A Mode of Assembly of P0, P1, and P2 Proteins at the GTPase-associated Center in Animal Ribosome

Ribosomal P0, P1, and P2 proteins, together with the conserved domain of 28 S rRNA, constitute a major part of the GTPase-associated center in eukaryotic ribosomes. We investigated the mode of assembly in vitro by using various truncation mutants of silkworm P0. When compared with wild type (WT)-P0, the C-terminal truncation mutants CΔ65 and CΔ81 showed markedly reduced binding ability to P1 and P2, which was offset by the addition of an rRNA fragment covering the P0·P1-P2 binding site. The mutant CΔ107 lost the P1/P2 binding activity, whereas it retained the rRNA binding. In contrast, the N-terminal truncation mutants NΔ21-NΔ92 completely lost the rRNA binding, although they retained P1/P2 binding capability, implying an essential role of the N terminus of P0 for rRNA binding. The P0 mutants NΔ6, NΔ14, and CΔ18-CΔ81, together with P1/P2 and eL12, bound to the Escherichia coli core 50 S subunits deficient in L10·L7/L12 complex and L11. Analysis of incorporation of 32P-labeled P1/P2 into the 50 S subunits with WT-P0 and CΔ81 by sedimentation analysis indicated that WT-P0 bound two copies of P1 and P2, but CΔ81 bound only one copy each. The hybrid ribosome with CΔ81 that appears to contain one P1-P2 heterodimer retained lower but considerable activities dependent on eukaryotic elongation factors. These results suggested that two P1-P2 dimers bind to close but separate regions on the C-terminal half of P0. The results were further confirmed by binding experiments using chimeric P0 mutants in which the C-terminal 81 or 107 amino acids were replaced with the homologous sequences of the archaebacterial P0.

The ribosomal large subunits from all organisms contain an active site termed the "GTPase-associated center" that is responsible for the GTPase-related events in protein biosynthesis. This active site is composed of the two highly conserved domains around 1070 and 2660 (Escherichia coli numbering is used throughout) of 23 S/28 S rRNA and the ribosomal proteins bound to the 1070 region (1)(2)(3). The protein components of this site in prokaryotic ribosomes constitute a characteristic pentameric complex, L10(L7/L12) 2 (L7/L12) 2 (4,5), in which two L7/L12 homodimers bind to the C-terminal regions of L10 (6) and constitute a highly flexible and functionally important lateral protuberance, the so-called "stalk" (7). Although the ribosomal stalk is observed by cryo-electron microscopy (8), the detailed structure of this pentameric complex has not been resolved by x-ray crystallography of ribosomes (9 -11). The chemical features of protein-protein and protein-rRNA interactions in the GTPase-associated center remain to be clarified.
The animal ribosomal phosphoproteins P0 and P1/P2 (P proteins) are counterparts of prokaryotic L10 and L7/L12, respectively, although P1 and P2 are related but different proteins, unlike prokaryotic L7/L12 (12)(13)(14). In yeast cells, there are two P1-type proteins, P1␣ and P1␤, and two P2-type proteins, P2␣ and P2␤ (15). It is believed that P proteins constitute a pentameric complex, designated here as P0⅐P1-P2, in the GTPase-associated center of eukaryotic ribosomes (16,17). This complex binds not only to eukaryotic 28 S rRNA but also cross-binds to prokaryotic 23 S rRNA (18) and determines the specificity of the ribosome for eukaryotic elongation factor 1␣ (eEF-1␣) 2 and 2 (eEF-2) (19). This strong dependence on the P0⅐P1-P2 complex for the factor accessibility suggests the direct interaction between the protein complex and elongation factors. It has also been suggested that the P0⅐P1-P2 complex modulates the functional structures of the sarcin/ricin domain around 2660 as well as the 1070 regions of 23 S/28 S rRNA and makes them accessible to eukaryotic elongation factors (20). Knowledge of the molecular details on the assembly of P0, P1, and P2 proteins onto rRNA is essential to clarify protein-dependent function of the GTPase-associated center.
Current biochemical and genetic evidence indicates that P1 and P2 proteins form the heterodimer (16,(21)(22)(23)(24) and P1-P2 dimers bind to a specific region within the C-terminal domain of P0 (25)(26)(27). On the other hand, the rRNA binding site seems to be located within the N-terminal region comprising about 200 amino acids (25), although direct evidence has not been shown. In the case of mammalian counterparts, isolated P0 is insoluble in aqueous solution, but the P1-P2 binding to P0 makes P0 soluble (23,28). We recently showed that silkworm P0, however, is soluble and useful for biochemical assays (16). By using the silkworm proteins, we demonstrated that binding of P1 and P2 to P0 induced the binding activity of P0 to rRNA (16). It is therefore conceivable that binding of P1-P2 to P0 at its C-terminal region affects the overall structure of P0.
To clarify the individual binding sites for two P1-P2 dimers and for rRNA on P0 in vitro, we here constructed 11 kinds of truncation mutants of silkworm P0 and used them for protein-protein and protein-rRNA binding experiments. We identified two neighboring sites for P1-P2 heterodimers within the C-terminal half and a crucially important site for rRNA binding at the N terminus of P0 and suggested that the protein-protein and protein-RNA bindings mutually affect each other. To evaluate the in vitro binding data on the basis of ribosome function, we used a hybrid ribosome system developed previously (19) in which E. coli L10⅐L7/L12 complex and L11 on the 50 S subunit were replaced with the eukaryotic counterparts P0⅐P1-P2 complex and eL12, respectively. Whenever efficient RNA binding could be observed, complexes containing the truncated P0 mutants bound to E. coli core ribosomes and induced activities dependent on eukaryotic elongation factors. It is interesting that ribosomes carrying a P0 variant accessible to only one P1-P2 heterodimer retained reduced but significant activity.

MATERIALS AND METHODS
Plasmid Construction, Protein Expression, and Purification-The cDNAs for Bombyx mori ribosomal proteins P0, P1, P2, and eL12 (a eukaryotic homologue of prokaryotic L11) were provided by Dr. K. Mita (National Institute of Agrobiological Sciences). The coding region in each cDNA was amplified by PCR (29), inserted to E. coli expression vector pET28c or pET3a (Novagen), and cloned. Proteins expressed in E. coli cells were purified, as described previously (16). The DNA fragments coding for various truncated P0 were also amplified by PCR using cDNA encoding full-length (WT, 1-316 amino acids) of P0 as a template (see Fig. 1). The overlapping PCR method (30) was used to construct the chimeric P0 mutants composed of the N-terminal 1-235 (C⌬81) and 1-209 (C⌬107) amino acid sequences of silkworm P0 fused to the C-terminal sequences of the archaebacterial (Pyrococcus horikoshii) P0-like protein, which are homologous to the 236 -316 and 210 -316 sequences of silkworm P0, respectively. The genomic DNA was used as a template of PCR to amplify the DNA fragments for the two C-terminal amino acid sequences of P. horikoshii P0. 3 Each DNA fragment was cloned into pET28c, and the P0 mutant was expressed and purified as described above.
In Vitro RNA Synthesis-The rat rDNA fragment containing residues 1841-1939 that correspond to 1029 -1127 of E. coli 23 S (designated here the 1070 domain) was amplified by PCR and inserted into the HindIII and XbaI sites of pSPT 18 (Roche Applied Science). The RNA fragment was synthesized using the plasmid DNA and SP-6 RNA polymerase and purified, as described previously (31).
P0⅐P1-P2 Complex Formation-After the amounts of isolated proteins were determined with the Micro BCA protein assay reagent kit (Pierce), the concentrations of individual protein samples (pmol/l) were estimated considering that 1 g of WT-P0 (or P0 mutants), P1, and P2 correspond to 29.3 (or 26.6 -42.2), 87.3, and 86.6 pmol, respectively. The protein samples were mixed together at a molar ratio of P0 sample: P1:P2 of 1:3:3, and the complex was reconstituted, as described previously (16). The P0⅐P1-P2 complex formation in the presence or absence of a 2-fold excess of the rRNA fragments of the 1070 domain was confirmed by 6% polyacrylamide (acrylamide/bisacrylamide ratio 39:1) native gel electrophoresis at 6.5 V/cm with a buffer system containing 5 mM MgCl 2 , 50 mM KCl, and 50 mM Tris-HCl (pH 8.0). Samples were electrophoresed for 6 h at constant voltage and 4°C with buffer recirculation (16). The gel was stained with Coomassie Brilliant Blue.
Ribosomal Subunits and the 50 S Core Particles-E. coli ribosomal subunits were prepared as described previously (20). 50 S cores deficient in L10⅐L7/L12 and L11 were prepared by extraction of the 50 S subunits from the L11-deficient E. coli mutant AM68 (32) in a solution containing 50% ethanol and 0.5 M NH 4 Cl solution at 0°C, as described previously (19).
Gel Retardation-[ 32 P]RNA fragments covering the 1070 domain (5 pmol) synthesized as described above were mixed with 10 pmol of P0⅐P1-P2 complex sample and 10 pmol of eL12 (28) and incubated at 30°C for 5 min in 10 l of a solution containing 20 mM MgCl 2 , 300 mM KCl, 20 mM Tris-HCl, pH 7.6 (16). The RNA-protein complexes were separated by 6% polyacrylamide native gel, as described above. The gel was dried and subjected to autoradiography.
Quantitative Analysis of P1 and P2 Incorporated into the Ribosome-Isolated P1 and P2 (25 g of each) were incubated with 500 units of casein kinase II (New England Biolabs) and 5 nmol of [ 32 P]ATP (250 Ci/mol) for 30 min at 30°C in a solution (50 l) containing 50 mM KCl, 10 mM MgCl 2 , 20 mM Tris-HCl, pH 7.5. The 32 P-labeled P1 (86 cpm/pmol) was mixed with non-labeled P0 (or C⌬81) and P2, and the complex was reconstituted as described above. For the E. coli 50 S core (78 pmol), excess amounts (195 pmol) of the P0⅐[ 32 P]P1-P2 complex were added, together with 154 pmol of eL12. The sample was then layered on a 10 -28% sucrose gradient in a solution containing 50 mM NH 4 Cl, 5 mM MgCl 2 , 5 mM 2-mercaptoethanol, and 20 mM Tris-HCl, pH 7.6, and fractionated after centrifugation at 40,000 rpm and 4°C for 3 h in a Hitachi P-45 ST rotor. The 50 S fraction was collected, and the amount of the associated P1 was estimated by its radioactivity. The 32 P-labeled P2 (105 cpm/pmol) was mixed with non-labeled P0 (or C⌬81) and P1, and its incorporation into 50 S core particles was analyzed as described for 32 P-labeled P1.

Preparation of Various Truncated Mutants of Animal Ribosomal
Protein P0-The eukaryotic ribosomal protein P0 plays a central role in the assembly of the active GTPase-associated center. Here we investigated the structural elements required for binding of two P1/P2 dimers and rRNA by using truncated mutants that are summarized in Fig. 1A. All P0 mutants including five C-terminal mutants (C⌬18, C⌬55, C⌬65, C⌬81, and C⌬107) and six N-terminal truncation mutants (N⌬6, N⌬14, N⌬21, N⌬48, N⌬66, and N⌬92) were expressed in E. coli cells but found to be insoluble. The proteins were solubilized in 6 M urea and purified by using ion exchange high pressure liquid chromatography. The purity of all the isolated P0 samples, P1, P2, and eL12, is shown in Fig. 1B. By adding P1 and P2 to individual P0 samples, P0⅐P1-P2 complexes were reconstituted as described under "Materials and Methods" and used in the following experiments.
Effect of the C-terminal Truncation of P0 on P0⅐P1-P2 Assembly-The formation of P0⅐P1-P2 complexes was examined by native polyacrylamide gel electrophoresis ( Fig. 2A). The complex of WT-P0 with P1 and P2 was detected as a shifted band ( Fig. 2A, lane 1), in a similar manner as the authentic proteins from silkworm ribosomes (16). The complexes were also formed efficiently even with P0 mutants, the C-terminal amino acids of which . Each protein mixture was tested for rRNA binding by gel mobility shift assay using a small amount of the 32 P-labeled RNA fragment covering the 1070 region (Fig. 2B). Strong RNA binding was observed for all reconstituted complexes (Fig. 2B, lanes 2-6). The C⌬107 mutant, which had no binding ability to P1/P2, showed reduced but significant binding affinity to the RNA (lane 7). The same mobility shift was observed in the isolated C⌬107 without P1 and P2 (not shown), suggesting that the C⌬107 mutant retains rRNA binding. This is an unexpected result because the WT-P0 fails to bind rRNA without P1 and P2 (16).
The same experiment as Fig. 2A was performed after the addition of an excess amount of the rRNA fragment (Fig. 2C). In the presence of the rRNA fragment, bands of the P0⅐P1-P2 complexes with all the P0 samples except C⌬107 were more distinct (Fig. 2C) than those in the absence of the RNA ( Fig. 2A). Particularly, the complexes with C⌬65 and C⌬81 were stabilized markedly with the RNA fragment (Fig. 2C,  lanes 4 and 5). The complex formation of C⌬107 with P1-P2 was not detected even after the addition of the RNA (lane 6). To confirm that the complexes formed in Fig. 2C all contain P0 (or its mutants), P1 and P2, the bands of the complexes were cut out of the gel and subjected to SDS gel electrophoresis. The gel was stained with a fluorescent dye, SYPRO Orange, to quantitate roughly the relative amounts of P1/P2 by fluorescence intensity (Fig. 2D). The three protein components were detected in all complexes formed. The intensity of both P1 and P2 in the complexes with C⌬65 and C⌬81 (lanes 4 and 5) was approximately half of that in the other complexes (lanes 1-3).
To eliminate a possibility that the effects of the truncation shown in Fig. 2 result from severe alteration of the tertiary structure of P0 rather than from deletion of the binding sites for P1/P2, we also performed binding experiments using chimeric P0 instead of truncation mutants. Because archaebacterial (P. horikoshii) ribosomes contain a eukaryotic P0-like protein that does not cross-bind to silkworm P1/P2, we replaced the C-terminal 81-and 107-amino acid sequences of silkworm P0 with the homologous sequences of P. horikoshii P0. As shown in Fig. 3, chimeric P0 samples did not form stable complexes with silkworm P1/P2 in the absence of the rRNA fragments (lanes 2 and 3). In the presence of the rRNA fragment, however, the chimeric P0, the C-terminal 81-amino acid sequence of which is from the corresponding region of P. horikoshii protein, could form a complex with silkworm P1/P2 (lane 5). In contrast, the chimeric P0, the C-terminal 107-amino acid sequence of which is from P. horikoshii, had no ability to bind silkworm P1/P2 (lane 6). The abilities of chimeric P0 mutants to complex with    1 and 4), a chimeric mutant composed of 1-235 amino acid residues of silkworm P0 fused with 238 -342 residues of P. horikoshii P0 (lanes 2 and 5) and another chimeric mutant composed of 1-209 residues of silkworm P0 fused with 212-342 residues of P. horikoshii P0 (lanes 3 and 6). The complexes formed in the absence of the rRNA fragment (lanes 1-3) and in its presence (lanes 4 -6) were analyzed by native gel electrophoresis, as in Fig. 2, A and C. P1/P2 are comparable with those of the C⌬81 and C⌬107 P0 mutants (Fig. 2C). It should be added that the chimeric P0, the C-terminal 107amino acid sequence of which is from P. horikoshii, could bind the P. horikoshii stalk dimers and form a functional complex. 3 These results supported the binding data with the C-terminal truncation mutants.
Effect of the N-terminal Deletion of P0 on P0⅐P1-P2 Assembly-The formation of P0⅐P1-P2 complexes was also examined with the N-terminal truncation mutants of P0 (Fig. 4A). Unlike the C-terminal truncation, all the N-terminal mutants N⌬6 -N⌬92 formed complexes that appeared as distinct bands in the presence of P1-P2 (lanes 2-7). The SDS gel electrophoretic analysis of the complexes showed that all contained P0 (or its mutants), P1, and P2 (data not shown), suggesting that the N-terminal truncations did not disrupt the binding potentials between P0 and P1-P2. In contrast, the N-terminal deletions caused marked effect of the rRNA binding (Fig. 4B). When N-terminal amino acids 1-6 (N⌬6) were truncated, the rRNA binding ability of P0⅐P1-P2 complex was reduced (Fig. 4B, lane 3). By further deletions (N⌬14 -N⌬92, lanes 4 -8), the binding ability was lost. However, the rRNA bindings of the complexes with N⌬6 and N⌬14, but not with N⌬21-N⌬92, were recovered by the addition of eL12 (data not shown).
Assembly of the Truncated P0 Mutants onto the E. coli 50 S Particles-Unlike prokaryotic L10, eukaryotic P0 was hardly released from the animal 60 S subunit by standard high salt/ethanol conditions, and reconstitution experiments of the GTPase-associated center with animal ribosomes are much harder than those with E. coli ribosomes. We have established conditions to form a hybrid ribosomal particle in which L10⅐L7/L12 complex and L11 within the E. coli 50 S subunit are replaced with animal P0⅐P1-P2 and eL12, respectively (19). We therefore used the E. coli 50 S subunits to investigate the assembly of animal P0⅐P1-P2 and    A and B, left) and the C⌬81  mutant (A and B, right) by sucrose density gradient centrifugation. A, the 50 S core particles (78 pmol) were mixed with excess amounts of P0⅐P1-P2 complexes consisting of [ 32 P]P1 (86 cpm/ pmol P1), non-labeled P2, and WT-P0 (left) or the C⌬81 mutant (right), together with eL12. The samples were separated by sucrose gradient centrifugation, as described under "Materials and Methods." The amount of P1 incorporated to the 50 S core was estimated by radioactivity of P1. B, the same experiment as in A was performed by using P0⅐P1-P2 complexes consisting of [ 32 P]P2 (105 cpm/pmol P2), non-labeled P1, and WT-P0 (left) or the C⌬81 mutant (right). eL12 forming an active GTPase-associated center. The formation of hybrid 50 S particles by mixing E. coli 50 S cores with various truncated P0 mutants, P1/P2, and eL12 was confirmed by acrylamide/agarose composite gel electrophoresis (Fig. 5). The gel mobility of E. coli 50 S cores (Fig. 5, A and B, lanes 1) was shifted upwards by the addition of only recombinant silkworm eL12 (Fig. 5, A and B, lanes 2) and then supershifted by the further addition of WT-P0⅐P1-P2 complex (Fig. 5, A  and B, lanes 3), indicating efficient binding of both proteins to E. coli 50 S core particles. Likewise, the complexes with the C-terminal truncation mutants of P0 (C⌬18-C⌬81) strongly bound to the core particles (Fig.  5A, lanes 4 -7), whereas a clear supershift was not detected by the addition of C⌬107 (lane 8). In the case of the complexes with the N-terminal deletion mutants of P0, the supershifts were observed only in the N⌬6and N⌬14-containing complexes (Fig. 5B, lanes 4 and 5). By further deletions of P0, the ability of ribosome binding was completely lost (lanes 6 -9).
As shown in Fig. 2A, there is a notable difference in P1/P2 binding property between a group including WT-P0, C⌬18, and C⌬55 and another group of C⌬65 and C⌬81. The former members could bind efficiently P1/P2 without the RNA fragment, whereas the latter mutants required the rRNA fragment. To quantitate the amounts of P1 and P2 bound to P0 and compare them between the two groups, P1 and P2 were labeled in vitro by 32 P phosphorylation and then incorporated into the hybrid 50 S particles with WT-P0 (Fig. 6, A and B, left) and C⌬81 (Fig. 6,  A and B, right). After sucrose gradient centrifugation, the hybrid particles were collected, and the incorporated P1 and P2 were quantified. When WT-P0 was used, 2.2 copies of P1 (Fig. 6A, left) and 1.7 copies of P2 (Fig. 6B, left) were incorporated per 50 S particle. However, when C⌬81 was used, only 1.1 copies of P1 (Fig. 6A, right) and 1.0 copy of P2 (Fig. 6B, right) were present on the 50 S particles. The results are consistent with those in Fig. 2D and indicate that the WT-P0 binds two copies of both P1 and P2 but that C⌬81 binds only one copy each. Considering previous evidence that P1 and P2 form a heterodimer (16,(21)(22)(23)(24) together with the present data, we infer that WT-P0, C⌬18, and C⌬55 bind two heterodimers, but C⌬65 and C⌬81 bind only one.
Activities of the Hybrid Ribosomes with the Truncated P0 Mutants-The hybrid 50 S particles carrying various P0 mutants shown in Fig. 5 were tested for their functions with E. coli 30 S subunits in the presence of eukaryotic elongation factors eEF-1␣ and eEF-2, as described previously (19). Concerning eEF-2-dependent GTPase, the hybrid ribosomes carrying C⌬18, C⌬55, N⌬6, and N⌬14 P0 mutants showed activity comparable with that with WT-P0 (Fig. 7A). In polyphenylalanine synthesis (Fig. 7B), the activities of the C⌬18-and C⌬55-containing hybrid ribosomes were 80 -90% of those with WT-P0, whereas the activities of the N⌬6-and N⌬14-containing hybrid ribosomes were slightly higher than those of WT-P0 ribosomes. The C⌬65-and C⌬81-containing hybrid ribosomes carrying only one P1-P2 heterodimer showed similarly reduced activities in both the GTPase assay and polyphenylalanine synthesis (70 and 40% of the WT, respectively). The C⌬107-containing hybrid ribosomes as well as the N-terminal deletion mutants N⌬21-N⌬92 showed only background activities comparable with those detected in the absence of P0, P1, and P2.

DISCUSSION
Two P1-P2 Heterodimers Bind to the Restricted C-terminal Regions of P0-It has been shown that the C-terminal fragment of E. coli L10, composed of amino acid residues 71-164, participates in bindings of two L7/L12 dimers (35). More recently, Griaznova and Traut (6) have clarified that 20 amino acids at the C terminus of E. coli L10 determine the binding sites for all four copies of L7/L12 proteins, i.e. the region of amino acids 145-154 is involved in the binding of one L7/L12 dimer, and the following amino acid sequence 155-164 is for another L7/L12 dimer (Fig. 8A). After completion of this manuscript, Diaconu et al. (36) reported the crystal structure of the L10⅐L7/L12 complex in Thermotoga martima, and the results surprisingly revealed three repetitive binding sites for L7/L12 dimers on the discrete C-terminal helix ␣-8. In yeast, the regions including residues 230 -290 (25), 212-262 (26), and 213-250 (27) within the C-terminal half of P0 have been identified as sites for P1/P2 binding (Fig. 8A), although individual sites for two P1-P2 dimers could not be resolved.
Our present evidence suggested that there are two neighboring and distinct binding sites for P1-P2 in the C-terminal half of silkworm P0, i.e. the regions of residues 252-261 and 210 -251 of silkworm P0 bind either one P1-P2 heterodimer. To compare the present data with those in E. coli (6), amino acid sequences of the C-terminal halves of silkworm (B. mori) P0 and E. coli L10 are aligned, together with yeast (Sc) and human (Hs) P0 (Fig. 8A), basically according to Shimmin et al. (37), who compared the sequences of L10 equivalent proteins from eubacteria, archaebacteria, and eukaryotes. In this alignment, the region of residues 252-261 of silkworm P0, participating in binding of P1-P2 at the C-terminal side, neatly corresponded to the sequence 155-164 of E. coli L10, required for binding of one L7/L12 homodimer at the C-terminal side. The region of residues 210 -251 of P0, identified as site-essential for another P1-P2 binding, covered the site corresponding to region of residues 145-154 of L10, which is required for the binding of another L7/L12 dimer. There may be an analogous feature in interaction between L10 equivalent proteins and the stalk dimers, although the amino acid sequence identity of the putative binding site for the stalk dimers in the L10 orthologs are very low.
The N Terminus of P0 Is Crucial for rRNA Binding-It has been reported that E. coli L10 and L10⅐L7/L12 complex bind directly to the 1070 region of 23 S rRNA (3,38,39), and the L10 fragment lacking the N-terminal 1-69 amino acids fails to assemble into the ribosome, together with L7/L12 (35). In eukaryotes, the yeast P0 variant lacking the C-terminal 87 residues can assemble into the ribosome in vivo with weakly bound P1/P2 (25). Moreover, P0⅐P1-P2 complexes formed in vitro have strong rRNA binding ability (28). These previous data suggest that the N-terminal halves of prokaryotic L10 and eukaryotic P0 contain the binding site for the 1070 domain of 23 S/28 S rRNA. However, further detailed biochemical studies on binding between L10/P0 and rRNA have not been carried out in prokaryotes or in eukaryotes. Here we have reported a crucial role for the N terminus of P0 in both rRNA binding and assembly into the ribosomal GTPase-associated center.
The N⌬6 and N⌬14 mutants seem to retain reduced rRNA binding ability that is enhanced with eL12, presumably by their cooperative rRNA binding (40). However, the N⌬21-N⌬91 variants showed no ability in rRNA binding and failed to assemble into 50 S core particles even with eL12. The present results strongly suggested that the N terminus of P0 including residues 1-21, particularly residues 15-21, plays an essential role in binding of the protein complex to the 1070 rRNA domain.
A possible alignment of the N-terminal amino acid sequences of P0 homologues among five eukaryotic species is shown in Fig. 8B. There are highly conserved amino acids Lys-10, Tyr-13, Phe-14, and Lys-16 (BmP0 numbering). Two conserved acidic amino acids Asp/Glu-22 and Asp/Glu-23 are also notable. These amino acids seem to constitute at least a part of the rRNA binding site in P0. In the currently published crystal data of an L10-like protein (36), this N-terminal region (residues 8 -21 of BmP0) corresponds to helix ␣1 that contacts rRNA.
Cooperative Relation among Bindings of two P1-P2 Dimers and rRNA to P0-As discussed above, silkworm P0 protein seems to have two binding sites for P1-P2 heterodimers defined by residues 210 -251 and 252-261. The P0 truncation mutants C⌬18 and C⌬55, as well as WT-P0, which appear to retain the two binding sites, form a stable complex with P1 and P2, whereas C⌬65 and C⌬81, which seem to have a single binding site, show only weak binding property to P1-P2 in the absence of the rRNA to which P0 binds (Fig. 2A). These findings indicate that the residues 210 -251-mediated P1-P2 binding is markedly enhanced in the presence of residues 252-261 and binding of the latter to another P1-P2 dimer. There may be two possible explanations for the cooperativity of the binding of the two P1-P2 dimers: 1) the 252-261-mediated binding of P1-P2 may cause a conformational change of P0 that stimulates the 210 -251-mediated binding and 2) the pentameric complex may be stabilized by interactions between the two P1-P2 dimers through P1-P1 and P2-P2 interactions, which has been suggested by previous biochemical assays (41,42) and two hybrid assays (15,21), although the latter assay shows that efficiency of P1-P1 and P2-P2 interactions is much lower than that of P1-P2 interaction. The present study also shows that the 210 -251-mediated P1-P2 binding of P0 is stabilized by its rRNA binding (Fig. 2C). On the contrary, strong rRNA binding of P0 is induced by P1/P2 binding to P0, as described previously (16). Furthermore, the present study shows that the truncation of two P1-P2 binding sites of P0 (C⌬107) increases its rRNA binding ability without P1/P2 (Fig. 2B and see also "Results"). These results indicate a correlation between P1-P2 binding to the C-terminal regions of P0 and rRNA binding to the N-terminal region. The rRNA binding site in P0 may be blocked with a part of the P1-P2 binding sites, when P1 and P2 are absent. It is also likely that the N-terminal end constituting a part of the rRNA binding site affects a state of the P1-P2 binding site. It is generally believed that the ribosomal stalks, eukaryotic P1/P2, and prokaryotic L7/L12 participate in interaction with translation factors. The present study suggested an additional role of P1/P2 as a modulator in P0 binding to rRNA.
The Significance of the Presence of Two P1-P2 Dimers in the Ribosome-It is interesting that the hybrid ribosome carrying C⌬81, which binds only one P1-P2 dimer, retains reduced but significant activity dependent on eukaryotic elongation factors (Fig. 7). The results indicate that amino acid residues 236 -316 of P0 including one of the two P1-P2 binding sites and the C terminus homologous to that of P1 and P2 (13) are not essential for basal ribosome activity and suggest that two P1-P2 dimers are required for full activity of both GTPase and polyphenylalanine synthesis. A dispensable feature of one of two stalk dimers has also been shown in E. coli L7/L12 (6,43,44). We infer that, at least in eukaryotes, the presence of two P1-P2 dimers in the ribosome improves efficiencies of GTPase turnover and thus the elongation cycle, as well as the efficiency of the assembly of the GTPase-associated center, as described above.