ppgpp inhibition of elongation factors tu, g and ts during polypeptide synthesis

Mol Gen Genet (1984) 197:36-45

© Springer-Verlag 1984

ppGpp inhibition of elongation factors Tu, G and Ts during polypeptide synthesis Aria-Maria Rojas, M~ns Ehrenberg, Siv G.E. Andersson, and C.G. Kurland Department of Molecular Biology, The Biomedical Center, Box 590, S-751 24 Uppsala, Sweden

Summary. The inhibition of elongation factors G, Tu and Ts by ppGpp was studied in vitro in a translation system with missense frequency and elongation rate similar to those in vivo. ppGpp inhibits EF-G with K I = 6 x 10- 5 M. When ppGpp is in twofold excess over GTP and EF-G is the rate-limiting component, the elongation rate is reduced twofold by ppGpp. EF-Tu is inhibited with K~=7 x 1 0 - 7 M in the absence of EF-Ts. When EF-Ts is added, the binding of ppGpp to EF-Tu becomes successively weaker. 1/K l de- pends linearly on 1/[Ts] and the intercept at the abscissa gives K~ = 4 x 10- 5 M. This reflects the binding of ppGpp to the binary TuTs complex. The slope reveals that the binding of EF-Ts to the TuMS binary complex is strong (10 -6 M), ppGpp may thus inhibit the cycling of EF-Tu indirectly by the removal of the free EF-Ts by its adsorption to TuMS, as well as directly by simple binding to Tu. EF-Tu inhibition by ppGpp can be fully reversed by high levels of aminoacyl-tRNA only in the presence of EF-Ts and at low ribosomal activity. Our in vitro observations have been extrapolated to in vivo conditions with conclusions as follows: Under strong amino acid starvation ppGpp in twofold excess over GTP cannot reduce significantly the elongation rate of ribosomes and thereby restore the errors to their normal levels as in the stringent response. Under weak starvation, in contrast, a significant rate reduction can be achieved by the trapping of EF-Ts in complex with TuppGpp.

Introduction

Among the pleiotropic effects of relA mutations in Escherichia coli is a marked enhancement of translational error during amino acid limitations (Gallant 1979; O'Far- rell 1978). Since the relA gene determines the structure of the stringent factor, which catalyzes in turn the synthesis of the guanine nucleotide analogs called magic spot I (ppGpp) and magic spot II (pppGpp), it has been concluded that these are involved normally in restricting translation errors during amino acid limitation (Gallant and Foley 1979). We are concerned here with how ppGpp influences translational accuracy, and in particular, how it interacts with the elongation factors.

The interactions between the elongation factors (EF-Tu and EF-G) and the magic spots have been the subject of numerous studies (Arai et al. 1972a, b; Hamel and Cashel

Offprint requests to." M. Ehrenberg

1973; Hamel 1976; Legault et al. 1972; Miller et al. 1973). It is generally agreed that both factors can be inhibited by magic spots, and that the inhibition of one or the other might be involved in the restriction of the errors of translation (O'Farrell 1978). For example, if the inhibition of other components of the protein synthetic machinery is sufficient, the charging level of the tRNA species corresponding to the limiting amino acid will increase because it is no longer the rateqimiting component in the system. The redress of the charging level would then restore the error frequency to its normal level. Unfortunately, careful measurements of the charging levels of the tRNA's during amino acid limitation in vivo suggest that they are not influenced significantly during the accumulation of magic spots (B6ck et al. 1966; Piepersberg et al. 1979; Yegian and Stent 1969). The problem then is to explain how the presence of magic spots can influence the accuracy of translation without perturbing the charging levels of the relevant tRNA species during amino acid limitation.

To address this problem we have developed a steady state in vitro system that, when limited for a cognate amino acid, responds with an increased error frequency in the absence of ppGpp but not in its presence (Wagner and Kur- land 1980; Wagner et al. 1982a). Furthermore, there is a concentration range of ppGpp under which the restriction of missense errors occurs without a concomitant increase of the charging level of the limiting tRNA species. The data obtained with this system were rationalized in the following way (Wagner et al. 1982a). First, it was suggested that the basis of this effect was the interaction of EF-Tu with ppGpp, and not the interaction with EF-G. Second, a kinetic argument was forwarded to account for the nonre- sponse of the charging level of the limiting tRNA species to magic spot. Since then, several new elements have entered our view of the effect of the stringent response on the accuracy of translation.

First, we have discovered that the kinetic mechanism we suggested previously cannot be correct because it contra- dicts the requirement for the charging level of the limiting tRNA in the steady state to be related reciprocally to the error frequency at the corresponding codon. The only way our suggestion could be valid is for special cases in which, for example, the affinity of one aminoacyl-tRNA for EF-Tu is very different from that of its competitor. The lack of generality disqualifies our earlier model, but there are alter- natives that also involve the interaction of EF-Tu with ppGpp.

Second, if magic spots were to change the accuracy of the ribosomal selection of tRNA's, the fact that the charg-

ing levels are not altered by ppGpp would be explained. Indeed, we have shown elsewhere that it is possible in prin- ciple to increase the accuracy of the proofreading flows with the aid of magic spot (Kurland and Ehrenberg 1984). According to this interpretation, EF-TuppGpp can increase the selectivity of ribosomal proofreading flows by retarding the rate at which peptide bonds are formed from the aminoacyl-tRNA and, thereby, permitting more time for the dis- carding of mismatched tRNA species.

In the present experiments we have studied the kinetic interactions of ppGpp with EF-G and EF-Tu during polypeptide synthesis. We show that the interaction with EF-G is considerably weaker than with EF-Tu. In addition, we have studied the influence of aminoacyl-tRNA as well as EF-Ts on the competition between GTP and MSI for EF- Tu. The data lead to an important negative conclusion: when starvation in vivo for a given amino acid is nearly complete, the stringent response cannot influence translational accuracy by the inhibition of either EF-Tu or EF-G activity. Nevertheless, under these conditions, EF-TuMSI will accumulate in significant amounts, and we believe that these are adequate to influence the accuracy by modulating the proofreading flows.

Materials and methods

Chemicals

Phenylalanine, phosphoenolpyruvate (PEP), GTP, putrescine, spermidine, myokinase (E.C. 2.7.4.3), pyruvate kinase (E.C. 2.7.3.40) and nucleoside 5'-diphosphate kinase (E.C. 2.7.4.6) were purchased from Sigma, Saint Louis, Missouri, USA. Poly(U) and E. coli tRNA from strain MRE600 were obtained from Boehringer-Mannheim, FRG. ppGpp from Sigma and Calbiochem-Behring Corp, was used.

Buffer

Polymix buffer (Jelenc 1980; Jelenc and Kurland 1979) contained 5 mM Mg 2+, 0.5 mM Ca 2+, 8 mM putrescine, 1 mM spermidine, 5raM phosphate, 5raM NH~-, 95 mM K +, I mM dithioerythritol, pH 7.5. Working strength buffer was prepared by mixing, 100 mM potassium phosphate (pH 7.5) and a tenfold concentrate without phosphate (pH 7.5) prepared according to the procedure described in Jelenc (1980).

Purification and preparations

Ribosomes were prepared from frozen MRE 600 cells (Public Health Laboratory service) as described by Jelenc (1980) except that the ribosomes were stored at - 8 0 ° C in polymix buffer. The purification of EF-G (Wagner and Kurland 1980), EF-Tu (Lebermann et al. 1980; Wagner et al. 1982b), EF-Ts (Arai et al. 1972a), Phe-tRNA synthetase (Phe S.) (Wagner et al. 1982b) have been described earlier.

Preparation of tRNA vhe and N-acetyl-Phe-tRNA Ph° was according to Ruusala et al. (1982b).

Translation assay

The standard protocol was as follows. Two mixtures, I and II, were prepared on ice. Both of them contained the poly-

37

mix buffer components. In addition, mixture I had ribosomes and EF-Tu at the concentrations specified for each experiment with Nac-Phe-tRNA phe in larger amounts than the ribosomes in each case. Poly(U) was added, 0.3 gg per pmol ribosome used. Mixture II contained 1.67 mM ATP, 10 mM PEP, 16.7 gg/ml pyruvate kinase, 16.7 units/ml myokinase, 16.7 units/ml nucleoside 5"-diphosphate kinase. 200 pmol/60 1~1 EF-G, unless specified, 50 units/60 gl Phe S., 0 .5mM 14C-Phenylalanine. EF-Ts, bulk tRNA or tRNA phe, GTP and ppGpp are specified for each experiment.

Mixture I was preincubated for 10 min and mixture II for 5 rain at 37 ° C. The elongation reaction was started by adding 40 ~tl mixture I to 60 gl mixture II. The assay was stopped by adding 5 ml trichloroacetic acid containing 0.5% w/v phenylalanine. Filtration and calculations were done as described by Jelenc and Kurland (1979). Back- grounds were measured by omission of Poly(U) in control samples.

Kinetic measurements

In the experiments that follow we shall be studying the influences of MSI on the elongation factors Tu, Ts and G. The assays for these effects were made in a steady state polypeptide elongation system, with one or other of the elongation factors present at a rate-limiting concentration. In this way, the effects of MSI on the rate-limiting component of the system were reflected in the rates of polypeptide elongation. The interpretation of such kinetic effects requires explicit rate equations for the different limiting concentrations; these are described as follows:

a) The EF-Tu cycle. The conventional view of the interactions between EF-Tu, EF-Ts, GTP, GDP, MSI, and AA- tRNA (Lipman 1969), can be summarized as follows:

R k l

T2 T2Ts R - " k 1

k2jrk2

T 3 + R A .

k°llk-0 T u G T P .

k 4 k 3 TuTsGTP . TuTs

k4 k3

(I) qljlq 1

TuTsMS

q211q-2

TuMS

Here, ternary complex (T3) interacts with codon-programmed ribosomes (RA) and releases the product TuGDP (T2); it is assumed for convenience that the system is being driven sufficiently by the aminoacyl tRNA (AT) as well as by the nucleoside triphosphate regeneration systems, that the back reactions over the ribosome governed by R_ are negligible, and that only the forward reaction governed by R will influence the flows over the ribosome. R, the rate factor, is the ratio kc,t/k M that describes the ribosome's interaction with ternary complex (Ehrenberg and Blomberg 1980). The interaction of the resulting binary complex with

38

EF-Ts leads to the formation of an intermediate (T2Ts), which again we assume forms a product (TuTs) in an effectively irreversible manner, i.e., the back reaction governed by k_ 2" [GDP] is negligible. In this scheme, the TuTs complex provides a critical branch point: in the absence of the magic spot analogs, it leads simply to tlie reformation of T 3 by the binding first of GTP and then the acquisition of an aminoacyl-tRNA. In the presence of a magic spot, a competitive nonproductive branch is established. This in- volves, first the formation of a TuTsMS complex and then the formation of a TuMS complex, each governed by the rate constants indicated•

A considerable simplification in the rate equations can be obtained when we consider the flows in this scheme under the two extreme conditions that most interest us ex- perimentally. One of these is the "equilibrium" approxima- tion, which corresponds to the case in which the concentration of ribosomes with an open A-site becomes negligibly small compared to the other concentration terms, IRA] ~ 0. In this limit, the rate of EF-Tu flow through the system (Jeq) is given by:

0 Tu. R- IRA]

J e q - - (1 E sl EMs )E S Ko 3 + ~ - 4 Q~ K4[AT]EGTP ] +E~-+I

(2) The dissociation constants in Eq. (2) are defined by the rate constants in scheme (1) as follows:

1/Qa=ql /q-1, 1 /Qz=q 2/q2, 1/Ko=ko/k-o,

1/K3=k3/k_ 3 and 1/K4=k 4/k4.

When the concentration of amino acyl-tRNA (AT) becomes very large in Eq. (2), all of the EF-Tu (Tu) is effectively trapped as ternary complex (T3), so that at this limit all the inhibitory effects of MS can be reversed. Accordingly for [AT] ~ m, it follows that:

Jeqo = T~ R [RA] = J +q~ (3)

Je~ and Jeq are the EF-Tu flows in the presence and absence of MS. A plot of (l/J + - 1/J~q) Jeqo versus 1/[AT] will yield a straight line. When the second term within brackets in the denominator in Eq. (2) dominates, this slope is independent of [Ts], and it is given by the expression:

Q2KoKa[MS] Q1K¢[GTP]

The detailed balance constraint requires that:

Q2 K3 _ K G T f '

Q 1 K4 KMS (4)

Here, KGT P and KMs are the dissociation constants for the formation of the corresponding binary complexes with EF- Tu. The ratio of these dissociation constants has been estimated (see Discussion).

The second limiting condition of interest to us is what we refer to as the "kinetic" case, which is seen when both the AT and R A are effectively at infinite concentrations compared to the other concentration terms in the scheme. In this limit the flow of EF-Tu (Jkin) is given by:

0 Tu

Jkin-- ( [MS] [MS]Q 2~ 1 1 1 1

l + ~ q - el [ Ts] ] R3[GTP] ~ R~[Ts] 1-k22+ ~

(5)

Here, the parameters R 1 and R 3 are defined from the rate constants in scheme (1) as:

k lk 2 k3k4 R l - k _ l_}_k2 ' R a - k _ 3 + k '

The complication inherent to Eq. (5) is that EF-Ts may be present free, or in a variety of complexes (see scheme). If, however, we consider the case in which EF-Tu is in excess of EF-Ts and EF-Ts is the rate limiting factor, we may ignore all terms in the denominator of Eq. (5) that do not contain l/[Ts], and we obtain:

o Ts

Jkin-- ( [MS] [MS] Q2) 1 ~ - ~ + - - + 1 1 1

I + Q ~ - I -I- Q1 ~uu R3[GTP] R1Tu k2 k4

(6)

This expression may be obtained directly from Eq. (5) by replacing T~ with T] and Ts with T~.

b) K 1 determinations. When the reciprocal flows (l/J) are plotted as a function of 1/[GTP] straight lines are obtained in the presence as well as absence of MS; the interpretations of these plots will depend on whether Eq. (2), (5) or (6) are relevant to the experiment• The slopes of such plots in the presence of MS (A +) and in its absence (A -) can be used to define the K~ value for the MS inhibition.

1 A + - A - 1 K, A- [MS] (7)

In addition, from Eqs. (2) and (5) we may write:

1 1 1 Q2 (8) Kl Q1 Q l [ T s ]

and from Eq. (6) we obtain

1 1 Q2 K~=Q~ + o' (9)

Q1Tu

At high concentrations of Ts in Eq. (8) and of T~ in Eq. (9), the value of K I approaches Q1. Likewisea l/K1 is a linear function of 1/ITs] in Eq. (8) and of 1/Tfi in Eq. (9)• The slopes of these lines are equivalent to the ratio Qz/Q1.

c) Inhibition of EF-Ts in vitro. When the concentration of MS becomes sufficiently high, Eq. (6) may be rewritten approximately as:

0 0

Ts. Tu. kc, t (10) J k i n - - 0

Tu + KM

where

[GTP] _ kc"t =-~MS]L Q1R3' KM= Q2' (11)

39

d) In vivo estimates. When the concentration of EF-Tu is high, that of EF-Ts is low, and the Tu factor is found either in complex with MS or GDP, the flow in Eq. (6) may be written:

T O

J k i n : [MS] 1 1 1 (12)

COTe] QIR3 ~k2 ~ k~

Now, the data described in the Results section (Table 2) allow us to estimate Q1Ra and, as we shall see, when [MS] is 0.9 mM, and [GTP] is 0.1 raM, the denominator in Eq. (12) is dominated by its first term. Hence,

0 TsQ1 R 3 [GTP] (13)

J k i n - - [MS]

Accordingly, the kc, t value in Table 2 is 4.6 s-1, and from this we deduce that QIR3 =40 s-1.

We may now ask under what conditions in vivo the formation of a TuTsMS complex will depress the concentration of T a below that of TuGDP and TuMS so that the inhibition of the flows is primarily due to the trapping of Ts in such a complex. We recall that the flow over the ribosome will be equal to that over EF-Ts, because the rate of the EF-Tu cycle in the absence of EF-Ts is negligible compared to that in its presence (Ruusala et al. 1982a). Thus,

R° (14) J k i n - - 1 1 "

QR[T3] 1 kR

This expression contains several implicit averages that we must explain. First, [T3] i is the concentration term corresponding to ternary complex for aminoacyl-tRNA cognate with a given codon; we assume that [T3]t=[T3]/20. The second average concerns QR and kR for the ribosome; when starvation for an amino acid is weak and/or the level of MS is very high, QR will be approximately equal to the R-factor for the interaction with cognate ternary complex. Our in vitro results indicate that QC=4 x 10 +TM 1 s-1 (T. Ruusala, unpublished data). The maximum turnover ra te (kcat) for ribosomes translating in vivo is close to 20 s- 1 (Maaloe 1979).

In contrast, if the degree of starvation for the amino acid is nearly complete, and the [MS] is low, the ribosome will be primarily stalled at °° starved" codons. Under these conditions, QR will correspond to the R factor for those mismatched ternary complexes that most easily pass the initial discrimination step in the ribosome (Ruusala et al. 1982b). We estimate that for the codon UUU, the rate factor for such noncognate ternary complexes is roughly 40 times smaller than that of ternary complexes containing Phe-tRNA. Thus, Q r = Q ~ = 1 0 6 M - 1 s - 1

It is known that 1 /k2+ l /k4=0 .03s (Ruusala etal. 1982a). In addition, we take the values for T ~ = R o = 10 -s M for growing bacteria (Gouy and Grantham 1980). Hence, we may now set Eq. (12) and (14) equal, and use these to determine [Ta]:

40O [Ta] = (15)

QR [ [ M S ] 1 ' \ [ G T P ] 2 ~)

One rough check on the relevance of the values of [T3] obtained in this way is that [T3] must be less than T °. Now, we take Q r = Q ~ = 4×107 M -1 s -1 and [MS]/ [GTP]=2 for the ratio during the stringent response in vivo (Gallant and Harada 1969).

From these figures and Eq. (15) we obtain [T3]= 1.5 x 10 5 M, a value approximately one-tenth of Tfi. Here, as long as the inhibition of protein synthesis is not too strong, MS will have a pronounced effect on the ternary complex concentration. In contrast, when starvation is nearly complete and only noncognate species are determining the flow over the ribosome, Q~=QW=106 M -1 s -1, there will be virtually no inhibition of the IT3] level when the [MS]/ [GTP] ratio is two. In other words, this mechanism for protein synthesis inhibition is ahnost exclusively linked to the turnover of EF-Tu at cognate codons. It can enhance significantly the accuracy of translation only when the flow in Eq. (12) is rate-limiting for protein synthesis under starvation. It can be calculated from Eq. (15) that the EF-Ts inhibition in the absence of starvation reduces the elongation rate by less than a factor of two. Accordingly, in a situation in which amino acid starvation reduces the total rate of peptide bond formation down to 10% of normal levels it does not appear likely that the inhibition of elongation factors can restore the accuracy of the bacterium.

Results

The novelty of the present study is that we have attempted to measure the kinetic effects of MSI on individual factors during steady state polypeptide synthesis. To this end we have used an in vitro system that translates poly(U) or poly(UG) at rates and accuracies comparable to those ex- pressed by ribosomes in vivo (Jelenc and Kurland 1979; Wagner et al. 1982b; Andersson et al. 1984).

The inhibition of EF-G function by MSI was studied under conditions in which the rate of poly(Phe) synthesis in the poly(U)-primed system was strictly proportional to the EF-G-concentration. Here, the GTP concentration-dependence of the polypeptide synthetic activity in the presence and absence of MSI was used to estimate a K~ value for the interaction with EF-G. When the reciprocal rate of poly(Phe) synthesis is plotted versus the reciprocal GTP concentration, the slopes of the lines in the presence (A +) and absence (A-) of MSI can be related to the K I value as defined in Eq. (7).

The data for two different concentrations of MSI as well as the control are summarized in Fig. 1; these yield a value of 8.5 x 10 .5 M for K~. A larger set of independent experiments yielded an average value of 6.7 x 10-5 M.

The inhibitory effects of MSI on EF-Tu were measured in the same way, i.e., with rate-limiting concentrations of this factor present rather than EF-G, and initially, in the absence of EF-Ts to simplify the kinetics of the system. The results of a typical experiment for this condition are shown in Fig. 2. Here reciprocal plots of poly(Phe) synthesis rate versus [GTP] in the presence as well as in the absence of MSI yield a value of 1.5 x 10 .6 M for K~. Six independent experiments yielded the average value of 7 x 10- 7 M.

In addition, filter binding assays were used to check the kinetic values obtained for EF-Tu. First the EF-TuGDP binding constant was measured; the corresponding Scat- chard plot shown in Fig. 3a yields a value of 6x 1 0 - 8 M for this equilibrium constant. Next, the affinity of MSI

40

30-

x

._= E v

% 2o ,p- X

O =, ® lO e- n

I I 110 20 3=0 40

1/ [GTP] (mM) -1

Fig. 1. GTP-dependent poly-(Phe) synthesis in the presence and in the absence of ppGpp with EF-G being rate limiting. Elongation rates were determined at various GTP concentrations as described in Materials and methods. Aliquots of 100 gl contained 70 pmol ribosomes; 200 pmol EF-Tu, 140 pmol EF-Ts, 0.2 mg bulk tRNA and 2 pmol EF-G. The reaction mixtures were incubated for 5 rain at 37 ° C. GTP concentrations were varied in the absence of ppGpp (e) and in the presence of either 0.3 mM ppGpp (A) or 0.6 mM ppGpp (*). The K l value was calculated as described in Materials and methods and is estimated to be 8.5 10 -5 M

L

o 40- • E Q. x .E E

~o 30-

_o 2 0

Q-

1 0

4 ; 810 120 1 ; 0

1/[-GTP-[ (mM) -1

Fig. 2. GTP-dependent poly-(Phe)-synthesis in the presence and in the absence of ppGpp with EF-Tu as the rate-limiting component. Elongation rates were determined at various GTP concentrations as described in Materials and methods. Aliquots of 100 gl contained 55 pmol ribosomes, 80 pmol EF-Tu, 0.5 mg bulk tRNA. No EF-Ts was present. The reaction mixtures were incubated for 30 min at 37 ° C. GTP concentrations were varied in the absence of ppGpp (o) and in the presence of 0.1 mM ppGpp (e). A K I

value was calculated as described in Materials and methods and is estimated to be 1.5 I0 -s M

was measured relative to that for GDP in a competition experiment; this yielded an equilibrium constant of 6 x 10 .7 M (Fig. 3b). This is very close to the average value of 7 x 10 v M obtained for the K I with EF-Tu.

The binding of aminoacyl- tRNA to an EF-Tu complex containing GTP is more stable than to one containing MSI (Miller et al. 1973; Pingoud etal . 1983). Therefore, we

u. oo 0. n C3

- 5 I I I

0 5 10 15 20

E G D P - I F r e e X 108 (M)

D.

I-- I

la. W

I kl. m

1 0

5-

I I 50 100

Fig. 3. a Determination of the EF-Tu GDP dissociation constant. The reaction mixture contained in 5 ml polymix buffer, 400 pmol EF-Tu and 450-850pmol GDP (o), 150pmol EF-Tu and 170M50 pmol GDP (e). The mixtures were incubated for 15 rain at 37 ° C before filtration through nitrocellulose filters. The EF-Tu GDP dissociation constant was determined from the intercept at the abscissa (see Methods). b Displacement of GDP from EF- TuGDP by ppGpp. The reaction mixture contained 30 pmol EF- TuGDP complex, 200 pmol [3H]-GDP and ~ 2 x 104 pmol ppGpp in 200 gl polymix buffer. The mixture was incubated for 15 rain at 37°C before filtration through nitrocellulose filters. The ratio of the dissociation constants was obtained from the slope of the line

might expect that under equilibrium conditions high concentrations of amino acyl- tRNA would favor the binding of GTP and oppose the inhibitory effects of MSI. This expectation was tested by studying the influence of charged tRNA on the rate of poly(Phe) synthesis in the presence and absence of MSI when EF-Ts is omitted from the system. The reciprocal elongation rates are plotted versus the reciprocal Phe- tRNA concentrations for one such experiment shown in Fig. 4. Two straight lines are observed: one for the presence and the other for the absence of MSI. Nevertheless, no significant influence of the charged tRNA on the degree of inhibit ion by MSI was detected even though very high concentrations of the Phe- tRNA were present. This result indicates that in the absence of EF-Ts, the EF-Tu does not equilibrate fast enough with the nucleo-

A I n o E

x

.E

% x

o = ,=

6 0

40

20

• w • C

I I 0 , 0 1 0,02

F - - I

, / / , 0 , 0 3 O, 1

I / t R N A Phe (pmol) -1

Fig. 4. Poly-(Phe)-synthesis was measured in the presence and in the absence of ppGpp with different concentrations of tRNA Ph°. Aliquots of 100 gl contained 12 pmol ribosomes, 300 pmol EF-Tu. No EF-Ts was present. The reaction mixture was incubated 10 min at 37 ° C. GTP 1 mM (e) and GTP 0.3 mM and ppGpp 0.7 mM (o)

tide in this steady state for there to be a detectable influence of P h e - t R N A on the degree of inhibition. We will return to this kinetic effect after we have considered the influence of EF-Ts on MSI inhibition.

EF-Ts has been shown to accelerate the release of MSI from EF-Tu as well as that of GTP (Miller et al. 1973). We could study the expected influence of EF-Ts on the compet i t ion of GTP and MSI for EF-Tu by keeping the EF-Tu concentrat ion so low that the rate of poly(Phe) elon- gat ion is l imited by this factor both in the presence and absence of EF-Ts. The da ta in Fig. 5 a summarize measurements of the K~ for MSI at several concentrat ions of EF-Ts. The K I values increase with increasing EF-Ts concentrat ion. Indeed, when reciprocal plots are made of the K~ values and EF-Ts concentrat ions (Fig. 5b), a linear relat ionship is seen. The intercept of this plot gives the limiting value of K~ at infinite [EF-Ts]; 4.6 x 10 5 M. Thus, EF-Ts in- creases the K I of MSI with EF-Tu by almost two orders of magni tude from its value of 7 x 10-7 M obtained in the absence of EF-Ts.

As suggested by Miller et al. (1973), a natura l intermediate for the exchange react ion of the nucleotides with EF-Tu would be the binary complex of EF-Tu and EF-Ts. Accord- ingly, the influence o f EF-Ts on the K I of MSI could result f rom the fact that this factor changes the target of MSI from EF-Tu to the b inary factor complex. In the lat ter case the format ion of an E F - T u M S I complex would involve the release of EF-Ts from a complex of E F - T u E F - T s M S I . According to such a scheme the K I value for MSI inhibit ion of EF-Tu is given by Eq. (8) in Mater ia ls and methods.

The da ta in Fig. 5b yield a value of Q 1 = 4 . 6 x 10 .5 M and of Q 2 = 7 x 10 .7 M. Q1 is the dissociation constant of MSI from the E F - T u E F - T s M S I complex and Q2 is the dissociation constant of EF-Ts from the same ternary complex (see Mater ia ls and methods). Averages from multiple experiments are 4.3 x 10 - s M for Q1 and 4.9 x 10 -7 M for Q2.

These da ta suggest that the E F - T u E F - T s M S I complex is rather stable. Therefore, the format ion of this ternary complex may be an impor tan t par t of the effect of MSI on t ransla t ion in vivo. We can now return to the influence of charged t R N A on the inhibi t ion by MSI.

We repeated the t i t rat ion of the poly(Phe) synthesizing system with MSI as a function of the P h e - t R N A concentration, but this time we included high concentrat ions of EF-

41

J m o E Q. × 2 4 0 " .E E

% 200 x

0 3" 1 6 0 ,

120-

80 -

40~

10 I I

20 30

I / [ G T P ] ( m i ) -1

0 , 6

x

~- 0 ,4

0.2 ̧

I f I I 1 2 3 4

I / E F - T s x 10 =6 (M) -1

Fig. 5. a GTP-dependent poly-(Phe)-synthesis in the presence and in the absence of ppGpp with EF-Tu as the rate limiting component. Elongation rates were determined at various GTP concentrations as described in Materials and methods. Aliquots of 100 gl contained 100 pmol ribosomes; 3 pmol EF-Tu; 400 pmol EF-G, 0.5 mg bulk tRNA. The reaction mixtures were incubated for 1 min at 37 ° C. The GTP concentration was varied in the absence of ppGpp and in the presence of 30 (0); 50 (zx) and 100 (o) pmol EF-Ts and in the presence of 0.1 mM ppGpp and 30 ( .) ; 50 (A) and 100 ( .) pmol EF-Ts. b EF-Ts dependence of the Kl-value. From experiment in a a K I value for each EF-Ts concentration was calculated. The data are represented in a double reciprocal plot where Q1 and Q2 are obtained as described in Materials and methods

Ts. Two such experiments are depicted in Fig. 6: one at 5 x 10 - s M ribosomes and the other at 10 .6 M ribosomes. When reciprocal plots are made for the rate of elongation (J) versus P h e - t R N A concentrat ion, straight lines are obtained both in the presence (J+) and absence ( J - ) of MSI. At the lower r ibosome concentrat ion, P h e - t R N A at high concentrat ion fully reverses the MSI-dependent inhibit ion; this is indicated by the coincidence of the intercepts: 1/Jg and 1/J o. In contrast , at higher r ibosome concentrat ions

42

A

x 1,8- "- • • .~:

v

1,4- o

I © t-.

n

= 1

kk m ©

X 0,6

0,2-

I I I I I 0,002 0,006 0,O1

I / tRNA Phe (prnol) -1

Fig. 6. Poly-(Phe)-synthesis was measured in the presence and in the absence of ppGpp at different concentrations of tRNA vh°. Ali- quots of 100 gl contained 150 pmol EF-Ts. One set of reaction mixtures was incubated 1 min at 37 ° C and contained 5 pmol ribosomes, 400 pmol EFG and either (a) 3 pmol EFTu, I mM GTP (zx) or (b) 10 pmol EF-Tu, 0.8 mM ppGpp, and 0.2 mM GTP (o). Another set of mixtures contained 100 pmol ribosomes, 1000 pmol EF-G and either (a) 6 pmol EF-Tu, 1 mM GTP (A) or (b) 20 pmol EF-Tu, 0.8 mM ppGpp, 0.2 mM GTP (e)

Table 1. Experiments as described in Fig. 6 with additional concentrations of ribosomes

[Ribosomes] x 108 M Ratio of intercepts J o / J ~

100 4.86 50 2.3 25 1.63 25 1.51 20 1.38 20 1.14

5 1

there is only a part ia l reversal of the inhibit ion at infinite [Phe-tRNA] i.e. l / J o + > 1/J o-

The da ta obtained in experiments performed with different r ibosome concentrat ions are summarized in Table 1. Here, the maximum rates of poly(Phe) elongation extrapolated to infinite [Phe-tRNA] from the reciprocal plots in the presence (1/J~-) and absence ( l / Jo ) o f MSI are tabu- lated. These da ta substantiate in detail the tendency of high Phe - tRNA concentrat ions to reverse the inhibi tory effect of MSI at low r ibosome concentrations.

We would expect that if there is stable complex formation between EF-Tu, EF-Ts and MSI, it will be reflected in the ways that the system responds to MSI when EF-Tu is in large excess over EF-Ts, and EF-Ts is limiting the rate of poly(Phe) synthesis. Equat ion (9) in the Materials and methods section shows how K~ should vary with the concentrat ion of EF-Tu (T~).

The da ta from such a variat ion are shown in Fig. 7a, which displays the reciprocal plots of elongation rate versus 1/[GTP] at different EF-Tu concentrations. Figure 7 b sum-

8O x t -

60

x

i 40 •

2 0

4 8 12 I

16 20

I/~GTP'] (mM) -1

g "?

O ,e- x 1

0,~

I I I I I 1 2 3 4 6

1 / E F - T u x 10 -6 (M) -1

Fig. 7. a GTP-dependent poly-(Phe)-synthesis in the presence and in the absence of ppGpp with EF-Tu as the rate-limiting component. Elongation rates were determined at various GTP concentrations as described in Materials and methods. Aliquots of 100 Ixl contained 100 pmol ribosomes; 3 pmol EF-Ts; 400 pmol EF-G, 0.5 mg bulk tRNA. The reaction mixtures were incubated for 1 rain at 37 ° C. The GTP concentration was varied in the absence of ppGpp and in the presence of 25 (,); 50 (e); 100 (n); 150 (4) pmol EF-Tu respectively, and in the presence of 0.1 mM ppGpp and 25 ( . ) ; 50 (e); 100 (o); 150 (o) pmol EF-Tu, respectively. b EF-Tu dependence of K] values. The K= value for each EF-Tu concentration was calculated from the data in a. The data are represented in a double reciprocal plot where Q1 and Q2 are obtained as described in Materials and methods

marizes the reciprocal plots of the corresponding values for the K~ as a function of EF-Tu concentrat ion: The slope and intercept of this plot yield values of Q1 = 5 x 10 5 M and Qz = 7 x 10-7 M. These values are very similar to those obtained in the EF-Ts t i t rat ion (Fig. 5).

The agreement between the two independent estimates of Q1 and Q2 provides strong support for the existence of the ternary complex containing EF-Tu, EF-Ts and MSI. Fur ther support was obtained from yet another sort of experiment. Here, the system was arranged with EF-Ts limiting the rate of poly(Phe) incorporat ion, and the EF-Tu- dependence of the activity in the presence as well as the absence of MSI was measured. Figure 8 summarizes these da ta in the form of reciprocal plots, from which the kca t and K M values for EF-Ts-media ted regeneration of ternary

43

i m o E e~

~<

E v

% ×

o

I

g.

lO

8

6

4

2

I I 0 1 2

1 / E F - T u x 10 2 ( pmo l ) -1

Fig. 8. EF-Tu dependent poly-(Phe) synthesis in the presence and in the absence ofppGpp with EF-Ts being rate limiting. Elongation rates were determined at various EF-Tu concentrations. Aliquots of 100 ~1 contained 100 pmol ribosomes, 200 pmol EF-G; 1.5 pmol EF-Ts; 300 pmol tRNA Phe. The reaction mixtures were incubated for different times at 37 ° C. The EF-Tu concentration was varied and the nucleotide concentrations were: 1 mM GTP ( , ) ; 0.3 mM ppGpp and 0.7 mM GTP (A); 0.6 mM ppGpp and 0,4 mM GTP (*); 0.9 mM ppGpp and 0.1 mM GTP (e). K M values and kca t values were calculated as described in Materials and methods

Table 2. K M and kea t from 5 different experiments as in Fig. 8

[ppGpp]mM kc. t s- i

0.9 5.7 6.5 5.4 4.6 7.5 0.8 8.5 9.6 8.4 9.1 0.6 12.0 21.7 18.5 13.9 14.0 0.3 31.3 0.0 83.8 115.0 75.5 46.5 51.5

[ppGpp]mM K M 106 M

0.9 4.6 7.3 4.9 2.0 8.4 0.8 4.0 4.8 3.4 5.3 0.6 1.2 4.6 3.7 2.6 3.0 0.3 2.5 0.0 2.2 4.5 2.5 1.6 1.6

complex are calculated at the different MSI concentrations. These calculations are summarized in Table 2.

In the absence of MSI an average K M = 2 . 2 x 1 0 - 6 M and an average koa t = 74 s - 1 are obtained. These values may be compared with those repor ted by Ruusala et al. (1982a), who found KM=2.5 x 10 .6 M and koat=33 s 1. As can be seen in Table 2, the K M values are more reproducible than the koa t values. We at t r ibute the variabi l i ty of the ko, t values to trace amounts of EF-Ts that contaminate our E F - G prepara t ions to varying extents. This contamina t ion leads to a small, systematic overest imation of the k~a t values so that we are more confident in the earlier value obtained by Ruusala et al. (1982a) with exceptionally clean E F - G preparat ions. I t is nevertheless clear from the da ta in Ta- ble 2 that MSI influences the KM values only to a slight extent, but that it has a systematically stronger effect on

¢..

_o 3"

"S

I IJ . W

° ~

I I I I 0,005 0,01 0,015 0,02

I / t R N A Phe (pmol) -1

Fig. 9. Poly-(Phe)-synthesis was measured in the presence and in the absence of ppGpp along with different concentrations of tRNA Ph°. Aliquots of 100gl contained 90pmol ribosomes, 120 pmol EF-Tu; 6 pmol EF-Ts. The reaction mixtures were incubated for different times at 37 ° C. GTP 1 mM (o) and GTP 0.4 mM, ppGpp 0.6 mM (e)

the kca t values. This is precisely what would be expected if the adsorpt ion of EF-Ts to a complex with EF-Tu and MSI were taking place.

Similar results have been obtained at both higher and lower EF-Ts concentrat ions (data not shown). In the limit of very high MSI concentrat ions the affinity between EF-Ts and the E F - T u M S I complex is approximate ly equal to the KM value (see Materials and methods). Indeed the average value o f K M is close to 2 x 10 .6 M at all MSI concentrat ions (Table 2). Given the approximate character o f this compar i - son and the accuracy of our calculations, we take the values of 5 x 1 0 - T M and 7 x 1 0 - 7 M obta ined above for Q2 as being in reasonable agreement with this K M estimate of

Q2- Finally, we have studied the effect of increasing concen-

trat ions of P h e - t R N A on the inhibit ion by MSI under the condit ions of the experiment described in Fig. 8. Here too the influences of charged t R N A concentrat ion on the degree of inhibit ion is relatively small (Fig. 9).

Discussion

We intend to use the da ta obtained in the present experiments to estimate the effect of MSI on the t ranslat ion system in vivo. I t is, therefore, relevant to stress two aspects of this da ta set. First , we have used an in vitro system that mimics the kinetic characteristics of t ranslat ion in vivo (Wagner et at. 1982b; Andersson et al. 1984). While this does not guarantee the validity of our results, it certainly reduces the l ikel ihood that they are completely irrelevant. Second, we have systematically explored the effects of MSI under well defined but different kinetic conditions. This permits us to assess the impact of the different effects of MSI under those condit ions that are relevant to the stringent response in vivo.

The inhibit ion of E F - G by MSI seems to be uncompli- cated, and it can be characterized by a KI = 6 × 10-5 M. This weak interact ion can lead to a nearly twofold reduction in the active E F - G pool when the [MSI]/[GTP] ratio is two (Fig. 2), which sets the upper limit for the expected inhibit ion of t ranslat ion through E F - G by MSI in vivo.

44

Thus, it has been estimated that the free EF-G concentration in vivo is close to 10-5 M (Gouy and Grantham 1980), and the rate factor (kc,t/KM) for the EF-GGTP interaction with ribosomes has been estimated at 3 x 10 ~ M - 1 s- 1 (T. Ruusala, unpublished data). These values are consistent with a 3 ms waiting time for a translating ribosome to ac- quire an EF-GGTP complex. In contrast, the whole elongation cycle in vivo takes more than 50 ms. Therefore, a dou- bling of the waiting time for this factor most probably will have only marginal effects on the overall rate of elongation in vivo. In addition, the effect of EF-G inhibition would be further reduced during starvation for an aminoacyl- tRNA species, which would tend to lower the concentration of pretranslocational ribosomes and, in equal measure, reduce the kinetic dependence of elongation on the effective EF-G concentration.

Pingoud et al. (1983) assumed that the EF-GMSI interaction has the same binding constant as that for EF-GGDP, which led them to greatly overestimate the affinity of MSI for EF-G. Furthermore, they assumed in their calculations that the translation system is nearly at equilibrium during the stringent response in vivo. Accordingly, their assertion that the critical factor interaction with MSI is that for EF-G and not that with EF-Tu would require that the EF-GGTP concentration is reduced to negligible amounts during the stringent response. This is not possible if the interaction of MSI with EF-G is as weak as the present data suggest.

The interactions of MSI with the EF-Tu cycle are considerably more complicated because they can be influenced by the aminoacyl-tRNA, by the GTP levels and, most im- portantly, by the interactions with EF-Ts. It is, accordingly, helpful to try to sort out these different effects by distinguishing two kinetic states of the EF-Tu, each characterized by the turnover rate of the factor. In the first state, the turnover rate is very low, as for example when there is a relatively low concentration of ribosomes, and there is a large excess of EF-Ts. In this extreme state the EF-Tu will approach equilibrium with the nucleotides competing for a binding site in the factor. The other extreme is characterized by a rapid turnover of the EF-Tu, as when there are relatively high concentrations of ribosomes and low concentrations of EF-Ts. The latter case is of greatest po- tential relevance to the analysis of the stringent response, since the problem that concerns us there is to determine how polypeptide synthesis is carried out accurately even though there is an imbalance in the pools of ternary complexes cognate to as well as noncognate to the "starved codon".

A distinguishing characteristic of the inhibition of the Tu-cycle in the "equilibrium" state is its sensitivity to aminoacyl-tRNA, which has been emphasized by Pingoud etal. (1983). Thus, by going to sufficiently high Phe- tRNA phe concentrations and to sufficiently low ribosome concentrations the inhibition by MSI can virtually be abol- ished (Fig. 6; Table 1). Pingoud et al. (1983) used some of the data of Gouy and Grantham (1980) to estimate the following concentrations in vivo: EF-Tu, 1.7× 1 0 - 4 M ; aminoacyl-tRNA, 1.5 × 10 -4 M; GTP, 2 × 10 -3 M; and MSI, 4 × 10 -3 M. I f we assume that EF-Tu during the stringent response near equilibrium can be found only as free ternary complex or as binary complex with MSI, we can use these concentrations as well as the inhibition curves described in Fig. 6, and Table 1 to calculate the partitioning of EF-Tu between these two states (see Materials and meth-

ods). We obtained results qualitatively similar to those of Pingoud et al. (1983); at an MSI/GTP ratio of two the ternary complex concentration has been reduced by roughly 10% and this is accompanied by the accumulation of EF- TuMSI at a concentration of 3 x 10 -5 M. Thus, the present data confirm in part the conclusions of Pingoud et al. (1983) that at equilibrium, the influence of MSI in vivo on the rates of elongation should be marginal.

In contrast, the influence of even high aminoacyl-tRNA concentration on the MSI-mediated inhibition of EF-Tu is marginal in the high turnover, kinetic domain (Fig. 4, 9). This sort of aminoacyl-tRNA-insensitive inhibition takes place when the activity of the ribosomes is not negligible, and they can drive the EF-Tu far from equilibrium. Thus, this state is relevant when the degree of starvation is not very great. Here, the consumption of EF-TuGTP at unstarved as well as at the starved codons must be main- tained at rates close to the nonstarved state. If this is so, the stringent response will lead to a preferential partitioning of EF-Tu between complexes with GDP and those with MSI. The critical condition here is that EF-Ts is bound to these binary complexes to such a degree that the regeneration of EF-TuGTP is retarded. As we have shown in Mate- rials and methods, an MS/GTP ratio of two in this kinetic domain can lead to a tenfold reduction of ternary complex concentration.

In summary, the idea (O'Farrel 1978) that inhibition of elongation rates by MSI will maintain the accuracy of translation during amino acid starvation seems in doubt. Miller et al. (1973) as well as Pingoud et al. (1983) have favored the view that inhibition of EF-G but not EF-Tu might lead to a significant reduction in the elongation rate. In contrast, we had favored the view that inhibition of EF- Tu might do the same and thus account for the stringent control of accuracy (Wagner et al. 1982 a). The present data persuade us that neither suggestion will work during strong starvation for an amino acid. We suggest that the influence of the stringent response on the accuracy of translation does not involve the lowering of the activity of the elongation factors per se.

Our working hypothesis is that the accuracy of translation is enhanced during the stringent response by a direct effect of the EF-TuMSI complex on the proofreading rates of the ribosome (Kurland and Ehrenberg 1984). We have observed some effects in vitro that are consistent with this suggestion (Rojas, unpublished data). The present results suggest that the concentrations of EF-TuMSI complex during the stringent response (circa 3 x 10 -5 M) will be adequate to support such a proofreading enhancement in vivo. Finally, systematic variations in the magic spot concentrations have been observed for bacteria growing exponentially in different media (Ryals et al. 1982). This might be related to the present demonstration that the rate of Tu cycling can be influenced by MSI through its effect on the availabil- ity of EF-Ts when the system is close to saturation. Thus, MSI may affect a medium-dependent fine tuning of the translation system through its influence on the EF-Ts-EF- Tu cycle under nonstarvation conditions.

Acknowledgements. This work was supported by grants from the Swedish Cancer Society and the Swedish Natural Science Research Council. Ana-Maria Rojas would like to acknowledge a Venezue- lian scholarship from the Consejo Nacional de Investigaciones Cientificas y Tecnologicas (CONICIT).

45

R e f e r e n c e s

Andersson SGE, Buckingham RH, Kurland CG (1984) Does codon composition influence ribosome function? EMBO J 1 : 9 1 - 9 4

Arai KI, Arai N, Kawakita M, Kaziro Y (1972a) Interaction of guanine - 5'-diphosphate, T-(or 3'-)disphosphate (ppGpp) with elongation factors from E. coli. Biochem Biophys Res Commun 48:190-196

Arai KI, Kawakita M, Kaziro Y (1972b) Studies on polypeptide elongation factors from Escherichia coli. J Biol Chem 247 : 7029-7037

B6ck A, Faiman LE, Neidhardt FC (1966) Biochemical and genetic characterization of a mutant of Escherichia coli with a tempera- ture sensitive valyl ribonucleic acid synthetase. J Bacteriol 92:1076-1082

Ehrenberg M, Blomberg C (1980) Thermodynamic constraints on kinetic proofreading in biosynthetic pathways. Biophys J 31:333 358

Gallant J (1979) Stringent control in E. coli. In: Roman HLK, Campbell A, Sandier LM (eds) Annual review of genetics, vol 13. Ann Rev Inc, Palo Alto, California, USA

Gallant J, Harada B (1969) The control of ribonucleic acid synthesis in Escherichia coli. III The functional relationship between purine ribonucleoside triphosphate pool sizes and the rate of ribonucleic acid accumulation. J Biol Chem 244: 3125-3132

Gallant J, Foley D (1979) On the causes and prevention of mis- translation. In: Koch G, Richter D (eds) Regulation of macromolecular synthesis by low molecular weight mediators. Aca- demic Press Inc, New York

Gouy M, Grantham R (1980) Polypeptide elongation and tRNA cycling in Escherichia coli: A dynamic approach. FEBS Lett 115:151-155

Hamel E (1976) Derivatives of guanosine triphosphate with ribose T-hydroxyl substituents. Interactions with the protein synthetic enzymes of Escherichia coll. Eur J Biochem 70:339-347

Hamel E, Cashel M (1973) Role of guanine nucleotides in protein synthesis. Elongation factor G and guanosine 5'-triphosphate, 3'-diphosphate. Proc Natl Acad Sci USA 70:3250-3254

Jelenc PC (1980) Rapid purification of highly active ribosomes from Escherichia coli. Anal Biochem 105:369-374

Jelenc PC, Kurland CG (1979) Nucleoside triphosphate regeneration decreases the frequency of translation errors. Proc Natl Acad Sci USA 76: 3174-3178

Kurland CG, Ehrenberg M (1984) Optimization of translational accuracy. Progr Mol Biol Nucl Acids Res, in press

Leberman R, Antonsson B, Giovanelli R, Guariguata R, Schu- mann R, Wittinghofer A (1980) A simplified procedure for the

isolation of bacterial polypeptide elongation factor EF-Tu. Anal Biochem 104:29-36

Legault L, Jeantet C, Gros F (1972) Inhibition of in vitro protein synthesis by ppGpp. FEBS Lett 27 : 71 75

Lipmann F (1969) Polypeptide chain elongation in protein biosyn- thesis. Science 164:1024-1031

Maaloe O (1979) Regulation of the protein-synthesizing machinery-ribosomes, tRNA, factors and so on. In: Goldberger RF (ed) Biological regulation and development. Plenum Press, New York, p 487-542

Miller DL, Cashel M, Weissbach H (1973) The interaction of guanosine 5'-diphosphate, 2'(3')-diphosphate with the bacterial elongation factor Tu. Arch Biochem Biophys 154:675-682

O'Farrell PH (1978) The suppression of defective translation by ppGpp and its role in the stringent response. Cell 14:545-557

Piepersberg W, Geyl D, Buckel P, B6ck A (1979) Studies on the coordination of tRNA-charging and polypeptide synthesis activity in E. coli. In: Koch G, Richter D (eds) Regulation of macromolecular synthesis by low molecular weight mediators. Academic Press Inc, New York

Pingoud A, Gast F-U, Block W, Peters F (1983) The elongation factor Tu from Eseherichia eoli, amino acyl-tRNA, and guanosine tetraphosphate form a ternary complex which is bound by programmed ribosomes. J Biol Chem 258 : ~[4200-14205

Ruusala T, Ehrenberg M, Kurland CG (1982a) Catalytic effects of elongation factor Ts on polypeptide synthesis. EMBO J 1 : 75-78

Ruusala T, Ehrenberg M, Kurland CG (1982b) Is there proof reading during polypeptide synthesis? EMBO J 1:741-745

Ryals J, Little R, Bremer H (1982) Control of rRNA and tRNA syntheses in Escherichia coli by guanosine tetraphosphate. J Bacteriol 151:1261-1268

Wagner EGH, Kurland CG (1980) Translational accuracy enhanced in vitro by (p)ppGpp. Mol Gen Genet 180:139-145

Wagner EGH, Ehrenberg M, Kurland CG (1982a) Kinetic suppression of translational errors by (p)ppGpp. Mol Gen Genet 185: 269-274

Wagner EGH, Jelenc PC, Ehrenberg M, Kurland CG (1982 b) Rate of elongation of polyphenylalanine in vitro. Eur J Biochem 122:193-197

Yegian CD, Stent GS (1969) An unusual condition of leucine transfer RNA appearing during leucine starvation of E. coli. J Mol Biol 39:45-58

Communicated by A. B6ck

Received March 12 / April 27, 1984

ppgpp inhibition of elongation factors tu, g and ts during polypeptide synthesis

Documents