turnover. The Ec-Tt-L12-containing ribosome was reconstituted with Ec-Tt-L12, Ec-L10, Ec-L11, and the E. coli 50 S core, as in Fig. 4. EF-G-dependent GTPase activity was assayed in the presence and absence of 1 mM fusidic acid. B, footprinting of the sarcin/ricin loop of 23 S rRNA upon EF-G binding. The Ec-Tt-L12-containing ribosomes (lanes 4 -6) and the control Ec-L12-containing ribosomes (lanes 1-3) were incubated with fusidic acid alone (lanes 2 and 5), with EF-G, GTP, and fusidic acid (lanes 3 and 6), or without them (lanes 1 and 4) at 37°C for 5 min. These samples were then treated with dimethyl sulfate at 37°C for 10 min. The dimethyl sulfate modifications were detected by primer extension and sequencing gel, as described under "Experimental Procedures."