The GTPase Center Protein L12 Is Required for Correct Ribosomal Stalk Assembly but Not for Saccharomyces cerevisiaeViability*

Protein L12, together with the P0/P1/P2 protein complex, forms the protein moiety of the GTPase domain in the eukaryotic ribosome. In Saccharomyces cerevisiae protein L12 is encoded by a duplicated gene, rpL12A andrpL12B. Inactivation of both copies has been performed and confirmed by Southern and Western analyses. The resulting strains are viable but grow very slowly. Growth rate is recovered upon transformation with an intact copy of the L12 gene. Ribosomes from the disrupted strain lack protein L12 but are able to carry out translationin vitro at about one fourth of the control rate. The L12-deficient ribosomes have also a defective stalk containing standard amounts of the 12-kDa acidic proteins P1β and P2α, but proteins P1α and P2β are drastically reduced. Moreover, the affinity of P0 is reduced in the defective ribosomes. Footprinting of the 26 S rRNA GTPase domain indicates that protein L12 protects in different extent residues G1235, G1242, A1262, A1270, and A1272 from chemical modification. The results in this report indicate that protein L12 is not essential for cell viability but has a relevant role in the structure and stability of the eukaryotic ribosomal stalk.

The ribosomal region involved in the hydrolysis of the elongation factor-bound GTP molecule upon its interaction with the ribosome during translation is generally called the ribosomal GTPase domain. In bacteria, a number of elements, RNA and proteins, have been shown to participate in this process to a different extent (see Ref. 1 for a review). At least two RNA components have been identified, which include the ␣-sarcin loop in domain VI of the 23 S rRNA and a highly conserved double hairpin in domain II, frequently referred to as the RNA GTPase center (2,3). Two proteins, L10 and L11, which bind to partially overlapping sites in the conserved domain II region, are also structural components of this active domain. On the other hand, protein L10 forms a very stable association with two dimers of the acidic protein L7/L12, the 6 M urea-resistant pentameric complex L10-(L7/L12) 4 , which is the main component of a typical protuberance of the large ribosomal subunit called the ribosomal stalk. The stalk is directly involved in the interaction of the elongation factors, as has been clearly shown by electron microscopy (4), participating in the translocation mechanism. Protein L11, which binds to the rRNA in a coop-erative way with pentameric protein complex L10-(L7/L12) 4 (5), is important for the GTPase activity, and in determining the right conformation of the GTPase center (6). This protein also has a key role in the mechanism by which thiostrepton and similar antibiotics block the elongation factor functions (7,8). Protein L11 is not essential for cell viability, and L11-defective bacterial strains have been obtained that are also resistant to thiostrepton (7).
Despite the large amount of data accumulating on its structure and function, we are still far from understanding the molecular mechanism supporting the bacterial GTPase center function. Much less is obviously known in the case of the eukaryotic organisms. Prokaryotic and eukaryotic ribosomes carry out the same basic functions, and, consequently, data obtained about the former have usually been extrapolated to the latter. Although the individual components show substantial differences (P0 is larger than L10, and L7/L12 has evolved to two families of closely related proteins, P1 and P2), they play similar roles. In fact, some of the data available on the structure of the eukaryotic GTPase domain fit nicely with those from bacterial ribosomes. Thus, there is a pentameric complex, P0-(P1) 2 -(P2) 2 (9), and a L11-like protein (10), presently called L12 (11), both binding to the 26 S/28 S GTPase RNA domain at sites that are equivalent to those found in bacteria (12). Moreover, both components are required for the correct conformation of the rRNA as determined by in vitro binding studies (13).
Nevertheless, the eukaryotic organisms have developed new regulatory elements missing from less evolved systems. In this sense, the GTPase domain, and more specifically the ribosomal stalk, might be a paradigmatic example. In eukaryotes, the stalk also seems to be involved in the interaction and activity of the elongation factors, but, in addition, a number of interesting features not present in the bacterial ribosome strongly suggest its implication in a regulatory mechanism of the ribosomal activity (see Ref. 14 for a recent review).
In general, the data underline the lower stability of the eukaryotic stalk as compared with the equivalent bacterial structure, and this feature is in the base of the mechanism that eukaryotes seem to have developed to regulate the ribosome translational activity in certain conditions (15). The structural differences existent between bacterial L10 and eukaryotic P0 must be to a great extent responsible for the dissimilar stability of their respective ribosomal stalks. In fact, it has been found that the eukaryotic specific C-terminal extension plays an important role in the formation of the pentameric complex (16). Moreover, the evolution from a unique L7/L12 bacterial protein to a family of proteins, P1 and P2, apart from introducing potentially useful structural diversity into the system, seems to have affected the self-association properties of these eukaryotic proteins, which, in contrast to the bacterial equivalents, can be found as monomers in solution (17).
The two other components of the active domain, the rRNA and protein L12, although not directly forming part of the stalk, can obviously affect its stability. In the first case, however, the full in vivo functional interchangeability of the bacterial and yeast domain II GTPase RNA (18,19) excludes this component as being important in order to explain their functional differences. The second component, protein L12, has an important role in determining the in vitro RNA GTPase center structure (13). The Saccharomyces cerevisiae protein L12, formerly called L15, is encoded by an intron-less duplicated gene (20). In contrast to the other yeast GTPase domain components, protein L12 has been studied in less detail. As a way to carry out an in vivo analysis of its function, especially its role in the GTPase structure and function, the obtention of S. cerevisiae strains defective in protein L12 was undertaken. The results of this study are shown in this report.
Escherichia coli DH5␣ was used for handling cloning vectors and was grown in LB medium. Bacteria were transformed according to standard procedures (22).

Genetic Manipulations
S. cerevisiae mating, sporulation, and tetrad analysis were performed following published methods (21).

Recombinant DNA Techniques
Restriction endonucleases, T4 DNA ligase, Klenow DNA polymerase I fragment, and other enzymes were purchased from Boehringer Mannheim, New England Biolabs, or Amersham Pharmacia Biotech.
DNA preparation, restriction enzyme digestions, agarose gel electrophoresis, ligation of DNA fragments, Southern blots, etc., were carried out according to standard techniques (23). DNA was sequenced by the dideoxy chain termination method using universal primers and complementary oligonucleotides. Probes were labeled by the random initiation method using Klenow DNA polymerase fragment and [␣-32 P]dCTP.
PCR was performed in 100 -50-l reaction mixtures in the conditions optimal for the used polymerase. The PCR products were purified by electrophoresis in agarose gels.

Gene Disruption Strategies
Inactivation of the rpL12 genes was performed by transforming S. cerevisiae with appropriate disruption cassettes following the methods described by Philippsen and co-workers (24, 25). Transformation was carried out by the lithium acetate method as described previously (26).
The disruption cassettes contained either the KanMX4 or the Schizosaccharomyces pombe HIS3 genes as selection markers, flanked by DNA fragments homologous to regions next to the fragment that has to be deleted. Cassettes with either 40 -45-bp flanking fragments (SFH) or 400 -560-bp flanking fragments (LFH) were obtained by PCR using chimeric oligonucleotides (24, 25). Plasmids pUG7-L12A-A 1.4-kilobase pair fragment, including the coding region, 528 bp from the 5Ј flanking region and 367 bp from the 3Ј flanking region of the rpL12A gene, was obtained from genomic S. cerevisiae W303 DNA by PCR using Pfu DNA polymerase. The PCR fragment was subcloned by blunt end ligation into the EcoRV site of the pUG7 polylinker. The absence of PCR induced mutations was confirmed by DNA sequencing.
pYES2-L12-A 498-bp XhoI-BamHI PCR fragment containing the rpL12A gene coding region was inserted in the corresponding sites of plasmid pYES2 (Invitrogen) under the control of the GAL1 promoter.

DNA Blots (Southern)
Yeast DNA was prepared as described (21). DNA, after digestion with restriction enzymes, was resolved by electrophoresis in 0.8% agarose gels, and blotted to nylon membranes (Amersham Pharmacia Biotech). Hybridization was performed according to standard procedures (23).
The S-30 was centrifuged to obtain the ribosomes and supernatant fractions as described previously (30). As a source of supernatant factors for in vitro protein synthesis, the fraction precipitated between 20% and 50% saturation of ammonium sulfate, and called S-100, was used. The acidic P proteins (SP fraction) were extracted from the ribosomes by ammonium-ethanol treatment (30); the extracted fraction was dialyzed against 10 mM Hepes, pH 7.4, and 0.5 mM phenylmethylsulfonyl fluoride and concentrated by filtration through Centricon SR3 membranes (Amicon).

Electrophoretic Methods
Proteins were analyzed by either SDS-PAGE or by isoelectrofocusing. SDS-PAGE was performed according to standard procedures. Isoelectrofocusing was carried out on vertical 5% polyacrylamide, 8 M urea isoelectrofocusing gels in the 2.5-5.0 pH range as described previously (31).
Proteins were either detected by silver staining or blotted to either PVDF or nitrocellulose membranes by electrophoresis in a semidry system using Novablot LKB buffer. Proteins in membranes were immunodetected following standard procedures (32) and using specific antibodies. Antibodies to L11, L12, and P proteins have been previously described (10,33,34).

Disruption of the Genes rpL12A and rpL12B
Encoding Ribosomal Protein L12-Disruption of the gene rpL12A in S. cerevisiae W303 was readily achieved by transforming the cells with an SFH disruption cassette containing the KanMX4 marker and selecting for Geneticin (G418)-resistant colonies (24). For rpL12B, an LFH cassette with HIS3 flanked by more than 400 bp at each end was required. In both cases disruptants were checked by Southern blotting (data not shown).
Preliminary attempts to obtain a double disruptant strain failed suggesting that protein L12 might be essential for cell viability; therefore, the preparation of conditional null mutants was approached. A heterozygous diploid yeast strain, S. cerevisiae A6H1, having one of the rpL12A and rpl12B copies interrupted by the KAN and HIS3 markers, respectively, was transformed with plasmid pYES2-L12, which carries a L12 coding region under the GAL1 promoter. One transformed strain, S. cerevisiae A6H1O, was sporulated, and a number of tetrads were dissected in galactose. Three of them were analyzed in detail (Table I), and they segregated as tetratype with respect to the KAN and HIS3 markers.
To confirm the disruption of the genes, the four haploid strains, A6H1O-6E, A6H1O-6F, A6H1O-6G, and A6H1O-6H, derived from tetrad 6, were analyzed in detail. The genes were tested by PCR, and the results indicated that rpL12A was disrupted in strains A6H1O-6E and A6H1O-6G, and rpL12B in strains A6H1O-6E and A6H1O-6F (data not shown). Consequently, 6E has both genes disrupted, 6F and 6G are monodisruptants, and 6H has a wild-type genotype, in agreement with the segregation of the genetic markers.
Unexpectedly, the three putative double disruptant kan r , his ϩ spores grew very poorly in glucose medium. As our preliminary results had suggested that protein L12 was essential for the cell viability, the effect of removing the pYES2-L12 plasmid was tested to exclude the possibility that a low background expression of the GAL1-controlled L12 gene could allow cell survival. The A6H1O-6E strain was grown on SD plates containing 5-FOA at 30°C for 4 days. The colonies that grew in these conditions were checked for the presence of the disruptant KAN and HIS3 markers as well as for the absence of URA3, by growing them in appropriate media. In addition, the persistence of the disrupted genomic genes, and the absence of the pYES2-borne L12 gene was confirmed by PCR. The viability of these 5-FOA-resistant strains clearly indicates that protein L12 is, indeed, not an absolute requirement for vegetative growth, although the cells lacking the protein grow very slowly.
Effect of the Disruption of the L12 Genes on Cell Growth-The four haploid strains were grown in YEP-glucose liquid medium at 30°C to estimate the effect of the different disruptions in the cell growth rate. The estimated exponential doubling time in YEP medium was around 95, 153, 170, and 325 min for strains A6H1O-6H, A6H1O-6G, A6H1O-6F, and A6H1O-6E, respectively.
To confirm that the effect on growth rate is due to the absence of protein L12, one of the 5-FOA-derived strains, S. cerevisiae A6H1O-6EA1, was transformed with centromeric plasmid pFL36-L12 containing the rpL12A gene yielding the strain S. cerevisiae A1-L12. The transformation restored the growth in glucose of A1-L12 to rates similar to those of S. cerevisiae A6H1O-6F which contains the same gene copy.
To look for a possible temperature-sensitive phenotype, the disruptant strains were grown at 15°C, 30°C and 37°C on YEP-galactose and YEP-glucose plates, together with strain W303 as control (Fig. 1). As expected, all strains grew similarly in galactose at the different temperatures. In glucose, except A6H1O-6E and A6H1O-6F, whose growth was similar at the three temperatures, the other strains grew faster as the temperature increased. This was especially notable for strain A6H1O-6G, which at 15°C grew as slowly as A6H1O-6F but at 37°C grew almost as fast as A6H1O-6H. These results indicate that rpL12B in strain A6H1O-6G is quite sensitive to temperature changes whereas rpL12A in strain A6H1O-6F does not respond to these alterations.
Expression of Protein L12 in the Disruptant Strains-The protein L12 present in the cells of the different strains was  1. Effect of temperature on the growth of the different L12 disrupted strains. Cells from strains A6H1O-6E, A6H1O-6F, A6H1O-6G, A6H1O-6H, and the wild-type W303-1b were grown up to A 600 ϭ 1 at the indicated temperatures in YEP with either galactose or glucose as a carbon source, and the cultures were serially diluted by a factor of 10. Aliquots (5 l) of each dilution were applied to the plates containing the same media and incubated at the indicated temperature. estimated by Western. Total extracts from the four haploid strains and from the diploid wild-type parental strain were resolved by SDS and detected using monoclonal antibodies specific to protein L12 (Fig. 2A). The protein was not found in the extracts from A6H1O-6E confirming the inactivation of both gene copies in this strain. On the other hand, the amount of L12 is higher in the extracts of strain A6H1O-6H and about the same in A6H1O-6F and A6H1O-6G.
To confirm the expression of protein L12 from pFL36-L12, a Western analysis of cell extracts from A1-L12 resolved by SDS-PAGE was similarly performed. Protein L12, detected in the pFL36-L12 transformed cells and in the parental W303 strain, is missing in strain A6H1O-6EA1 (Fig. 2B).
Effect of Protein L12 Absence on the Ribosomal GTPase Center Structure-An analysis of the GTPase center proteins in the ribosomes from the different strains was performed by polyacrylamide gel electrophoresis. Using a monoclonal antibody specific to the conserved C-terminal peptide present in all the P proteins, the amount of these components in the ribosomes from the double disruptant S. cerevisiae A6H1O-6E and the parental strain was estimated. No appreciable differences were found in the amount of protein P0 in both samples; on the contrary, the total amount of 12-kDa proteins (in S. cerevisiae formed by four proteins, P1␣, P1␤, P2␣, and P2␤), which cannot be resolved by this technique, seemed to be reduced in the disruptant cells (see below).
Protein P0, unlike its bacterial equivalent, protein L10, binds very strongly to the RNA, and it is difficult to remove from the ribosome by washing with salt (35). To test whether this interaction is affected by protein L12, ribosomes from the double disruptant and the control cells were washed with increasing concentrations of ammonium chloride in the presence of 50% ethanol, and the released proteins were analyzed by Western blot. As shown in Fig. 3, whereas P0 started being released from the wild-type ribosomes only by a 1 M salt buffer, the protein was already detected in the 0.5 M washing from the L12-defective particles.
The effect of L12 in the binding of the 12-kDa proteins was checked by isoelectrofocusing the L12-defective ribosomes. The ribosomes from the double disruptant S. cerevisiae A6H1O-6E were found to contain less amount of acidic proteins. Although proteins P1␤ and P2␣ were only slightly reduced, proteins P1␣ and P2␤ were practically absent from the ribosomes (Fig. 4). Protein P1␤ can be found either in the unprocessed or in the processed form (P1␤Ј) depending on the preparation (36).
Transformation with plasmid pFL36-L12 increases the proportion of the missing acidic proteins but not to the wild-type level. This is in agreement with the partial recovery of the growth rate induced by the rpL12A gene copy in the plasmid.
Footprinting of the GTPase Center in the 26 S RNA-To test the effect of protein L12 removal on the structure of the RNA, the GTPase region in ribosomes from the double disrupted strain was subjected to chemical probing. Ribosomes were treated with dimethyl sulfate, CMCT, and kethoxal, and mod-ifications were detected by reverse transcriptase using oligonucleotides homologous to the 3Ј end of the GTPase center (28). As shown in Fig. 5, nucleotides G1235 and A1262 were totally protected, and nucleotides G1242, A1270, and A1272 partially protected from modification by the presence of protein L12 in the wild-type ribosomes.
Polymerizing Activity of L12-defective Ribosomes-Ribosomes from the double disrupted strain were prepared and tested in a poly(U)-dependent system using as a source of supernatant factors either its own S-100 fraction or a similar fraction from the parental S. cerevisiae W303. Ribosomes from the S. cerevisiae A6H1O-6EA1 transformed with pFL36-L12 were similarly tested. As shown in Fig. 6, the polymerizing activity of the defective ribosomes was reduced to 10 -20% of the control activity. Transformation with pFL36-L12 restored notably the activity of the particles but not to the control values, in agreement with the effect of the transformation in the growth rate and in the stalk composition (Fig. 4). DISCUSSION Protein L12 is encoded in S. cerevisiae by a duplicated gene, rpL12A and rpL12B, which express identical polypeptides (20). Sporulation and tetrad analysis of a heterozygous diploid strain carrying a plasmidic L12 gene under the control of the GAL1 promoter yielded haploid strains carrying either one or both L12 gene copies simultaneously inactivated and a galactose-inducible copy.
Expression of both L12 gene copies is required for optimal growth indicating that none of them provides enough ribosomal protein and suggesting the absence of a compensatory regulatory mechanism reported in other instances (37). It is interesting, however, that the absence of rpL12A can be almost completely compensated by increasing the growth temperature of the corresponding disrupted strain. This might indicate that either a real temperature-dependent regulatory process takes place or, simply, that the expression of rpL12B copy is temperature-sensitive. Further experimental work, now in progress, is required to clear up different pending questions related to the expression of the two yeast protein L12 genes.
Transfer of the cells to a glucose medium to repress the plasmidic L12 gene expression drastically reduced but did not abolish the growth of the double L12 disruptant. In fact, it was possible to obtain viable double disrupted strains in which the L12 carrying plasmid had been cured by growing in 5Ј-FOA. It seems, therefore, clear that protein L12 is not an absolute requirement for ribosome activity and cell viability. However, the protein must have an important role in ribosome function as cells lacking L12 have a doubling time close to 6 h.
In contrast to E. coli, in which a significant number of ribosomal proteins are dispensable for cell viability (38), most proteins seem to be essential in the yeast ribosome. Excluding the exchangeable acidic proteins P1/P2, out of 32 proteins so far studied, only three, S31 (UBI3) (39), L24 (formerly L30) (40), and L39 (UBI1/UBI2) (41), are shown to be dispensable for cell viability. These results indicate that eukaryotic ribosomes have tighter structural requirements than the bacterial one, and the changes resulting from the protein absence are lethal for the ribosome activity.
Protein L11 is also dispensable for viability in bacteria (7). There are, however, significant differences in the role played by bacterial L11 and eukaryotic L12, the most important being the different involvement in the stability of the stalk. Thus, although the bacterial mutants lacking protein L11 contain standard amounts of L7/12 (42), in the L12-defective yeast strains two of the four acidic proteins, P1␣ and P2␤, are missing from the ribosomes. These results indicate that protein L12 is either directly or indirectly involved in the interaction of the 12-kDa acidic proteins with the eukaryotic ribosome. Physical proximity of the C-terminal domain of bacterial L7/12 to protein L11 has been reported, confirming the high flexibility of the acidic proteins (43). However, these interactions are probably irrelevant to the binding of the acidic proteins to the ribosome, which takes place only through the N-terminal domain of the proteins in bacteria (44) as well as in yeast (45). In any case, although the acidic proteins have been cross-linked to protein P0 (46), a direct interaction between L12 and proteins P1/P2 has not been so far reported in the ribosome, although they have been shown to be able to associate in solution (33).
The preferential release of P1␣ and P2␤ from L12-defective ribosomes indicates the asymmetrical structure of the yeast stalks as compared with the bacterial one, and stresses the different structural role of each protein of the same type, P1␣/ P1␤ and P2␣/P2␤. Previously, the analysis of different disrupted mutants had shown that the effects caused by the absence of one of the proteins cannot be suppressed by an excess of the other one (47)(48)(49). All the functional and structural data confirm, therefore, that the two members of the same protein family are performing specific functions.
Independent of its physiological meaning, which is not obvious, the existence of four acidic proteins in yeast ribosomes together with the assumed dimeric character of these proteins raises some interesting structural questions. Thus, as only four copies of acidic proteins have been detected per eukaryotic ribosome (9,50), in the cell there must be either a homogeneous population of ribosomes carrying one monomer of each protein type or a heterogeneous population of particles carrying two dimers in different combinations, (P1␣) 2 /(P2␣) 2 , (P1␣) 2 /(P2␤) 2 , (P1␤) 2 /(P2␣) 2, (P1␤) 2 /(P2␤) 2. The specific release of P1␣ and P2␤ by the absence of L12 fits in better with an asymmetrical ribosomal stalk made by one protein of each type, P0, P1␣, P2␤, P1␤, P2␣, in which the four 12-kDa proteins do not have the same role.
Contrary to what could be expected, there seems to be a closer relationship between two proteins of a different type, P1␣/P2␤ and P1␤/P2␣, the binding of the first pair being more directly affected by the presence of protein L12. These two pairs might play the same function than the standard P1 and P2 dimers in mammals. The P1␣/P2␤ and P1␤/P2␣ associations are, however, not an absolute requirement for the formation of the stalk, as it is possible to obtain, by gene disruption, yeast mutant strains with ribosomes exclusively containing the P1␣/ P2␣ and P1␤/P2␤ pairs (48,49). In any case, the P1␣/P2␤ pair seem to be especially relevant for ribosome activity as its absence is more harmful for cell growth than the absence of P1␤/P2␣ (49).
Protein L12 also affects the interaction of protein P0, although not so obviously as in the case of the acidic proteins. P0 is present in similar amounts in disruptant and in wild-type ribosomes, but the protein can be released from the particles more easily in the first case. Thus, although P0, unlike bacterial L10, is hardly washed off the eukaryotic ribosomes by ammonium/ethanol buffers (30,35), the protein can be detected in the mutant ribosome washes at relatively low ammonium chloride concentrations.
The effect of protein L12 in the protein P0 interaction is in agreement with the binding of both proteins to partially overlapping sites in the mammalian rRNA (51). Our footprinting results with ribosomes, in addition to confirm the absence of L12 from the double disrupted strain ribosome, clearly show that this protein protects nucleotides in the 26 S rRNA domain II GTPase center. These nucleotides are equivalent to those protected by mammalian protein L12 in the corresponding 28 S rRNA (12,51) and by bacterial L11 in the 23 S rRNA (3). On the other hand, the coincidence of our data, obtained with native ribosome, and those from in vitro reconstituted systems indicates that the structure of the GTPase active domain is probably independent from the rest of the particle and supports the idea that the ribosome is made of a number of structural domains that show a fairly large degree of structural autonomy. Thus, it was recently shown that it is possible to assemble the 30 S subunit platform domain specifically and independently from the rest of the subunit (52).
Altogether, the results in this report confirm previous data indicating that despite apparently performing similar basic functions during translation, the bacterial and the eukaryotic stalk show notable structural differences. One of the more significant divergences is the higher heterogeneity and asymmetry of the yeast structure, which results in lower stability. This characteristic is, on the other hand, an essential requirement for the regulatory role that it may be playing. Obviously, a highly stable structure like the bacterial stalk can hardly be involved in a dynamic exchange of its components, like the one that takes place in eukaryotes.