Ribosomal Protein L1 from Escherichia coli ITS ROLE IN THE BINDING OF tRNA TO THE RIBOSOME AND IN ELONGATION FACTOR G-DEPENDENT GTP HYDROLYSIS*

Two Escherichia coli mutants lacking ribosomal pro- tein Ll, previously shown to display 40 to 60% reduced capacity for in vitro protein synthesis (Subramanian, A. R., and Dabbs, E. R. (1980) Eur. J. Biochem. 112, 425-430), have been used to study partial reactions of protein biosynthesis. Both the binding of N-acetyl- Phe-tRNA to ribosomes and the 6 to 8-fold stimulation of the elongation factor G (EF-G)-dependent GTPase reaction by mRNA plus tRNA, assayed in the presence of wild type 30 S subunits, were low with L1-defficient 50 S subunits. Addition of pure protein L1 to the assay restored both reactions to 100% of the control. By contrast, the basic EF-G GTPase reaction in the absence of mRNA and tRNA was not at all affected (mRNA alone had no effect). None of the following partial reactions were more than moderately modified by the lack of protein L1: binding to ribosomes of EF-G .GDP plus fusidic acid; the translocation reaction catalyzed by EF-G plus GTP; poly(U)-dependent binding to ribosomes of Phe-tRNAPh‘ (whether dependent on elongation factor Tu plus GTP or not); and the EF-Tu-dependent GTPase activity. It is concluded that protein L1 is involved in the interaction between ribosomes and peptidyl-tRNA (or tRNA) in the peptidyl site and consequently in the ribosomal GTPase activity depending on the simulta- neous action of tRNA and EF-G.

RD19 and MV17-10, respectively, have been described (9,lO). Subramanian and Dabbs (1 1) analyzed ribosomes from these strains in vitro for protein synthesis using poly(U) and phage f2 RNA as messengers and found them to be 40 to 60% as active as wild type ribosomes; full activity could be recovered by adding pure protein L1. Here we report on the analysis of partial reactions of polypeptide synthesis using these L1deficient 50 S subunits.

MATERIALS AND METHODS'
Assay.y.-Unless otherwise stated, assays were carried out a t 30 "C in 50-pl reaction mixtures containing 50 mM imidazolium acetate (pH 7.5), 100 mM KCI, 20 pmol of control 30 S subunits (heatactivated hy a 30-min incubation at 40 "C), and 3 to 20 pmol of 50 S subunits. In most of the assays shown in the figures, equimolar amounts of 30 and 50 S subunits were used; however, all assays were also performed using limiting amounts of 50 S subunits and, where possible, conditions of linear kinetics. In some cases (e.g. enzymatic binding of Phe-tRNA to poly(U)-ribosomes), the mutant 50 S subunits had even higher activities than the controls under these conditions, while in most assays there was no difference in the results reported.
Translocation of N-A~etyl-r~ClPhe-tRNA from the Ribosomal A Incubation was for 10 min. For the binding of N-acetyl-Phe-tRNA to the A site (at 0 "C), the MgCI2 concentration was raised to 20 mM and 16 pmol of N-a~etyl-['~C]Phe-tRNA~~' were added. For translocation, immediately following the second step, the indicated amounts of EF-G plus 2.5 to 20 nmol of G T P were added and the reaction mixtures were incubated for 10 min at 30 "C, then cooled to 0 "C. For puromycin reaction, puromycin was added to 1 mM final concentration followed by a 30-min incubation a t 30 "C. following vigorous shaking for 30 s, an aliquot of the organic supernatant was counted in toluene/Triton X-100 (Serva) containing 10 g/liter of2,5-diphenyloxazole and 0.1 g/liter of 1,4-bis[2-(5-phenyloxazolyl)]benzene.
The EF-G GTPase reaction was carried out for 10 min in 100-pl reaction mixtures (160 mM KCI, MgC12 as indicated) with 100 pmol of EF-G and 20 to 3 0 nmol of [y-:"P]GTP (5 to 10 cpm/pmol). 10 pmol of tRNA"h' or N-acetyl-Phe-tRNA'he and 6 pg of poly(U) were added as indicated. G T P hydrolysis was measured as the liberation of 'T', as described (20).
The EF-Tu-dependent GTPase reaction was carried out at 20 mM KC1 (the optimum concentration) or 100 mM KCI, in 100-p1 reaction mixtures with Phe-tRNA (variable) and 6 pg of poly(U), o r Lys-tRNA and 6 pg of poly (A), 100 pmol of EF-Tu, 20 pmol of EF-Ts (this EF-Tu/EF-Ts ratio being optimal in our conditions), and 80 to :300 pmol of [r-'"P]GTP (1000 to 2500 cpm/pmol). The samples were processed as for the EF-G GTPase reaction. In both kinds of GTPase reaction, the highest activities never exceeded 205 hydrolysis of the total suhstrate.

RESULTS
The two-dimensional gel electrophoreses (22) of the proteins contained in 50 S subunits from E. coli wild type and mutant strains are shown in Fig. 1. Only the wild type contained the full complement of proteins, while 50 S subunits from both mutant strains lacked protein L1.
The individual reactions of polypeptide synthesis have been studied in the order in which they take place on the ribosome, starting with the binding of aminoacyl-tRNA to ribosomes.
Binding of Aminoacyl-tRNA to Ribosomes- Fig. 2 shows the poly(U)-directed binding of Phe-tRNA to ribosomes as a function of [Mg'"]. The EF-Tu-and GTP-dependent or "enzymatic" binding (upper curves) was nearly the same for all o f t h e strains. In the 3 to 15 mM range of [Mg'"], nonenzymatic binding was (reproducibly) slightly higher with the mutants than with the control 50 S subunits, but this effect disappeared with short incubation times (1 min) and was reversed when Lys-tRNA plus poly (A) replaced Phe-tRNA and poly(U) (data not shown). Moreover, with rate-limiting amounts of 50 S subunits and 10-s incubations a t 0 "C, the mutant 50 S subunits showed slightly higher stimulation of the enzymatic binding and lower nonenzymatic binding of Phe-tRNA than the controls (not shown). Therefore, binding of aminoacyl-tRNA to the A site (whether EF-Tu-and GTPdependent or not; Refs. [23][24][25] is not influenced by the lack of protein L1. N-Acetyl-Phe-tRNA is preferentially bound to the P site (Ref. 26; checked with the puromycin reaction). This reaction was studied at 7 and 15 mM M$+ by adding increasing amounts of the various 50 S subunits to control 30 S subunits (Fig. 3). The binding was very fast, and the kinetics could not be measured by our technique. A 30-s incubation was nonetheless chosen, at which point binding tapers off but still increases with time.
Lack of protein L1 strongly reduced N-acetyl-Phe-tRNA binding at 7 mM Mg'" (Fig. 3A) and almost totally abolished it at 15 mM Mi'+ (Fig. 3B). Adding back L1 restored this activity in both cases.
EF-Tu-dependent GTPme Activity-This reaction normally follows the enzymatic binding of aminoacyl-tRNA and precedes release of the factor. It normally requires both ribosomal subunits and aminoacyl-tRNA, whereas cognate mRNA is needed only at low (4 to 10 mM) Mi'' concentration (21). The activity was measured as a function of aminoacyl-tRNA concentration at 6 and 30 mM M$+ (see Miniprint); in no case was there a significant effect when 50 S subunits  lacking protein L1 replaced control subunits.
Binding of EF-G to Ribosomes-In the presence of the antibiotic fusidic acid, EF-G.GTP or EF-G-GDP bind to both 50 S subunits and 70 S ribosomes (27,28), which in turn bind to nitrocellulose filters. We studied this reaction using wild type and mutant 50 S subunits (see Miniprint) and found no significant difference. Therefore, the binding of EF-G to ribosomes does not require protein L1.
The Translocation Reaction and Peptidyltransferase-To study this reaction, the next in the sequence leading to the addition of an amino acid to the polypeptide chain, we made use of an artificial system involving a number of steps (see "Materials and Methods"); no dependence on the presence of protein L1 was found (see Miniprint).
The EF-G CTPase Reaction in the Absence of mRNA and  When present, a 2-fold molar excess of protein L1 over 50 S subunits was preincubated with the 50 S subunits for 10 min at 37 "C. 0, protein L1 without 50 S subunits.
tRNA-This reaction follows translocation and precedes release of the factor from the ribosome. Even though the natural partner of EF-G. GTP in this reaction is a ribosome carrying both an uncharged tRNA in the P site and a peptidyl-tRNA in the A site, this reaction proceeds readily with ribosomes alone. At low monovalent cation concentration, the 50 S subunit can support a turnover activity as high as that found with 70 S ribosomes at elevated, more physiological salt concentrations (29). In both of these assay systems, protein L1 had no significant influence (see Miniprint). The EF-G GTPase Reaction in the Presence of mRNA and tRNA-mRNA and tRNA have been reported both to stimulate (30,31) and to inhibit (32,33) the EF-G GTPase reaction, the latter most strongly at high [Mg2+]. As Parmeggiani et al. (34) and Chinali and Parmeggiani (35) have shown, the stimulation by tRNA of the EF-G GTPase is strongly enhanced at high concentrations of monovalent cations and low (5 to 10 mM) Mgz+ concentrations. We have confirmed this using isolated 30 and 50 S subunits. The conditions chosen for most of our experiments (160 mM K' , 7 mM Mg") maximize the stimulation of the EF-G GTPase reaction by tRNAPh' and poly(U). Fig. 4A shows the effect of poly(U) and tRNAPh" on the EF-G GTPase reaction at 160 mM K' . Their addition results both in a strong stimulation of the GTPase activity at moderate M$+ concentrations and a shift of the M$+ optimum toward lower concentrations. The addition of poly(U) alone showed neither of these effects (data not shown). Fig. 4B shows the same assay performed with 50 S subunits from mutant MV17-10 (similar results were obtained with RD19) supplemented with wild type 30 S subunits. The Mg2+ dependence of the basic GTPase reaction was similar to the control, while the stimulation by tRNAPhe and poly(U) was lower for the mutant 50 S subunits, particularly between 6 and 8 mM Mg2+. This pointed to a possible role of protein L1 in the interaction among ribosomes, tRNA, and EF-G. Fig. 5 shows the effect on G T P hydrolysis of adding purified protein L1 without and with poly(U) and tRNAPh'. In the absence of tRNAPh' and poly(U), protein L1 had no significant effect on GTPase activity with mutant and control 50 S subunits. The stimulation of the EF-G GTPase by tRNAPh' and poly(U) was fully dependent on protein L1 a t 7 mM Mg2+, while at 10 mM Mg2+ the importance of L1 for the tRNA effect was less pronounced, 50 S subunits from mutant MV17-10 displaying 50% of the full effect in the absence of L1. The protein did, however, restore activity to 100% of the control with both MV17-10 and RD19 50 S subunits.
The effect of increasing amounts of tRNAPh' at 7 mM M$+ is shown in Fig. 6. Maximum stimulation was achieved a t [tRNA]/[ribosomes] -0.5. At this low stoichiometry and in the ionic conditions employed, uncharged tRNA binds almost exclusively to the ribosomal P site (23)(24)(25)(26)36). N-Acetyl-Phe-tRNAPhe bound to the P site (checked by the puromycin reaction) also stimulated the EF-G GTPase reaction in the presence of protein L1.
The stimulation of the EF-G GTPase by tRNAPhe and

Ribosomal Protein
Ll in EF-G GTPase and tRNA Binding poly(U) has been found with several preparations of EF-G prepared by standard procedures (see Miniprint). Following a proposal by Leberman et al. (37), we have also made several preparations of EF-G using buffer systems containing sodium azide. However, even after prolonged dialysis against azidefree buffer, the resulting EF-G was irreversibly damaged in our hands. Its activity in poly(Phe) synthesis was reduced to less than 5%, the stimulation of the EF-G GTPase by tRNA at 160 mM K+ to approximately one tenth. This loss of activity is not apparent in "standard" EF-G GTPase assays not employing tRNA and mRNA or in binding of EF-G. GDP. fusidic acid to ribosomes, assays routinely used to detect EF-G during preparation. Incubation with e.g. 10 mM 2-mercaptoethanol or 2 mM dithiothreitol restored the stimulation of the EF-G GTPase reaction by tRNA to only 20 to 35% of that of control EF-G prepared by standard procedures.

DISCUSSION
The availability of E. coli mutants lacking one ribosomal protein a t a time greatly facilitates the task of assigning functions to these proteins. Reconstitution of activity then requires the incorporation of the missing protein only. For mutants MV17-10 and RD19, Dabbs et al. (38) have shown that protein L1 is missing both in the 50 S subunits and the supernatant; likewise, no evidence was found for a drastically altered protein L1. These mutants grow at approximately half the rate of wild type E. coli (11). Subramanian and Dabbs (11) have shown that polypeptide synthesis in uitro is slowed down accordingly, to about 40 to 60%, and that it can be fully restored by adding pure protein L1.
The present study suggests that two of the tested partial reactions of polypeptide synthesis are responsible for this effect: i.e. binding of peptidyl-tRNA (or tRNA) to the ribosomal P site and the coupled stimulation by tRNA and cognate mRNA of the EF-G GTPase reaction.
Protein L l has been localized on the 50 S subunit by immunoelectron microscopy using wild type, RD19, and MV17-10 50 S subunits (38). The binding site for EF-G has been similarly defined (39). As Fig. 7 (40) shows, L1 maps on the wide lateral protuberance opposite to the L7/L12 stalk (41). In the 70 S ribosome, the head region of the 30 S subunit, where most functional sites have been found, it positioned between the wide lateral proturberance and the central protuberance (1, 42). The binding site for EF-G as determined by immunoelectron microscopy (39) maps close to the origin of the L7/L12 stalk, far away from protein L1. On the other hand, Maassen and Moller (43), using a photochemical crosslinking reagent spanning about 10 A, have shown direct interaction between EF-G and L1.
From the present results, this protein appears to be functionally important for binding of tRNA to the ribosomal P site. Concerning the stimulation by tRNA of the EF-G Ribosomal Protein L1 in EF-G GTPase and tRNA Binding GTPase reaction, it seems possible that both tRNA and 20. Sander, G., Marsh, R. C., Voigt, J., and Parmeggiani, A. (1975) protein L1 act coordinately to anchor EF-G in a more favor-Biochemistry 14, 1805-1814 able position for GTP hydrolysis. Whether this effect involves 21. Sander, G. (1977) E m J . Biochem. 7 5 9 523-531 direct interaction between EF-G and ~1 , EF-G and tRNA, or 22. Kyriakopoulos, A., and Subramanian, A. R. (1977) Biochim. both remains to be elucidated.