Eur. J. Biochem. 220, 739-744 (1994) 0 FEBS 1994

Escherichia coli elongation-factor-% mutants with decreased affinity for aminoacyl-tRNA Carsten ANDERSEN and Ove WIBORG Institute of Chemistry, Aarhus University, Aarhus, Denmark

(Received November 10, 1993/January 12, 1994) - EJB 93 1638/2

The two evolutionary well-conserved histidine residues, His66 and His11 8, of Escherichia coli elongation factor Tu have been subjected to mutational analysis. The two histidines have each been replaced by alanines, denoted H66A and H11 8A, respectively. His118 has also been substituted by glutamate, H11 SE. The three mutants have been characterized with respect to thermostability, GTPase activity and affinity for aminoacylated tRNA. Most conspicuously, the tRNA affinity is reduced or almost abolished. k- , for dissociation of the ternary complex increases by factors of 14, 40 and 48 for H66A, H118A and H118E, respectively, when compared to the wild type. The half- lives for the non-enzymic deacylation of aminoacylated tRNA in the ternary complex are 391, 107, 69, 54 and 61 min for wild type, H66A, H118A, H118E and free aminoacylated tRNA, respectively. The Kd is about 20-times higher for H66A compared to wild type.

Our results strongly suggest that His66 and His118 play major roles in stabilization of the ternary complex.

Elongation factor Tu (EF-Tu) from Escherichia coli is a monomeric protein consisting of 393 amino acid residues, and with a molecular mass of 43 kDa. During the elongation step in protein biosynthesis EF-Tu promotes the binding of aminoacyl-tRNA to the A-site of the programmed ribosome in the process of decoding the mRNA [l]. In its active con- formation with GTP as a cofactor, EF-Tu has a high affinity for aminoacyl-tRNA with equilibrium dissociation constants in the range of 10-’M [2], the affinity for non-acylated tRNA being several orders of magnitudes lower [3].

The low intrinsic GTPase activity of EF-Tu is stimulated on the ribosome when EF-Tu is present in the form of a ternary complex with GTP and aminoacyl-tRNA. Upon GTP hydrolysis EF-Tu undergoes a conformational change to a GDP form with negligible affinity for aminoacyl-tRNA. The exchange factor EF-Ts catalyses the replacement of GDP with GTP and thus completes the cycle.

The crystal structure of E. coli EF-Tu in complex with GDP has been determined at high resolution [4] and so has the structure of EF-Tu :GTP from the thermophilic bacteria Thermus aquaticus and Thermus thermophilus [5, 61.

Theoretical speculations about the biochemical nature of the important intermediate in the elongation process, the ter- nary complex, has brought about many interesting studies over the last two decades [7, 81. No unequivocal determina- tion of macromolecular recognitory interaction sites has so far been established.

Correspondence to 0. Wiborg, Institute of Chemistry, Aarhus

Fax: +45 86196199. Abbreviations. EF, elongation factor; R value, correlation coeffi-

Enzyme. Pyruvate kinase (EC 2.7.1.40).

University, Langelandsgade 140, DK-8000 Aarhus C, Denmark

cient.

To map the interaction points of EF-Tu and aminoacyl- tRNA different chemical modification studies have been ap- plied in addition to cross-linking and site-directed mutagene- sis studies.

Modification studies of Cys81 indicated that this residue took part in tRNA binding [9, 101. This conclusion could not, however, be confirmed by mutagenesis studies [11].

In three different studies [12-141 several lysine and argi- nine residues have been indicated to be involved in tRNA binding, in particular Lys2, Lys4, and Lys263. As EF-Tu is expected to recognize some overall structural features of the many different tRNA molecules, e.g. the folding of the phos- phate-ribose backbone, lysines and arginines might indeed be good candidates for formation of salt-bridges with tRNA backbone phosphates in stabilization of the ternary complex.

Modification studies on the ternary complex have re- vealed that His66 and His118 and at least one of the methio- nines 91,98, or 112 are protected against chemical modificat- ion [15, 161. This indicates that these residues make up part of a tRNA-binding site. In support of this, His66 has been successfully cross-linked to a modified aminoacyl-tRNA ; W- bromoacetyl-Lys-tRNA [17] indicating that His66 binds the aminoacyl moiety of aminoacyl-tRNA.

In another attempt, EF-Tu fragments 56-68 and 118- 124 were identified as cross-linking to aminoacyl-tRNA [ 181. The cross-linking agent was trans-diamminedichloroplati- num and the cross-linking sites potentially His66 and Hisll8.

Taken together the results obtained so far in mapping of the tRNA-binding site on EF-Tu strongly points towards some basic residues and His66 and Hisll8. For this reason we have undertaken the task to characterize the importance of the two histidines by site-directed mutagenesis. His66 was mutated to alanine, H66A; and His118 to either alanine or glutamate, H118A and H118E.

740

MATERIALS AND METHODS Characterization of the ternary complex Construction of mutants In the following assays, 3 pM EF-Tu:GDP was incubated

for 20 min at 20°C in the presence of 1 mM GTP, 2 mM Site-directed mutagenesis was performed according to phosphoenolpyruvate and o.l mg/ml pyruvate kinase.

Taylor et al. [19] using a M13mpll clone containing the tufA tRNAPhe was aminoacylated by incubating 14 pM tRNAPhe gene encoding EF-Tu [201 as a template. The sequence of with 50 pM [3H]Phe, 5 mM ATP, 0.24 CTP and 1 p1 the mutagenic primer for mutating His66 to Ala was 5'- synthetase in 250 pl charging buffer Tris/HC1, CAACACTTCTGCTGTTGAATACG-3' and for mutational pH 7.5, 50 mM NH,C~, 12 mM MgC12, 2.8 m~ 2-mercapto- analysis of Hisll8,5'-GACTCGTGAGGMGATCCTGCTG- at 3 7 0 ~ for 10- 15 min. The synthetase was purified GGTC-3'3 where is Or A for mutating to Or crudely from Saccharomyces cerevisiae according to von der

[21]. The recA- strain JM109 [22] was used as a host. Expected mutations were identified by ssDNA sequenc-

ing of the M13mplltufA constructs, using Sanger's dideoxy- nucleotide chain-termination method [23] and confirmed on the expression vector, pGEXFXtufA, by the dsDNA cycle sequencing method [24].

Expression and purification

formed as described earlier 1201.

Upon mutagenesis the tufA mutant gene was Haar et al. [29]. For use in the non-enzymic hy&olysis assay the synthetase was removed, after charging, by phenol ex- recloned and in the PGEX gene fusion traction and Phe-tRNAPh" precipitated.

The equilibrium dissociation constant, Kd, for formation of the ternary complex EF-Tu :GTP: yeast [3H]Phe-tRNAPhc was determined by the ~bonuclease-resistance assay essen- tially as described by Louie and Jurnak [2]. The concentra- tions of EF-Tu and aminoacyl-tRNA were, however, kept in the micromolar range due to a considerably higher Kd for the mutants. 0.67 pM [3H]Phe-tRNAPhe (specific activity 574 cpdpmol) was incubated with EF-Tu :GTP (0.05 -

KCl, 30 mM NH,CI, 6 mM MgC1,) at 0°C for 30 min. 1 p1 RNase A (10 mg/ml) was added for 15 and the amount of

washing. The Kd values were determined from a Scatch- ard plot [30]. The ternary-complex dissociation-rate constant k- was determined by the ribonuclease-digestion rate assay as described by Louie and Jurnak [2] . In 100 pl PP-6 buffer, 2.5 pM EF-Tu:GTP was incubated with 0.7 pM [3H]Phe- tRNAPhe (specific activity 574 cpm/pmol) at ooc for 30 min. At time zero, R N ~ ~ ~ A was added to a final concentration

precipitated and filtered. As the dissociation of the complex follows first-order kinetics, k_ , may be determined as the slope of a plot of ln(lEF-Tu :GTP:Phe-tmA]M) time.

The protective effect of EF-T~ in the ternary complex against non-enzymic hydrolysis of the minoacyl bond was measured as prescribed by Pingoud and Urbanke [311.

Expression and purification of mutant proteins was per-

Protein concentrations were determined with an accuracy

serum albumin as a standard. The yields of the purified mu- tant proteins were in the range Of mg/g wet paste.

Protein activity and stability GDP-binding activity was taken as an overall measure

for the EF-Tu activity and its native state, although, as may

ability of EF-Tu to bind nucleotide is retained longer than the ability to bind aminoacyl-tRNA. GDP-binding activity was measured upon equilibration with [3H]GDP (specific ac- tivity 700 cpdpmol) and binding to nitrocellulose filters, essentially performed as described by Miller and Weissbach 1261. The thermal-inactivation profiles for the GDP and GTP forms were determined as described by Giimiisel et al. 1271.

pM) in 50 pl pp-6 buffer (60 mM Tris/HC1, pH 7.8,30

Of 5 % a modified method Of Bradford [25i using bovine ternary complex measured after precipitation, filter binding

be Seen from the characterization Of the the of 30 pg/ml. Aliquots of 15 pl were successively withdrawn,

10 pmol EF-Tu was incubated with 200 pmol [3H1GDP/GTP in 50 p1 binding buffer (50 200 pp-6 buffer, 1 pM EF-Tu:GTp was incubated with

0.3 pM [3H]Phe-tRNAPhe (specific activity 574 cpm/pmo]) at

20 pl were withdrawn, precipitated and filtered. From a plot

TrisHcl, pH 7.6* loo NH4Cl9 10 mM MgC12, 50 mM KCl, 1

ature, then on ice for 30 min.

GTPase activity assay The GTPase activity was measured as the amount of lib-

erated inorganic phosphate, using the isopropylacetate/mo- lvbdate method 1281. Iv"P1GTP and EF-Tu:GDP were incu-

dithiothreitol). ooc for 30 min. ~ ~ l l ~ ~ i ~ ~ incubation at 2 0 0 ~ aliquots of

of In c,/c, versus time, where c, and c, are the concentrations of the ternary complex at time t and 0, respectively, the slope may be taken as a relative measure of the protective effect of EF-Tu on the aminoacyl bond.

In vitro translation assay

The mixture was incubated for min at the assay temper-

L ~ -. ~

dated separately with phosphoenolpyruvate and pyruvate ki- nase prior to the assay. 45 pmol EF-Tu:GTP was incubated at 20°C in 50 mM TrisHCl, pH 7.6, 5 mM MgC12, 50 mM KCl, 0.5 mM dithiothreitol and 1.5-30 pM [Y-~~PIGTP (spe- cific activity 1700 cpdpmol). Samples were withdrawn ev- ery 5 min up to 30 min and the reaction stopped by addition of perchloric acid. K , and kcat were determined from Line- weaver-Burk plots. For characterization of the GTPase activ- ity 0.5 pM EF-Tu was incubated with 60 pM GTP and stim- ulated by addition of 50 pM kirromycin, 0.5 pM aminoacyl- tRNA or 0.5 pM poly(U)-programmed ribosomes blocked in the P-site with uncharged tRNAPhe. Ribosomes were removed by centrifugation prior to extraction of liberated inorganic phosphate.

The poly(U)-directed poly(Phe) synthesis assay was car- ried out as described by Swart and Parmeggiani [32]. A 240- pl reaction mixture containing 60 mM Tris/HCI, pH 7.8, 30mM KC1, 30mM NH4Cl, 6mM MgC12, 0.18 mM GTP, 0.4 pM E. coli ribosomes, 0.15 mg/ml poly(U) and 0.3 pM EF-G was mixed with 50 pl charging mix containing 12 pM [3H]Phe-tRNAPhe (specific activity 574 cpdpmol), 3 pM phosphoenolpyruvate, 30 pg/ml pyruvate kinase, 5 mM ATP, 0.24 mM CTP in charging buffer and 0.13 pM EF-Tu:GTP in a total volume of 300 pl. Following incubation at 20°C 5O-pl samples were withdrawn at 5-min intervals, transferred to 1 ml 16% trichloroacetic acid, incubated at 90°C for 30 min, cooled on ice for at least 5 min and finally filtered. Active ribosomes were prepared according to Jelenc (331.

74 1

1.2 , 1 - yo 1 .- Y

p 0.8

9 0.6 a n : 0.4 3 0.2 2

._ -0

> ._

0 0 10 20 30 40 50 60 7(

TemD. ("CI 3

1 s.1 o5

v) v

> L 5-104

0 I . . ,

Fig. 3. Lineweaver-Burk plot for the intrinsic GTPase activity. (01, Wild-type EF-Tu; (A), H66A; (+), H118A; (m), H118E, The

, s-l .

Fig. 1. Thermal inactivation profile for EF-Tu:GDP. (01, Wild-

GDP-binding is given as the relation between the activity measured at temperature (Temp.) t , c,, and at O'C, c,.

type EF-Tu; (A), H66A; (*I, H118A; (m), H118E. The relative hydrolysis rate is given in mol liberated phosphate . mol E F - T ~ - I

- , 1.2 oo . 0- ' - p 0.8 5

2 0.6

: 0.4 - z 0.2

0

-

I-

> ._

0 IT

0 10 20 30 40 50 60 Temp. CC)

Fig. 2. Thermal inactivation profile for EF-'h::GTP. (O), Wild- type EF-Tu; (A), H66A; (*), H118A; (m), H118E. The relative GTP-binding is given as the relation between the activity measured at temperature (Temp.) t , c,, and at O'C, c,.

RESULTS Protein activity and stability

The purity of EF-Tu wild-type and mutant protein prepa- rations determined by SDSPAGE was at least 95%. The ac- tivity measured by the GDP-binding activity assay was in the range of 80-90%. Equilibrium with [3H]GDP was obtained after 20 min at 20 "C. The protein concentrations wherever given are in units of active EF-Tu as measured by the GDP- binding assay.

The thermostability of both the GDP and GTP forms was investigated by irreversible heat-induced inactivation. The stability of the two forms are slightly different. The temper- ature, &, at which the GDP-binding activity of EF-Tu is reduced to 50% is 43-46°C for the mutants and 52°C for the wild type (Fig. l ) , indicating that each different mutation induces a minor structural destabilization of the inactive form of EF-Tu. With respect to the GTP form, H66A has the same stability as the wild type, i.e. = 45"C, whereas 4,,, for H118A and H118E is reduced to 43°C and 37"C, respective- ly, which for the H118 mutation again indicates a minor structural destabilization (Fig. 2). To justify ignoring differ- ences in stability the assay temperature has been kept at max- imum of 20°C in the following assays.

GTPase activity The kinetics constants for the intrinsic GTPase activity

were determined from Lineweaver-Burk plots (Fig. 3 ) . The plots were constructed from initial rates ; i.e. during hydroly-

Table 1. Kinetics parameters for the intrinsic GTPase activity. K,,, and k,,, were determined from Lineweaver-Burk plots with R values (correlation coefficients) above 0.91. wt, wild type.

EF-Tu type Km k,,, X 1 Ob k,,, X 1 OhlKBn

S-' pM-1 . s-1 PM wt EF-Tu 7.8 141 18 H66A 6.0 92 15 H118A 7.3 107 15 H118E 5.1 111 22

sis of the first GTP molecule. Table 1 shows that K,,, values are almost identical for wild type and mutants, as are k,,, values. Table 2 shows for both wild-type and mutants a mod- erate stimulation of GTPase activity with Phe-tRNA, ribo- somes and kirromycin, when added separately, and a more pronounced effect in particular when kirromycin is combined with Phe-tRNA.

Characterizatidn of the ternary complex

The equilibrium dissociation constant Kd for formation of the ternary complex was for wild-type EF-Tu and H66A determined by the ribonuclease resistance assay (Fig. 4 and Table 3). The Kd value for H66A being 20-times higher than for the complex with wild-type EF-Tu. The wild-type Kd value of 11 nM is in the range of published values, which have been determined using the same or different methods (Table 3). The variation in published values is probably mainly due to the use of different methods. The ribonuclease resistance assay [2, 341 and non-enzymic hydrolysis protec- tion assay [35] might be considered to give more relative binding constants than the fluorescence technique [36].

The dissociation rate constant k-, was determined by the ribonuclease-dgestion rate assay from plots of In( [EF- Tu: GTP: Phe-tRNAPh"]N) versus time (Fig. 5) . The values for the mutants being up to about 50-times higher than wild- type values (Table 3).

The ability of EF-Tu to protect the labile aminoacyl bond against non-enzymic hydrolysis has been used as a relative measure of the affinity of EF-TU for Phe-tRNA. The time kinetics was followed in the deacylation process both in the presence of excess amounts of EF-Tu and in the absence of the protein (Fig. 6). For wild-type EF-Tu 50% deacylation takes place within 391 min, for H66A the half-life is 107 min

742

Table 2. Stimulation of the GTPase activity. The table indicates the rate of GTP hydrolysis. The last column gives the general qualitative effect for wild-type (wt) as well as mutants. The rates of stimulated GTPase activity were all determined with R values above 0.90. Although measured at high substrate concentrations the rates should only be regarded to give a semi-quantitative measure of V,,, as they have been determined from single substrate concentrations only.

Stimulant GTPase activity Effect

wt EF-TU H66 A H118A HI 18E

mol . mol-' . min-'

None Phe-tRNA Ribosomes Phe-tRNA + ribosomes Kirromycin Kirrornycin + Phe-tRNA Kirromycin + ribosomes Kirromycin + aminoacyl-tRNA

0.030 0.109 0.179 0.093 0.088 0.93 0.49

+ ribosomes 1.23

0.040 0.088 0.091 0.073 0.24 0.64 0.25 1.2

0.021 0.081 0.040 0.064 0.23 1.62 0.23 2.0

0.043 0.105 + 0.127 + 0.083 + 0.094 + 0.68 ++ 0.25 + 0.69 ++

40 -- 35

f 3 0

f 25

2 15

- - s 2 0

'L 10

= 2=

\

5

0 0.2 0.3 0.4 0;s 0.6 0.7 0.8

1

7- 0.9

0.8 - r, - -

Q: z 0.7 5

0.6

' 0.5

0. \ u

B A

0.4 1 I I I I I 0.15 0.18 0.21 0.24 0.27 0.3

Fig. 4. Scatchard plot for the binding of Phe-tRNAPhe to EF-m. A and B are for wild-type EF-Tu and H66A, respectively. I> denotes the average number of Phe-tRNAPhe molecules boundmolecule EF- Tu. The maximal value of v is given by the abscissa intercept. An intercept below one indicates that the protein preparation is not as active in tRNA binding, 70% for wild-type EF-Tu and 30% for H66A, as it is in GDP binding (set to 100%).

and for H118 mutations about 60 min, which is the same as for non-protected free Phe-tRNA.

In vitro translation assay Under the experimental conditions used in the poly(Phe)

synthesis assay there were no significant differences between wild-type EF-Tu, H66A and H118A. For H118E, the incor- poration rate was, however, considerably lower (Fig. 7).

DISCUSSION The thermostability of the mutants H66A, H118A and

H118E is identical to wild-type EF-Tu at the temperatures

Table 3. Kinetics parameters for the ternary complex. n.d., not determined. Kd for H118A and H118E could not be determined due to methodical limitations (see text). The R values were all above 0.91.

EF-TU k, X 10' Kd

S - ' nM wt EF-TU 0.25 11 H66A 3.4 220 H118A 10 n.d. H118E 12 n.d. Reported values 1.6 (21 0.94 [2], 79 [34],

20 [35], 0.60 [36]

'1 - * -4 'K I

\ - 3 ! x

I I I 0 50 100 150 200

Time (s)

Fig. 5. Dissociation rate for the ternary complex. (O), Wild-type EF-Tu; (A), H66A; (+), H118A; (W), H118E; (X), free aminoacyl- tRNA. c, and c, denotes the concentration of ternary complex at time t and 0, respectively. The dissociation rates are given in Table 3.

used in the assays. Also the GTPase activity, both intrinsic and stimulated, is conserved for the mutants, i.e. neither the nucleotide affinity nor the hydrolytic mechanism is affected by mutating the histidines. As far as formation of the ternary complex is concerned the importance of His66 and His118 becomes, however, most conspicuous. The rate constant k - , for dissociation of the ternary complex increases by factors of 14, 40 and 48 for H66A, H118A and H118E, respectively, when compared to wild type (Table 3 ) . An increase in k - , reflects a decrease in affinity for aminoacyl-tRNA and the results clearly prove the importance of the histidines in stabi- lization of the ternary complex, which could be effected

743

A

0.5

0

I

- 2

I I I I ; 50 100 150 200 Time (min)

Fig. 6. Protection of Phe-tRNAPh‘ within the ternary complex against non-enzymic hydrolysis. (o), Wild-type EF-Tu ; (A), H66A; (e), H118A; (W), H118E; (X), free aminoacyl-tRNA. c, and c, denote the concentration of ternary complex at time t and 0, respectively. Half-lives are 391, 107, 69, 54, 61 min for wild type, H66A, HI 18A, HI 18E and free aminoacyl-tRNA, respectively.

0

1 A

- 4 [/*

: / - I I I 1 I I

0 5 1 0 15 20 25 30 35 Time (min)

I I I 1 I I 0 5 1 0 15 20 25 30 35

Time (min)

Fig. 7. Activity of EF-lh in poly(Phe) synthesis. (O), Wild-type EF-Tu; (A), H66A; (+), H118A; (m), H118E.

through the formation of salt-bridgeshydrogen bonds to tRNA backbone phosphates. The drastic substitution of a his- tidine by a glutamate, H11 8E, could be argued to introduce an electrostatic repulsion towards the phosphates, however, k- , for mutating to the neutral alanine is almost as high and in this case the argument would not hold.

The results have been confirmed by measuring Kd for formation of the ternary complex. Kd increases by a factor of 20 when mutating His66 to alanine, proving that in addition to an increase in k- , the association rate k,, has also de- creased. The assay for determination of Kd is based on mea- suring the amount of ternary complex after removing free aminoacyl-tRNA by treating with RNase A. If k- , is high, the RNase treatment might disturb equilibrium. This is in fact what happens when attempting to determine Kd for aminoacyl-tRNA and either of the mutants H118A and H118E. The effect is so drastic that Kd cannot be determined at all by this method.

Also the half-lives for the non-enzymic deacylation of aminoacyl-tRNA in the ternary complex are in agreement with the k - , values. The affinity for aminoacyl-tRNA is al- most completely lost upon mutation of His118 and very much reduced upon mutation of His66.

From an evolutionafy point of view our results are fully comprehensible, since His118 is invariant and His66 is ex- tremely well conserved amongst procaryotes.

Using an in vitvo translation assay a decrease in synthesis rate would be expected for the mutants. However, a decrease is only observed for Hll8E (Fig. 7), which must be due to some rate-limiting factor other than the stability of the ter- nary complex in the instance of poly(Phe) synthesis with the mutants H66A and H118A. This was also apparent from the fact that poly(Phe) synthesis could not be stimulated by EF-

Fig. 8. Schematic diagram of the structure of EF-Tu:GTP from Thernius aquaticus. Using E. coli EF-Tu numbering the positions of His66, His118 and Arg58 are indicated and so is the domain numbering. The diagram was drawn using the MolScript program W I .

Ts. Probably the limiting factor is the activity of the ribo- somes.

Chemical modification studies [lS, 161 and cross-linking experiments [ 17, 181 have pointed to His66 and His1 18 being involved in the binding of aminoacyl-tRNA. Our kinetics data on the histidine mutants unequivocally verify these indications.

In the active GTP form of EF-Tu domains 1 and 2 form a deep cleft, which is not present in the GDP form [S, 61. This cleft is lined with several solvent-exposed basic residues and could be suggested to constitute part of a tRNA-binding site. The position and exposure of His66 speaks in favour of this theory as it is located at the bottom of this cleft with a completely solvent-exposed side chain (Fig. 8). The solvent- accessible area has been measured to about 1.10 nm’ in both the GDP and GTP form (Nyborg, J., personal communicat- ion).

In both structural forms the side chain of His118 forms hydrogen bonds to the backbone nitrogen of Gly18, for the GDP form also to the side chain of His84 and for the GTP form to a water molecule buried within the structure (Kjeld- gaard, M., personal communication). The hydrogen-bonding pattern to His118 is broken upon substitution by alanine or glutamate and explains why the thermostabilities of the mu- tants are reduced (Figs 1 and 2). From a structural point of view the role of His118 is unclear. The solvent-accessible area for His118 has been measured to approximately 0 nm2 for the two structural forms [5]. As His118 is completely buried in the interior of domain 1 (Fig. 8) [S, 61, its direct participation in tRNA binding thus implies the induction of yet another EF-Tu conformation. In a classical sense it would represent an induced-fit mechanism.

Arg58 forms a salt bridge to Asp86 in the GTP form [6]. Breakage of the salt bridge has been suggested to be one of the steps in stimulation of GTP hydrolysis [6] and could be the trigger for induction of a tRNA-binding conformation, including an exposure of Hisll8. This could place the tRNA-

744

acceptor stem close to Hisll8. Such an interpretation is in accordance with our kinetics data on the non-enzymic de- acylation, which indicates that His118 is most strongly in- volved in the specific protection of the aminoacyl bond.

If the overall domain organization is conserved in a theo- retical new tRNA-binding conformation, part of the tRNA, i.e. the anticodon stem, may still be located in the ‘tRNA- binding cleft’ between domains 1 and 2.

We thank Karen Margrethe Nielsen for skillful technical assis- tance, Dr Jens Nyborg and Prof. B. F. C. Clark for valuable discus- sions and comments on the manuscript, and Dr Morten Kjeldgaard for providing Fig. 8. The work was funded by the Danish Biotech- nology programme (PERC) and the NOVO fund.

REFERENCES 1.

2. 3.

4.

5.

6.

7.

8.

9.

10.

11.

12. 13.

Miller, D. L. & Weissbach, H. (1977) in Molecular mechanisms of protein biosynthesis (Weissbach, H. & Petska, S., eds) pp. 323-373, Academic Press, New York.

Louie, A. & Jurnak, F. (1985) Biochemistry 24, 6433-6439. Faulhammer, H. G. & Joshi, R. L. (1987) FEBS Lett. 217,203-

211. Kjeldgaard, M. & Nyborg, J. (1992) J. Mol. Biol. 223, 721-

742. Kjeldgaard, M., Nissen, P., Thirup, S. & Nyborg, J. (1993)

Structure I , 35-49. Berchtold, H., Reshetnikova, L., Reiser, C. 0. A., Schirmer, N.

K., Sprinzl, M. & Hilgenfeld, R. (1993) Nature 365, 126- 132.

Wiborg, O., Andersen, C., Knudsen, C. R., Kristensen, T. J. & Clark, B. F. C. (1993) Biotechnol. Appl. Biochem. 19, 3-15.

Clark, B. F. C., Kjeldgaard, M., Barciszewski, J. & Sprinzl, M. (1993) in tRNA (RajBhandary, T. & Soll, D., eds) American Society for Microbiology, Washington DC, in the press.

Jonik, J., Petersen, T. E., Clark, B. F. C. & Rychlik, I. (1982) FEBS Lett. 150,485-488.

Jon&, J., Smrt, J., Holy, A. & Rychlik, I. (1980) Eul: J. Bio- chem. 105, 315-320.

Anborgh, P. H., Parmeggiani, A. & Jon&, J. (1992) Eul: J. Bio- chem. 208, 251-257.

Kraal, B. & Hartley, B. S. (1978) J . Mol. Biol. 124, 551-564. Antonsson, B. & Leberman, R. (1984) Eul: J. Biochem. 141,

483-487.

14. Marschel, A. H. & Bodley, J. W. (1980) Arch. Biochem. Bio- phys. 203, 489-495.

15. Jon& J., Petersen, T. E., Meloun, B. & Rychlik, I. (1984) Eul: J. Biochem. 144, 295-303.

16. Jon&, J. & Rychlil, I. (1987) Biochim. Biophys. Acta 908, 97- 102.

17. Duffy, K. L., Gerber, L., Johnson, A. E. & Miller, D. L. (1981) Biochemistry 20, 4663-4666.

18. Metz-Boutigue, M.-H., Reinbolt, J., Ebel, J.-P., Ehresmann, C. & Ehresmann, B. (1989) FEBS Lett. 245, 194-200.

19. Taylor, J. W., Ott, J. & Eckstein, F. (1985) Nucleic Acids Res.

20. Knudsen, C. R., Clark, B. F. C., Degn, B. & Wiborg, 0. (1992)

21. Smith, D. B. &Johnson, K. S. (1988) Gene (Amst.) 67, 31-40. 22. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular

cloning: a laboratory manual, 2nd edn, Cold Spring Harbor Laboratory, Cold Spring Harbor NY.

23. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl Acad. Sci. USA 74, 5463-5467.

24. Murray, V. (1989) Nucleic Acids Res. 17, 8889. 25. Sedmark, J. J. & Grossberg, S. E. (1977) Anal. Biochem. 79,

544-552. 26. Miller, D. L. & Weissbach, H. (1974) Methods Enzymol. 30,

219-232. 27. Gumusel, F., Cool, R. H., Weijland, A., Anborgh, P. H. & Par-

meggiani, A. (1990) Biochim. Biophys. Acta 1050, 215-221. 28. Parmeggiani, A. & Sander, G. (1981) Mot. Cell. Biochem. 35,

129- 158. 29. von der Haar, F. (1979) Methods Enzymol. 59, 257-267. 30. Scatchard, G. (1949) Ann. NYAcad. Sci. 51, 660-672. 31. Pingoud, A. & Urbanke, C. (1979) Anal. Biochem. 92, 123-

32. Swart, G. & Parmeggiani, A. (1987) Biochemistry 26, 2047-

33. Jelenc, P. C. (1979) Anal. Biochem. 105, 369-374. 34. Tanada, S., Kawakami, M., Yoneda, T. & Takemura, S. (1981)

J. Biochem. (Tokyo) 89, 1565-1572. 35. Pingoud, A,, Urbanke, C., Krauss, G., Peters, F. & Maass, G.

(1977) Eul: J. Biochem. 78,403-409. 36. Ott, G., Schiesswohl, M., Kiesewetter, S., Forster, C., Arnold,

L., Erdman, V. A. & Sprinzl, M. (1990) Biochim. Biophys. Acta 1050, 222-225.

13, 8764-8785.

Biochem. Int. 28, 353-362.

127.

2054.

37. Kraulis, P. J. (1991) J. Appl. Crystallogr. 24, 946-950.