The RNA Interacting Domain but Not the Protein Interacting Domain Is Highly Conserved in Ribosomal Protein P0*

Protein P0 interacts with proteins P1α, P1β, P2α, and P2β, and forms the Saccharomyces cerevisiaeribosomal stalk. The capacity of RPP0 genes fromAspergillus fumigatus, Dictyostelium discoideum, Rattus norvegicus, Homo sapiens, and Leishmania infantum to complement the absence of the homologous gene has been tested. In S. cerevisiae W303dGP0, a strain containing standard amounts of the four P1/P2 protein types, all heterologous genes were functional except the one from L. infantum, some of them inducing an osmosensitive phenotype at 37 °C. The polymerizing activity and the elongation factor-dependent functions but not the peptide bond formation capacity is affected in the heterologous P0 containing ribosomes. The heterologous P0 proteins bind to the yeast ribosomes but the composition of the ribosomal stalk is altered. Only proteins P1α and P2β are found in ribosomes carrying the A. fumigatus, R. norvegicus, and H. sapiens proteins. When the heterologous genes are expressed in a conditional null-P0 mutant whose ribosomes are totally deprived of P1/P2 proteins, none of the heterologous P0 proteins complemented the conditional phenotype. In contrast, chimeric P0 proteins made of different amino-terminal fragments from mammalian origin and the complementary carboxyl-terminal fragments from yeast allow W303dGP0 and D67dGP0 growth at restrictive conditions. These results indicate that while the P0 protein RNA-binding domain is functionally conserved in eukaryotes, the regions involved in protein-protein interactions with either the other stalk proteins or the elongation factors have notably evolved.

The ribosomal stalk is an important structural element of the large subunit, which has been proposed to have a role in the translocation step of protein synthesis (1,2). It was shown to be involved in the activity of the elongation factors in bacteria (3) as well as in eukaryotes (4,5), and a direct interaction with EF-Tu and EF-G has recently been confirmed by electron microscopy (6,7). Moreover, the stalk may also have a regulatory role for the activity of the eukaryotic ribosome (8).
The bacterial stalk is formed by a pentamer made of two dimers of protein L7/L12 and protein L10, that interacts with one of the most highly conserved regions of the large rRNA, the so-called GTPase center (9,10). The similar eukaryotic pentameric complex is made of four 12-kDa acidic proteins and the 34-kDa protein P0, which are equivalents to L7/L12 and L10, respectively. In eukaryotes having two types of acidic proteins, called P1 and P2, they seem to be present as dimers in the stalk (11). The acidic proteins interact through their NH 2 -terminal domain (12) with P0 in a region close to the carboxyl end (13). The whole protein complex binds to the rRNA GTPase center through the NH 2 -terminal domain of protein P0 (13,14). The COOH-terminal end of the proteins is exposed to the medium and interacts with the elongation factors (6,7).
In quite a few eukaryotic species there are more than one protein of the P1 and P2 types. A third protein type, P3, has even been reported in plants (15). In Saccharomyces cerevisiae there are two P1 proteins, P1␣ and P1␤, and two P2 proteins, P2␣ and P2␤. The four yeast stalk proteins share the same COOH-terminal peptide, EESDDDMGFGLFD, which is also present in protein P0, but show substantial differences in the rest of the amino acid sequence. Thus, the overall sequence similarity in the two proteins of the same type is only close to 80% and their function is not totally equivalent (8).
As an average, there are about four 12-kDa acidic proteins per yeast ribosome (16). Therefore, if the yeast acidic proteins are present as dimers there cannot be two copies of each protein per particle, and, consequently, the ribosome population should be heterogeneous regarding the stalk composition. There is, nevertheless, evidence suggesting that, contrary to mammals, the acidic proteins can be as monomers in the yeast ribosome (17). However, further experimental data are required to totally resolve this question. In any case, independently of the copy number, the presence of at least one protein of the P1 type and one of the P2 type is required to form a complex with protein P0. Thus, in S. cerevisiae D67, a mutant in which the two genes encoding the P1␣/P1␤ proteins have been disrupted, the P2␣/P2␤ proteins accumulate in the cytoplasm but are not found bound to the ribosome (18).
Although the bacterial and eukaryotic stalk proteins seem to play a similar role with respect to their basic function in the ribosome, they have substantial structural differences (19,20,13,21). Protein P0 shows a higher degree of structural and functional complexity than the bacterial protein L10 (20,13). The eukaryotic protein has a carboxyl end extension of around 100 amino acids that resembles the amino acid sequence of proteins P1 and P2 (19). This extension plays an important role in the interaction with the P1 and P2 proteins and, therefore, in the formation of the pentameric complex (13). In addition, it is able to perform the functions of the complete stalk in the absence of the other components (22). Protein P0 by itself constitutes the minimal stalk sufficient to support accurate protein synthesis although at a lower rate than the complete stalk (23). The binding of P1 and P2 proteins, which are present also in an exchangeable cytoplasmic pool, increases the effi-ciency of the ribosome. This evolutionary peculiarity of the eukaryotic P0 protein seems to have provided the eukaryotic ribosome a way to regulate the translation process that is missing in the prokaryotic organisms (see Ref. 8, for review).
Protein P0 also binds to the GTPase site in the eukaryotic large rRNA. This highly conserved region in the RNA molecule has been shown to be interchangeable between bacteria and eukaryotes (24,25). However, the bacterial and the eukaryotic proteins differ in respect to the RNA binding since, while protein L10 is easily removed from the ribosome by washing with ammonium-ethanol buffers (26), the eukaryotic polypeptide resists this treatment and remains tightly bound to the ribosome (27).
Two additional functional domains can be clearly defined in the P0 protein in addition to the RNA-binding one, the region involved in the formation of the pentameric complex with P1 and P2, and the region connected to the interaction with the elongation factors, both probably located in the COOH-terminal part.
Data on the capacity of P0 protein from different organisms to complement the different functions of the endogenous protein in S. cerevisiae, together with a comparative analysis of the respective amino acid sequences can provide information relative to the functional role of the conserved and non-conserved regions and can help to define more precisely the protein active domains. This type of study has been carried out expressing the RPP0 gene from five different species in a conditional P0 null mutant of S. cerevisiae that carries the genomic RPP0 gene under the control of the inducible GAL1 promoter.
Yeasts were grown in either YEP medium (1% yeast extract, 2% peptone) or minimal YNB medium, supplemented with the necessary nutritional requirements. In both cases, the carbon source was either 2% glucose or 2% galactose as indicated. Stock cultures of all yeast strains were maintained in galactose medium. When required, cells were shifted to glucose medium and allowed to grow for at least 20 generations to reach steady state growth conditions. Escherichia coli DH5␣ was used as a host for the routine maintenance and preparation of plasmids and was grown in LB medium.

Enzymes and Reagents
Restriction endonucleases were purchased from Roche Molecular Biochemicals, MBI Fermentas, New England Biolabs, and Amersham Pharmacia Biotech, and were used as recommended by the suppliers. T4 DNA ligase, calf intestinal alkaline phosphatase, and the DNA polymerase I Klenow fragment were from Roche Molecular Biochemicals. DNA manipulations were done basically as described in Ref. 28. PCR 1 was carried out using PFU DNA polymerase from Perkin Elmer and custom made oligonucleotides from Isogen, following the recommendations of Dieffenbach and Dveksler (29).
Plasmids pFL37 was derived from pFL38 (30) by removing the URA3 marker with BglII and introducing in the same position a BamHI fragment carrying the HIS3 marker. BSP0 was obtained by inserting a 2.8kilobase pair EcoRI-AvaII fragment containing the yeast RPP0 gene in the MCS of Bluescript (13). BSP0Sc(NdeI) was derived from BSP0 by introducing a NdeI restriction site (CATATG) at the initiator ATG in the RPP0 gene by heteroduplex mutagenesis (31). pUC-P0Dd resulted from the insertion into the MCS of pUC19 of a 1-kilobase pair EcoRI cDNA fragment from a positive phage, containing 10 bases of the 5Ј-UTR, the coding region and 42 bases of the 3Ј-UTR of the Dictyostelium discoideum RPP0 gene. The positive phage was cloned from an expression library using antibodies specific to the D. discoideum protein (32).
pFL37-P0Sc-A SmaI-XhoI fragment, containing the 5Ј-UTR, the coding region and the 3Ј-UTR of S. cerevisiae RPP0 gene from plasmid BSP0, was inserted in the corresponding sites of the pFL37 vector.
pFL37-P0Dd-The D. discoideum RPP0 gene as a Klenow-filled 1 -kilobase pair EcoRI fragment from a pUC-P0Dd (32), was subcloned in the blunt-ended NdeI-EcoRI sites of pBSP0Sc(NdeI), substituting the coding region of the S. cerevisiae gene and yielding the plasmid BSP0Dd. When inserted in the correct orientation, the D. discoideum gene in BSP0Dd was under the control of the yeast RPP0 promoter. Afterward, the SmaI-XhoI fragment from BSP0Dd was introduced in the corresponding sites of pFL37.
pFL37-P0Rn, pFL37-P0Af, pFL37-P0Li, and pFL37-P0Hs-In the four cases a similar strategy was followed. Using appropriate oligonucleotides, the coding region of the P0 protein gene obtained by PCR from either a cDNA library (Aumigatus fumigatus) or previously prepared cDNA clones (Homo sapiens, Rattus norvegicus, and Leishmania infantum). The oligonucleotides were designed to introduce a NdeI site at the initiation end and either an EcoRI site (R. norvegicus and L. infantum) or a NehI site (A. fumigatus and H. sapiens) at the termination end of the coding region. The PCR product was digested with NdeI and either EcoRI or NehI and introduced in the corresponding sites of BSP0Sc(NdeI), obtaining plasmids with the heterologous gene under the control of the yeast P0 promoter. The fragment containing the hybrid gene was subcloned in pFL37.
pFL37-P0HHY-Using a conserved EcoRV restriction site present at an equivalent position in both S. cerevisiae and H. sapiens RPP0 genes, plasmids pFL37-P0Sc and pFL37-P0Hs were treated with EcoRV and EcoRI and the fragment 0.7-kilobase pair fragment derived from the yeast gene, containing the COOH-terminal region of the protein, was subcloned into the restricted pFL37-P0Hs. The resulting pFL37-P0HHY plasmid contains a chimera that encodes a 314-amino acid long protein composed by the first 205 amino acids and the last 109 amino acids from the human and the yeast polypeptide, respectively.
pFL37-P0HYY-This plasmid was prepared taking advantage of a 30-nucleotide long stretch starting around position 400 that is identical in the H. sapiens and S. cerevisiae genes. Using complementary oligonucleotides to both strands of this region as well as the universal primers at the other end, two PCR fragments were obtained from plasmids BSP0Hs and BSP0Sc. The fragments comprising the 5Ј and 3Ј regions were derived from the human and yeast genes, respectively. These two fragments together with the universal primers were used for an overlap extension PCR (33) to obtain a chimeric P0HYY gene, which was subcloned as a XhoI-EcoRI fragment into the pFL37 vector. The plasmid was checked by restriction analysis and by DNA sequencing. The protein encoded by the P0HYY gene contains the first 138 amino acids from H. sapiens and the last 176 amino acids from S. cerevisiae.

Cell Transformations
Bacterial transformations were performed according to the procedure of Hanahan (34). Yeasts were transformed using the lithium acetate method as described (35).

Cell Fractionation and Ribosome Preparation
Yeasts were grown exponentially in rich YEP medium up to A 600 ϭ 1 and cells were collected by centrifugation and washed with buffer 1 (100 mM Tris-HCl, pH 7.4, 20 mM KCl, 12.5 mM MgCl 2 , 5 mM ␤-mercaptoethanol). Cells in buffer 1 were supplemented with protease inhibitors (0.5 mol of phenylmethylsulfonyl fluoride, 1.25 g of leupeptin, aprotinin, and pepstatin per gram of cells), and broken with glass beads. The extract was centrifuged in a Beckman SS-34 rotor (12,000 rpm, 15 min, 4°C), yielding the supernatant 30 fraction which was afterward submitted to high speed centrifugation at 90,000 rpm for 30 min at 4°C in a Beckman TL100.3 rotor. The supernatant S100 fraction was stored at Ϫ80°C and the crude ribosome pellet was resuspended in buffer 2 (20 mM Tris-HCl, pH 7.4, 500 mM AcNH 4 , 100 mM MgCl 2 , 5 mM ␤-mercaptoethanol) and centrifuged through a discontinuous sucrose gradient (20/40%) in buffer 2 at 90,000 rpm for 120 min at 4°C in a TL100.3 rotor. The pellet of washed ribosomes was dissolved in buffer 1 and stored at Ϫ20°C.

Protein Analysis
Ribosomal proteins were analyzed either by 15% SDS-PAGE or isoelectrofocusing. Isoelectrofocusing was carried out as described previously (36). Particles were pretreated with RNase A (10 g/mg ribosomes) on ice for 30 -45 min. After lyophilization, the samples containing 0.5 mg of ribosomes were resuspended in loading buffer (6% ampholytes, 8 M urea) and directly loaded into a standard vertical gel (5% acrylamide, 0.2% bis-acrylamide, 6 M urea, 6% pH 2.5-5.0 ampholytes). As cathode and anode solutions 30 mM NaOH and 180 mM H 2 SO 4 were used at the upper and bottom part of the gel, respectively. Isoelectrofocusing was run in the cold room at 6 mA constant current until the voltage reached 600 V and then at 250 V for 16 h.
Proteins were usually detected by standard silver staining. Alternatively, gels were stained in a solution containing 0.25% Coomassie Blue R-250 (Sigma) dissolved in 45% ethanol, 10% acetic acid. After 30 min, gel was destained using the same solution but without Coomassie Blue. Scanning of the bands in the gels was performed using a Molecular Dynamics computing densitometer model 300A. Protein sequencing was performed by Edman degradation using an Applied Biosystems 447 automatic peptide sequenator at the Centro de Biología Molecular "Severo Ochoa" Protein Sequencing Service.

Western Blotting
Proteins in gels were transferred to membranes by electrophoresis in a semi-dry system using Novablot LKB buffer. The membranes were treated with 5% non-fat milk dissolved in TBS (10 mM Tris-HCl, pH 7.4, 200 mM NaCl) for 30 min and afterward they were incubated for 1 h with the antibody diluted in the same buffer. Subsequently, the membranes were washed 15 min in TBS containing 5% non-fat milk and 0.1% Tween 20. Then, the second antibody (R␣M/PO or G␣R/PO), diluted in the former buffer, was added and the membranes were incubated for 30 min. Finally, they were washed 15 min with 0.1% Tween 20 in TBS. Bound antibodies were located by detecting peroxidase activity using the ECL system (Amersham Pharmacia Biotech) and then exposed to film.
Peptide bond formation was estimated by the "fragment reaction" (37) using N-acetyl-tRNA as substrate. The reaction mixture, in 150 l of 20 mM Tris-HCl, pH 7.3, 270 mM NH 4 Cl, and 13 mM MgCl 2 , contained 1 mg/ml ribosomes, 2 mM puromycin, and 1 pmol of N-[acetyl-3 H]tRNA (2500 cpm/pmol). The reaction was initiated by the addition of 1 volume of methanol (33% final concentration), kept at 0°C for 30 min, and then stopped by the addition of 100 l of 0.3 M sodium acetate, pH 5.5, saturated with MgSO 4 . The samples were extracted with 1.5 ml of ethyl acetate, and 1 ml of the organic phase was checked for radioactivity. Samples lacking ribosomes were used as blanks and subtracted.
Binding of N-[acetyl-3 H]tRNA to the P site. The assay was performed using 10 pmol of 80 S ribosomes, 0.3 mg/ml polyuridylic acid, and 20 pmol of N-[acetyl-3 H]tRNA (2500 cpm/pmol) in the conditions described previously by Triana et al. (38). Blanks without ribosomes were used for background estimation. After incubation for 20 min at 30°C the samples were filtered using nitrocellulose membranes.

Complementation of a Conditional P0 Null Phenotype by
Heterologous Proteins-S. cerevisiae W303dGP0 carries the essential RPP0 gene under the control of the GAL1 promoter and it is viable only when grown in galactose as a carbon source (20). This strain has been transformed with the plasmids pFL37-P0Af, pFL37-P0Dd, pFL37-P0Rn, and pFL37-P0Li which, respectively, carry the coding sequence of the RPP0 genes from A. fumigatus, D. discoideum, R. norvegicus, and L. infantum under the control of the 5Ј regulatory region of the yeast RPP0 gene to assure the same level of expression for all of them. As a positive control, the strain was also transformed with the plasmid pFL37-P0Sc, carrying the S. cerevisiae RPP0 gene. The capacity of the transformed strains to grow in glucose media will define the ability of the heterologous P0 proteins to functionally substitute for the endogenous polypeptide.
Due to the existence of a large pool of active ribosomes carrying the yeast P0 protein, the transformants growing in galactose are able to grow in glucose for some time, and usually they have to be transferred to a second glucose plate to clearly detect the effect of the heterologous protein expression. In addition to the strain transformed with the homologous RPP0 gene, cells transformed with pFL37-P0Af, pFL37-P0Dd, and pFL37-P0Rn were able to continuously grow at 30°C on glucose plates. On the contrary, plasmid pFL37-P0Li did not support the growth of the transformed strain in glucose (Fig. 1A).
To test the complementation capacity of the heterologous P0 proteins in the absence of the P1/P2 proteins in the stalk, the S. cerevisiae D67dGP0 strain was used. This strain, like W303dGP0, carries the genomic RPP0 gene under the control of the GAL1 promoter but, as commented in the Introduction, its ribosomes are deprived of acidic P1/P2 proteins (18,13). None of the S. cerevisiae D67dGP0 transformed strains, except the control expressing the homologous P0, was able to grow in glucose (Fig. 1B). It seems, therefore, that the complementing effect of the P0 proteins from A. fumigatus, D. discoideum, and R. norvegicus requires the presence of the 12-kDa P1/P2 proteins in the ribosome.
Effect on the Ribosomal Stalk Composition-To test alterations in the stalk composition by the presence of the rat, A. fumigatus, and D. discoideum proteins, washed ribosomes from transformed S. cerevisiae W303dGP0 strains growing in glucose for more than 20 generations were fractionated by SDS-PAGE electrophoresis and the P0 proteins detected by Western using monoclonal antibody 3BH5, specific to the carboxyl end of the yeast P0 ( Fig. 2A). The intensity of the yeast and rat P0 protein was similar. The ribosomes carrying the A. fumigatus yielded the same results (not shown).
The D. discoideum P0 protein was not recognized by the monoclonal 3BH5 antibody since no band is detected in the corresponding sample (Fig. 2A). These results confirm differences in the sequence of the COOH-terminal peptide of D. discoideum P proteins (32), and, in addition, clearly indicate that the endogenous yeast P0 protein from the galactose culture is not present in the glucose-grown cells. The heterologous  (S.c.), A. fumigatus (A.f.), D. discoideum (D.d.), L. infantum (L.i.), and R. norvegicus (R.n.) were plated in medium YNB supplemented with the corresponding carbon source as indicated and incubated at 30°C.
P0 was detected in this case using a rabbit antiserum raised against the slime mold protein (32). The rabbit antiserum, which does not cross-react with the yeast and rat proteins, shows that the D. discoideum P0 is present in the corresponding ribosomes in a similar proportion than in the control D. discoideum ribosomes (Fig. 2B).
The results showing similar amounts of P0 in all the ribosomes suggest that the expression level of the heterologous proteins is similar in all transformed strains. This is not surprising since the heterologous genes are under the control of the yeast P0 5Ј-UTR in all the plasmids. However, to exclude the formation of a large cytoplasmic pool of heterologous proteins due to overexpression or poor incorporation into the ribosome, a Western analysis of the cell supernatant S100 extracts was performed using the monoclonal antibody 3BH5. The results confirmed the absence of free P0 in the cytoplasm of either the wild-type cells or the transformed yeasts, as shown in Fig.  3 for cell expressing the R. norvegicus protein.
The 12-kDa acidic proteins in washed ribosomes from the transformed strains were analyzed by isoelectrofocusing, and a densitometric estimation of the acidic protein bands in each sample was performed ( Table I). The results showed that while the control ribosomes from cells expressing the yeast P0 protein have standard amounts of the four proteins, P1␣, P1␤, P2␣, and P2␤, in the particles from cells expressing the rat and A. fumigatus P0, only proteins YP1␣ and YP2␤ are detected, and the amount is smaller than in the control. The ribosomes containing the D. discoideum protein show almost standard amounts of YP1␣ and YP2␤ but YP1␤ and P2␣ are notably reduced.
Expression of L. infantum Protein P0 -The protozoan P0 did not complement the absence of the yeast P0, but if it is expressed and binds to the yeast ribosome, it should be detected in the particles at short times after shifting the cells to a glucose medium. Unfortunately, the COOH-terminal end of the L. infantum P0, like the one from D. discoideum, is not recognized by the yeast monoclonal antibody, and there was no specific antibody available. However, the protozoan protein is larger than the yeast polypeptide and a new band at the expected position was detected in the gel from ribosomes of transformed cells growing in a glucose medium (Fig. 4).
The proteins in the band were transferred to a polyvinylidene difluoride membrane, treated with trypsin, and the peptides resolved by high performance liquid chromatography. Edman degradation sequencing confirmed the presence of specific L. infantum P0 peptides in the band.
Effect of Heterologous P0 Proteins on Cell Growth-The S. cerevisiae W303dGP0 strains containing the D. discoideum, A. fumigatus, or R. norvegicus genes, kept permanently in glucose plates, showed a doubling time of 125, 165, and 190 min, respectively, when growing in liquid YEP glucose medium at 30°C. The cells expressing the homologous RPP0 gene duplicated every time of 95 min in the same conditions. On the other hand, the growth rate of the D67dGP0 transformants growing in YEP-galactose liquid medium declined rapidly upon transferring to glucose medium, and they stopped growing completely after a few generations.
The capacity of the heterologous P0 proteins to support growth in stress conditions was tested by growing W303dGP0 transformants in high salt at 37°C. The strain expressing the RPP0 gene from R. norvegicus did not grow in the presence of either NaCl or sorbitol at 30 and 37°C. The cells expressing the A. fumigatus protein grew at high ionic conditions only at 30°C but not at 37°C (results not shown). Similar phenotypes have been related to other stalk protein mutations (36,39), suggesting that the translation of proteins involved in the cellular stress response is affected by changes in the ribosomal stalk.
In Vitro Activity of Ribosomes Carrying Heterologous Proteins-Ribosomes from the different S. cerevisiae W303dGP0 strains were tested in different in vitro assays trying to char-FIG. 2. Immunodetection of P0 protein in ribosomes from transformed S. cerevisiae W303dGP0. The same amount of ribosomes (80 g) from cells transformed with plasmids expressing P0 from S. cerevisiae (1), D. discoideum (2), and R. norvegicus (3) were resolved on SDS-polyacrylamide electrophoresis gels. Proteins were detected using either a monoclonal antibody specific to the yeast P proteins (A) or a rabbit serum raised against the D. discoideum P proteins (B). In B, the same amount of ribosomes from D. discoideum (4) was also included as a control.
FIG. 3. Immunodetection of P0 in the cytoplasm of cells expressing the R. norvegicus protein. Supernatant fraction (100 g) after removal of ribosomes (S-100 fraction) from strains expressing either S. cerevisiae P0 (1) or R. norvegicus P0 (3) were resolved by SDS-PAGE and the protein detected with monoclonal antibody 3BH5 to the COOH-terminal peptide. Ribosomes (20 g) from wild-type yeast were included as control (2).

TABLE I
Estimation of stalk proteins in ribosomes containing an heterologous P0 protein Estimation was carried out by scanning of either a SDS electrophoresis Western blot for protein P0 or silver-stained isoelectrofocusing gels for the 12-kDa proteins ("Materials and Methods"). The intensity of the bands in the ribosomes from the parental S. cerevisiae W303 strain was considered as 100. Average values from three determinations, rounded up to the closest significant figure, are shown. acterize the effect caused by the presence of the heterologous proteins. The overall polymerizing activity was tested in a poly(U)-dependent polyphenylalanine synthesis assay, the peptide bond formation capacity by a modified fragment reaction, and the interaction with substrates in a non-enzymatic Nacetyl-Phe-tRNA binding. An extract depleted of free 12-kDa proteins from S. cerevisiae D4567 (22) was used in the polymerization test to exclude a possible affect of these proteins on the ribosome activity. The results are summarized in Table II. The polymerizing activity is affected in all the ribosomes but especially in those carrying A. fumigatus and R. norvegicus P0 proteins; however, the capacity of the particles to form peptide bonds is practically unaffected in all cases. The binding of N-acetyl-Phe-tRNA is partially reduced also in A. fumigatus and R. norvegicus ribosomes but in less proportion than is their overall polymerizing capacity. These results are compatible with an effect of the heterologous P0 protein in the translocation step in agreement with the role that the stalk is proposed to have in the translation process.
Activity of Chimeric P0 Proteins Contains a Mammalian Amino-terminal Domain-To test whether the heterologous P0 proteins do really bind to the yeast ribosome but their complementing activity is actually determined by the carboxyl domain, chimeric proteins were constructed in which fragments of different size from the yeast P0 amino-terminal domain were replaced by the equivalent fragments from a mammalian protein. The H. sapiens RPP0 gene has been used for these constructions. The human P0 protein is 99% identical to the rat protein and both behave similarly when expressed in S. cerevisiae (data not shown). Two constructs, P0HYY and P0HHY, were obtained which encode chimeric P0 proteins of 314 amino acids containing, respectively, the first 138 and 205 amino acids from the human protein amino domain complemented with the corresponding yeast carboxyl domain. The fragment containing roughly the last 100 amino acids residues of the P0 carboxyl end has been shown to play a key role in the interaction of the P1 and P2 proteins (13).
S. cerevisiae W303dGP0 and D67dGP0 were transformed with both constructs and their capacity to support growth was tested. As summarized in Table III, both chimeric proteins are able to complement in glucose media the absence of native P0 in W303dGP0 almost as efficiently as the wild-type yeast P0 protein, even in the case of P0HHY that contains about twothirds of the human polypeptide. Similarly, both chimeric proteins are able to support growth of D67dGP0 on glucose plates although to a somewhat lower rate than the wild-type yeast protein. Moreover, the strains transformed with the chimeric proteins did not show the osmosensitive phenotype displayed by the strains expressing the H. sapiens proteins.
When the ribosomes from both glucose-grown W303dGP0 transformed strains were analyzed by isoelectrofocusing it was found that, in contrast to those from cells expressing the complete H. sapiens P0, they contained standard amounts of the P1/P2 proteins (Fig. 5). The human polypeptide, like the rat protein previously shown, allows the binding of a reduced fraction of the yeast P1␣ and P2␤ proteins.

DISCUSSION
Three functional domains can be defined in the eukaryotic ribosomal stalk P0 protein, one involved in binding to rRNA, a second one connected to P1/P2 protein interaction, and a third associated with the elongation factors. Useful information for the characterization of these domains has been obtained by studying the capacity of P0 from H sapiens, R. norvegicus, L. infantum, A. fumigatus, and D. discoideum to complement the absence of the endogenous protein in S. cerevisiae P0 conditional null mutants. Thus, the A. fumigatus, D. discoideum, and R. norvegicus genes allowed growth of S. cerevisiae W303dGP0 in conditions that repress the expression of the genomic yeast P0, although at a slower growth rate than the control, especially in the case of the rat protein. On the contrary, the gene from L. infantum did not support growth of the transformed strain in any condition.
The heterologous P0 proteins are present in the ribosomes from transformed S. cerevisiae W303dGP0 strains in similar amounts as the yeast P0 in the control particles. In contrast, the amount of 12-kDa P1 and P2 proteins is reduced and their relative proportion is altered. Proteins P1␤ and P2␣ are missing from the ribosomes carrying the A. fumigatus and rat P0 proteins and are notably reduced in ribosomes containing D. discoideum P0 while the binding of P1␣ and P2␤ is much less affected by the heterologous P0 proteins. These results confirm previous data indicating a different function for the two proteins of the same type (18,22,40), and suggest that the four proteins associate forming two heterodimers, P1␣/P2␤ and P1␤/P2␣, with different structural and functional roles. This idea is also supported by the fact that only P1␣/P2␤, but not P1␤/P2␣, are required for yeast P0-linked resistance to antifungal sordarin derivatives (41). Moreover, the P1␣/P2␤ pair is preferentially released from the ribosomes lacking protein L12 (42).   The in vitro polymerizing capacity of the ribosomes containing complementing heterologous P0 proteins is reduced, and there is a relationship between the ribosome activity and the amount of bound yeast 12-kDa acidic proteins. It seems, therefore, that the capacity of the heterologous P0 proteins to support cell growth depends on their potential to bind the yeast acidic proteins in the ribosomal stalk. This conclusion was totally confirmed when the heterologous genes were tested in the S. cerevisiae D67dGP0 mutant. In this strain, whose ribosomes are totally deprived of 12-kDa proteins, none of the heterologous proteins was able to support cell growth in glucose. Moreover, the substitution of the carboxyl-terminal domain in the mammalian P0 protein by its yeast equivalent fully activate the capacity of the chimera to complement the growth of S. cerevisiae W303dGP0 at the restrictive conditions. Similarly, the chimeric P0 allows the binding of standard amounts of the 12-kDa P1/P2 proteins to the stalk, confirming the importance of the last 110 amino acids for the formation of the ribosomal stalk, as previously reported (13).
Altogether, the reported results indicate that while the rRNA-binding domain of eukaryotic P0 proteins has been functionally conserved, the regions involved in protein-protein interactions have faster diverged. The RNA-binding domain is included in the NH 2 -terminal region comprising about 200 amino acids (13). A comparison of the P0 protein sequences indicates the existence of three amino acid clusters in this region, showing a conservation ranging from 50 to 75%, which might be involved in the interaction of these proteins with the rRNA GTPase region (Fig. 6). The highly conserved RNA binding capacity of P0 is not surprising since, as commented previously, the corresponding nucleic acid moiety of the GTPase center is also among the most conserved regions in the rRNA (43). In contrast, the region from approximately position 200 to position 275, involved in the interaction with the acidic P1/P2 proteins, has notably evolved as indicated by the different capacity of the heterologous P0 to bind the yeast 12-kDa proteins.
A third functional domain in protein P0 is defined by its interaction with the elongation factors during translation. P0 shares a highly conserved COOH-terminal peptide with the 12-kDa P1/P2, linked to the rest of the molecule through an alanine-rich hinge (19). This domain is essential for the protein activity, and it is proposed to have a relevant role in the interaction with the supernatant factors (13). Nevertheless, the presence of this common structural element is not sufficient for the heterologous P0 protein complementarity. Thus, although the carboxyl end is almost identical in yeast, A. fumigatus, R. norvegicus, and H. sapiens P0, none of them complement the absence of the yeast protein in D67dGP0 when the 12-kDa proteins are not present in the ribosome. It seems that other structural elements in P0, besides the conserved COOH-terminal peptide, play a role in determining a functional interaction of the elongation factor with the ribosome. These elements seem to be, at least in part, located in the amino-terminal domain since the chimeric protein P0HYY, carrying the first 138 amino acids from the H. sapiens sequence, only partially complement the absence of the native yeast P0 in the S. cerevisiae D67dGP0.